New optical and radiocarbon dates from Ngarrabullgan Cave, a Pleistocene archaeological site in Australia: implications for the comparability of time clocks and for the human colonization of Australia.

David, Bruno ; Roberts, Richard ; Tuniz, Claudio 等

The human settlement of Australia falls into that period where dating is hard because it is near or beyond the reliable limit of radiocarbon study; instead a range of luminescence methods are being turned to (such as thermoluminescence at Jinmium: December 1996 ANTIQUITY). Ngarrabullgan Cave, a rock-shelter in Queensland, now offers a good suite of radiocarbon determinations which match well a pair of optically stimulated luminescence (OSL) dates - encouraging sign that OSL determinations can be relied on.

Archaeologists have been divided over the antiquity of human occupation in Australia and Papua New Guinea since the publication of 50,000-60,000-year-old thermoluminescence (TL) and optically stimulated luminescence (OSL) dates from northern Australia (Roberts et al. 1990; 1994b). One group has argued that this time-range approximates the antiquity of habitation on this continent (Chappell et al. 1996; Roberts et al. 1994a; Roberts & Jones 1994). Another has argued that these dates are erroneous or not directly comparable to the 14C chronology from most other Australian sites (Allen 1994; Allen & Holdaway 1995); they argue that there is as yet no conclusive evidence for human occupation in Australia pre-dating 40,000 radiocarbon years b.p. Recently reported TL dates of 116,000[+ or -]12,000 years BP and beyond for artefact-bearing sands (Fullagar et al. 1996) have further lengthened the time-frame at issue, although a number of chrono-stratigraphic uncertainties make these results open to various possible interpretations. No paired 14C/optical dates beyond 30,000 b.p. have hitherto been available in Australia, partly because stratified sites possessing appropriate sediments and charcoal are rare. Charcoal-rich deposits within a silica-rich sandy matrix now enable two Pleistocene optical age determinations to be set alongside a suite of radiocarbon determinations from Ngarrabullgan(1) Cave in northern Australia.

Ngarrabullgan: site and stratigraphy

Ngarrabullgan is a large (18x6 km) table-top mountain 100 km west of Cairns in north Queensland. Rising 200-400 m above the surrounding plains and hills, it is bounded by high cliffs. Ngarrabullgan Cave is the largest archaeological site on top of the mountain, from which cultural deposits more than 37,000 radiocarbon years old have been obtained (David 1993). The deposits at this site are very dry, and there have been few terrestrial mammals on the mountain-top to disturb sediments; the integrity of strata and the preservation of organic materials [TABULAR DATA FOR TABLE 1 OMITTED] (especially charcoal and microscopic residues on stone tools) are exceptional (Fullagar & David in press).

Dating

The rock-shelter deposit is finely stratified but shallow; it possesses 27 distinct strata over a depth of only 43.5 cm, the top 35.7 cm of which is cultural [ILLUSTRATION FOR FIGURE 1 OMITTED]. Twenty-one 14C dates obtained by accelerator mass spectrometry (AMS) and four radiocarbon dates obtained by beta-counting are listed in TABLE 1.

Radiocarbon determinations

All of the Australian Nuclear Science and Technology Organisation (OZB) AMS 14C dates are on single pieces of charcoal; the Beta Analytic beta-counting radiocarbon results are on multiple pieces of charcoal (TABLE 1). The results show that the period of occupation before 37,000 b.p. may be restricted to a single, thin stratum (3E; [ILLUSTRATION FOR FIGURE 1 OMITTED]) consisting of sparse cultural deposits; this initial period of occupation was followed by one or more episodes of low intensity occupation at [approximately]32,500 radiocarbon years b.p. The site was then abandoned and appears not to have been re-occupied until [approximately]5400 b.p. The 14C dates from substrata 3A-3D exhibit a high degree of self-consistency, with none of the individual pieces of dated charcoal being intruded from the overlying Holocene units. This implies that stratum 3, or at least the extant portion, has remained largely undisturbed since deposition. The 18 AMS 14C determinations made on individual pieces of charcoal from this unit present a rare opportunity to conduct a statistical examination of age data for a single archaeological datum of Pleistocene antiquity [ILLUSTRATION FOR FIGURE 2 OMITTED]. The data appear to approximate a Gaussian distribution, suggesting that the charcoal pieces are indeed contemporaneous and that their range of ages can be satisfactorily explained in terms of random scatter.

The weighted mean age (Aitken 1990) for the 20 finite 14C age determinations from stratum 3 is 32,540[+ or -]110 years b.p. and the arithmetic average is 32,500[+ or -]360 years b.p. (where the error is characterized by the standard error on the mean; the standard deviation of the distribution is [+ or -]1630 years). Excluding the two most outlying ages (28,800 and 36,100 b.p.) has a negligible effect on these averages (32,510[+ or -]110 and 32,500[+ or -]280 b.p., respectively, with a standard deviation of the distribution of[+ or -]1180 years). The best estimate of the 14C age of substrata 3A - 3D is [approximately]32,500 years b.p.; a finite age for the underlying substratum 3E is not available.

Optical age determinations

The 14C ages of stratum 3 were cross-checked by optical dating (Aitken 1994; Huntley et al. 1985; Wintle 1993) of its quartz sedimentary component. A sample ([ANU.sub.OD]122a) was collected in two 1.8-cm diameter tubes from substratum 3C and the uppermost part of substratum 3E [ILLUSTRATION FOR FIGURE 1 OMITTED]. Due to the small quantity of quartz present in these units, the contents of both tubes were combined to provide a sufficient mass of quartz for dating. The sample was processed in subdued red light, and the quartz fraction was isolated after sample. treatment with hydrogen peroxide, hydrochloric acid, fluorosilicic acid, fluoboric acid, and the removal of heavy minerals using a sodium polytungstate solution. Quartz grains of 90-125 [[micro]meter] diameter were obtained by dry sieving and finally etched in 40% hydrofluoric acid for 45 minutes to remove the [Alpha]-irradiated outer rinds. 90Sr/90Y sources were used to deliver calibrated laboratory [Beta] doses ([+ or -]3% uncertainty at the 2[Sigma] level). Optical stimulation was provided by the green/blue wavelengths from a filtered tungsten-halogen lamp and the ultraviolet portion of the emitted OSL was measured; experimental details are given in Murray et al. (in press).

The [Gamma] dose rate from the decay of radio-nuclides in the 238U and 232Th chains, and 40K, was determined using a portable [Gamma]-ray spectrometer and high-resolution [Gamma]-ray spectrometry (Murray et al. 1987) of the surrounding material. The effective [Beta] dose rates were deduced from the high-resolution [Gamma]-ray spectrometry measurements and a [Beta] attenuation factor of 0.93[+ or -]0.03 (Mejdahl 1979). The dose rate due to the internal uranium and thorium content of the quartz grains was estimated from laser-ablation ICP-MS measurements on some etched grains, an assumed [Alpha]-efficiency "a" value of 0.10[+ or -]0.05, and the [Beta]-absorption factors of Mejdahl (1979). The dose rate conversion factors of Nambi & Aitken (1986) and Olley et al. (1996) were used, and the measured water content of the sample (2[+ or -]1%) was assumed to have prevailed throughout the period of sample burial. The cosmic ray dose rate was estimated (Prescott & Hutton 1994), making allowance for the rock thickness of the cave roof, geomagnetic latitude, altitude, and geomagnetic field changes during the late Quaternary.

Two methods of palaeodose determination were employed. The palaeodose obtained using the multiple-aliquot additive-dose method (Aitken 1990) was 47[+ or -]4 Gy, using 52 aliquots (each aliquot being normalized by an initial 0.1 second light exposure), a preheat treatment of 220 [degrees] C for 300 seconds (Rhodes 1988; Roberts et al. 1994c), and a saturating exponential dose-response function (Brumby 1992). A newly-devised single-aliquot additive-dose protocol (Murray et al. in press) was also applied to this sample, giving a mean palaeodose of 48[+ or -]2 Gy from six individual palaeodose determinations in the preheat temperature range 280-300 [degrees] C (held at these temperatures for 10s). The agreement between the multiple-aliquot and single-aliquot palaeodoses, and the small amount of scatter among the single-aliquot palaeodose estimates, indicates that there is no significant contamination of this sample due to the incorporation of decomposing bedrock. This posed a potentially serious problem for this sample as it was located only 5 cm above bedrock which might have liberated quartz grains whose OSL signal had never been zeroed by sunlight exposure since rock formation. Roberts & Jones (1994) had warned of such problems at rock-shelters where rubble and saprolite (in situ weathered bedrock) are present within, or in close proximity to, artefact-bearing deposits. Such 'contamination' may account for the discrepancies reported between 14C and luminescence chronologies at Puritjarra rock-shelter in central Australia (Smith et al. in press), and may similarly be a factor contributing to the old TL dates reported from Jinmium (Fullagar et al. 1996); in neither study was the possibility of contamination examined using single-aliquot methods (Duller 1995; Murray et al. in press), which have been extended recently to the dating of single grains (Lamothe et al. 1994; Murray & Roberts submitted). Such investigations are now in progress.

High-resolution [Gamma]-ray spectrometry of the sample material showed no evidence of dis-equilibrium in the 238U or 232Th decay chains (the activities of 238U, 226Ra, and 210Pb in the uranium chain being 23.9[+ or -]4.9, 18.1[+ or -]0.4, and 17.4[+ or -]2.9 Bq [kg.sup.-1] respectively, while those of 228Ra and 228Th in the thorium chain were 24.5[+ or -]0.9 and 24.4[+ or -]0.5 Bq [kg.sup.-1] respectively; 40K activity was 127.3[+ or -]6.0 Bq [kg.sup.-1]). The [Gamma]-ray dose rate measured in situ was, therefore, used in the age calculation to accommodate any inhomogeneity in the deposit; the in situ and high-resolution [Gamma]-ray dose rates (both for 2% water content) differed by less than 3%. The total dose rate was 1.38[+ or -]0.06 mGy [year.sup.-1], comprising 0.67, 0.56, 0.13 and 0.02 mGy [year.sup.-1] from [Beta], [Gamma], cosmic ray, and internal [Alpha] and [Beta] radiation, respectively.

Using the multiple-aliquot palaeodose estimate, the OSL age is calculated as 34,100[+ or -]3100 years BP (the latter 1[Sigma] total uncertainty being the quadratic sum of the random and systematic errors; Aitken 1990). This result is shown as [ANU.sub.OD]122a-M in FIGURE 1 and TABLE 1. The age obtained using the single-aliquot protocol ([ANU.sub.OD]122a-S) is 34,900[+ or -]2100 years BP. The two estimates are statistically indistinguishable at the 1[Sigma] level and give a combined (weighted) age of 34,700[+ or -]2000 years BP; the palaeodoses can be combined in this manner, having been obtained using different OSL equipment and laboratory sources calibrated against different international standards.

Paired age comparison

The best estimate of the optical age of sample AN[U.sub.OD]122a is 34,700[+ or -]2000 years BP, which is [approximately]2200 years greater than the mean 14C age for stratum 3 ([approximately]32,500 years b.p.). A similar degree of 14C-age underestimation for the 30,000-40,000 b.p. period has been reported previously on the basis of 234U/230Th ages on corals (Bard et al. 1993), TL ages on fireplaces (Bell 1991) and burnt flints (Boeda et al. 1996), and geomagnetic intensity variations (Guyodo & Valet 1996; Laj et al. 1996). When calibrated for changes in the 14C production rate, the AMS 14C ages for stratum 3 are consistent with the optical date. The agreement indicates that the two time clocks are broadly comparable over at least this time period (see also Smith et al. in press). By implication, the 50,000-60,000-year-old TL and optical dates from unheated sediments at Malakunanja II (Roberts et al. 1990) and Nauwalabila I (Roberts et al. 1994b) do not equate with 14C dates of less than 40,000 years b.p., as has been suggested (Allen 1994). The implication of the Malakunanja II and Nauwalabila I luminescence dates, coupled with the results presented here, is that Australian prehistory is considerably older than 39,700[+ or -]1000 years b.p., which is currently the oldest reliable radiocarbon evidence for human presence on the continent (O'Connor 1995).

Acknowledgements. We thank the Kuku Djungan Aboriginal Corporation for encouraging this research; Earthwatch, AIATSIS and AINSE for funding it; and the University of Queensland for a Post-Doctoral Fellowship to BD for this project. RR is currently in receipt of a Queen Elizabeth II Fellowship from the Australian Research Council. I. Faulkner kindly drew FIGURE 1. We thank N. Spooner and A. Murray for help with OSL equipment, A. Murray for access to high-resolution [Gamma]-ray spectrometry facilities, M. McCulloch for access to laser-ablation ICP-MS facilities, and E. Lawson, Q. Hua and G. Jacobsen for AMS radiocarbon analyses. We also thank J. Allen, R. Fullagar, S. Holdaway and A. Wintle for comments on earlier drafts of this paper.

1 Following instructions from local Aboriginal Elders, the spelling of Ngarrabullgan has changed over the years (Nurrabullgin, Ngarrabullgin). We will henceforth use the present spelling.

References

AITKEN, M.J. 1990. Science-based dating in archaeology. Longman: London.

1994. Optical dating: a non-specialist review, Quaternary Science Reviews 13: 503-8.

ALLEN, J. 1994. Radiocarbon determinations, luminescence dating and Australian archaeology, Antiquity 68: 339-43.

ALLEN, J. & S. HOLDAWAY. 1995. The contamination of Pleistocene radiocarbon determinations in Australia, Antiquity 69: 101-12.

BARD, E., M. ARNOLD, R.G. FAIRBANKS & B. HAMELIN. 1993. 230Th-234U and 14C ages obtained by mass spectrometry on corals, Radiocarbon 35: 191-9.

BELL, W.T. 1991. Thermoluminescence dates for the Lake Mungo Aboriginal fireplaces and the implications for radiocarbon dating, Archaeometry 33: 43-50.

BOEDA, E., J. CONNAN, D. DESSORT, S. MUHESEN, N. MERCIER, H. VALLADAS & N. TISNERAT. 1996. Bitumen as a hailing material on Middle Palaeolithic artefacts, Nature 380: 336-8.

BRUMBY, S. 1992. Regression analysis of ESR/TL dose-response data, Nuclear Tracks and Radiation Measurements 20: 595-9.

CHAPPELL, J., J. HEAD & J. MAGEE. 1996. Beyond the radiocarbon limit in Australian archaeology and Quaternary research, Antiquity 70: 543-53.

DAVID, B. 1993. Nurrabullgin Cave: preliminary results from a pre-37,000 year old rockshelter, Archaeology in Oceania 28: 50-54.

DULLER, G.A.T. 1995. Luminescence dating using single aliquots: methods and applications, Radiation Measurements 24: 217-26.

FULLAGAR, R. & B. DAVID. In press. Defining site use over [greater than]37,000 years of occupation at Ngarrabullgan Cave (Australia): residue and use wear analyses of stone artefacts, Cambridge Archaeological Journal.

FULLAGAR, R., D. PRICE & L. HEAD. 1996. Early human occupation of northern Australia: archaeology and thermoluminescence dating of Jinmium rock-shelter, Northern Territory, Antiquity 70: 751-73.

GUYODO, Y. & J-P. VALET. 1996. Relative variations in geomagnetic intensity from sedimentary records: the past 200,000 years, Earth and Planetary Science Letters 143: 23-36.

HUNTLEY, D.J., D.I. GODFREY-SMITH & M.L.W. THEWALT. 1985. Optical dating of sediments, Nature 313: 105-7.

LAJ, C., A. MAZAUD & J.C. DUPLESSY. 1996. Geomagnetic intensity and 14C abundance in the atmosphere and ocean during the past 50 kyr, Geophysical Research Letters 23: 2045-8.

LAMOTHE, M., S. BALESCU & M. AUCLAIR. 1994. Natural IRSL intensities and apparent luminescence ages of single feldspar grains extracted from partially bleached sediments, Radiation Measurements 23: 555-61.

MEJDAHL, V. 1979. Thermoluminscence dating: beta-dose attenuation in quartz grains, Archaeometry 21: 61-72.

MURRAY, A.S., R. MARTEN, A. JOHNSTON & P. MARTIN. 1987. Analysis for naturally occurring radionuclides at environmental concentrations by gamma spectrometry, Journal of Radio-analytical and Nuclear Chemistry, Articles 115: 263-88.

MURRAY, A.S. & R.G. ROBERTS. Submitted. Determining the burial time of single grains of quartz using optically stimulated luminescence, Earth and Planetary Science Letters.

MURRAY, A.S., R.G. ROBERTS & A.G. WINTLE. In press. Equivalent dose measurement using a single aliquot of quartz, Radiation Measurements.

NAMBI, K.S.V. & M.J. AITKEN. 1986. Annual dose conversion factors for TL and ESR dating, Archaeometry 28: 202-5.

O'CONNOR, S. 1995. Carpenter's Gap rockshelter 1:40,000 years of Aboriginal occupation in the Napier Ranges, Kimberley, WA, Australian Archaeology 40: 58-9.

OLLEY, J.M., A. MURRAY & R.G. ROBERTS. 1996. The effects of disequilibria in the uranium and thorium decay chains on burial dose rates in fluvial sediments, Quaternary Science Reviews 15: 751-60.

PRESCOTT, J.R. & J.T. HUTTON. 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations, Radiation Measurements 23: 497-500.

RHODES, E.J. 1988. Methodological considerations in the optical dating of quartz, Quaternary Science Reviews 7: 395400.

ROBERTS, R.G. & R. JONES. 1994. Luminescence dating of sediments: new light on the human colonisation of Australia, Australian Aboriginal Studies 1994 (2): 2-17.

ROBERTS, R.G., R. JONES & M.A. SMITH. 1990. Thermoluminescence dating of a 50,000 year-old human occupation site in northern Australia, Nature 345: 153-6.

1994a. Beyond the radiocarbon barrier in Australian prehistory, Antiquity 68: 611-16.

ROBERTS, R.G., R. JONES et al. 1994b. The human colonisation of Australia: optical dates of 53,000 and 60,000 years bracket human arrival at Deaf Adder Gorge, Northern Territory, Quaternary Science Reviews 13: 575-83.

ROBERTS, R.G., N.A. SPOONER & D.G. QUESTIAUX. 1994c. Palaeodose underestimates caused by extended-duration preheats in the optical dating of quartz, Radiation Measurements 23: 647-53.

SMITH, M.A., J.R. PRESCOTT & M.J. HEAD. In press. Comparison of 14C and luminescence chronologies at Puritjarra rock shelter, central Australia, Quaternary Science Reviews.

WINTLE, A.G. 1993. Recent developments in optical dating of sediments, Radiation Protection Dosimetry 47: 627-35.