Material and Methods

Twenty bear teeth from different levels of the Azokh 1 cave were analyzed by amino acid racemization. The Biomolecular Stratigraphy Laboratory (BSL) uses dentine for amino acid racemization dating of vertebrates. Bones are rejected because they are more prone to diagenetic interference (Masters 1986, 1987). Dentine collagen samples were obtained by drilling the root of the teeth with a dental diamond drill. A hole 2–3 mm in diameter was drilled near the tooth neck to reach the dentine, which is protected by the crown. Between 5 and 46 mg of dentine were obtained. The outermost part of the root (mostly cementum) was discarded. For the establishment of the chronology of Azokh Cave, we used only the aspartic acid content of the samples, because that racemizes fastest. Goodfriend (1991) noted that the analysis of more than one amino acid provides largely redundant information on sample age (Torres et al. 2002).

For details on sample pretreatment and amino acid extraction, see Kaufman and Manley (1998) and Kaufman (2000). The samples were pretreated with 2 N HCl at room temperature and a posterior dialysis step (Spectra/Por mnco 3500 D membrane) to eliminate dissolved mineral fraction and free amino acids (Marzin 1990; Torres et al. 1999, 2000). Subsequently, hydrolysis was performed under N2 atmosphere in 7 μl of 6 M HCl for 20 h at 100 °C. The hydrolysates were evaporated to dryness in vacuo, and then rehydrated in 7 μl 0.01 M HCl with 1.5 mM sodium azide and 0.03 mM L-homo-arginine (internal standard). For derivatization, the samples were mixed (2 μl) with the pre-column derivatization reagent (2.2 μl), which comprised 260 mM isobutyryl-L-cysteine (chiral thiol) and 170 mM o-phtaldialdehyde, dissolved in 1.0 M potassium borate buffer solution at pH 10.4. Eluent A consisted of 23 mM sodium acetate with 1.5 mM sodium azide and 1.3 mM EDTA, adjusted to pH 6.00 with 10 M sodium hydroxide and 10% acetic acid. Eluent B was HPLC-grade methanol, and eluent C consisted of HPLC-grade acetonitrile.

The amino acid concentrations and ratios were measured with an Agilent HPLC-1100, equipped with a fluorescence detector. Excitation and emission wavelengths were programmed at 335 nm and 445, respectively. A Hypersil BDS C18 reverse-phase column (5 μm; 250 × 4 mm i.d.) was used for the analysis. A linear gradient was obtained at 1.0

ml/min and 25 °C, from 95% eluent A and 5% eluent B upon injection to 76.6% eluent A, 23% eluent B, and 0.4% eluent C at min 31.

Results

The results of the individual analyses of the Azokh samples are shown in Table 16.3. The aspartic acid content can be used as an indicator of collagen integrity for further DNA analysis: low aspartic acid concentrations indicate deep degradation of the organic matrix of the dentine and, therefore, low DNA amount. The D/L ratios of other amino acids are not provided because in many cases the D enantiomers could not be identified in the chromatograms. The mean aspartic acid racemization ratios of the different levels are shown in Table 16.3.

Samples LEB-4685, 4686, 4490 and 4491 (Table 16.3) did not provide enough collagen to determine their amino acid content. Samples marked with an asterisk were not considered for the age calculation because of their high deviation from mean values, although we cannot rule out that they were affected by re-working processes. According to our own experience in the Amutxate cave in Spain (Torres et al. 2007), reworking of sediments and bones in ancient deposits constitutes the norm and not the exception.

The aspartic acid D/L ratios of dentine collagen samples were introduced in the age calculation algorithm (Table 16.3). For the age-calculation algorithm (Fig. 16.6), we used the racemization values from eight cave-localities dated through different radiometrical methods: 14C in bones (Eirós, Galicia; Grandal d’Anglade and Vidal Romaní 1997), Th/U in speleothems (La Lucia, Cantabria; Torres et al. 2001), electron spin resonance (ESR) and uranium series in bear teeth (Sima de los Huesos, Burgos; Bischoff et al. 1997), ESR results on bear teeth from Amutxate (Torres et al. 2007) and unpublished ESR data obtained from bear teeth (Navarra; Troskaeta and Santa Isabel, Vizcaya; La Lucia and La Pasada, Cantabria). We have calculated the mean age from the individual values obtained from each sample, and the age uncertainty is the standard deviation (Table 16.4, Fig. 16.6). In this sense, the use of monogeneric samples reduces taxonomically controlled variability in D/L ratios (Murray-Wallace 1995; Murray-Wallace and Goede 1995).

Table 16.4 Mean amino acid racemization aspartic acid D/L values obtained in the dentine of bear teeth from Units II and V together with their correspondent mean numerical age

Conclusions

Possible diagenetic processes linked to acid leaching and to guano accumulation could be responsible for the individual age scattering. The time-average of Units II and V of Azokh cave has been calculated and we can provisionally conclude that both correspond to the later part of the Middle Pleistocene.

Uranium-Lead (U-Pb) Dating

of Stalagmites from the Azokh Cave Complex (Robyn Pickering)

Introduction

Two stalagmite samples from the Azokh Cave Complex were selected for U-Pb dating (Table 16.1). The first attempt (sample 1) was focused on the large stalagmite boss near the entrance of the cave, but this material contained too much common 206Pb for an age determination to be made. During a second field season, a number of small stalagmites at the very back of the cave were considered, and the best preserved one selected for dating (sample 2). This second

sample consisted of clear to creamy coloured calcite and produced an age of 1.19 ± 0.08 Ma. All sample preparation and dating was undertaken at the School of Earth Sciences at the University of Melbourne, Australia.

Laser-ablation Pre-screening

Successful U-Pb dating depends on the concentration of U present in the sample, as well as the amount of Pb. Too much common 206Pb can mask the radiogenic daughter 206Pb, making it impossible to date the material in question. Laser ablation ICP-MS is used to map out the U and Pb concentrations in samples prior U-Pb analysis so that layers with high U and low Pb can be identified and selected for dating. Speleothem samples are cut, set in resin and polished into 10 × 5 × 1 cm blocks to fit into the laser cell. The laser ablation results of samples 1 and 2 are shown in Fig. 16.7. From the laser scans it is clear that sample 1 (Fig. 16.7a) is not suitable for U-Pb dating, given the dominance of the Pb signal and the low U content. Sample 2 (Fig. 16.7b), however, is highly suitable for dating, with several layers with U concentrations of close to 10 ppm, and very low Pb. Based on these results, dating was focused on sample 2.

Sample Preparation and Measurement

Once a U-rich layer has been identified from the laser ablation tracks, small (*3 mm3) blocks of speleothem

a

b

material are cut using a hand operated dentist drill. These small blocks are then etched in a mild HCl solution to remove the outer layer and decrease the risk of Pb contamination from the drilling and subsequent handling of the samples. Samples are spiked with a mixed 235U–205Pb spike and dissolved in 6 M HCl and left to equilibrate on a hot plate overnight. Uranium and lead are extracted and concentrated using a standard ion-exchange resin column separation, following the protocol outlines in Woodhead et al. (2006).

Uranium and lead from each of the multiple aliquots from the single U-rich layer are then measured on a Nu-Instruments MC-ICP-MS, again following Woodhead et al. (2006).

Results

Element concentrations and isotope ratios are obtained for sample 2 and are summarized in Table 16.5. The average U concentration for the layers analysed is 5.7 ppm, while the Pb is much lower at 0.03 ppm. 238U/206Pb ratios vary

between 389 and 849, giving enough spread to produce a range of 207Pb/206Pb ratios from 0.711 to 0.772. An age for sample 2 was calculated using the 238U/206Pb and

207Pb/206Pb ratios to construct a Tera-Wasserberg isochron (Fig. 16.8). Ideally the individual analyses should plot along a straight line, the slope of which is a function of the age of