A Middle Palaeolithic wooden digging stick from Aranbaltza III, Spain

Aranbaltza is an archaeological complex formed by at least three open-air sites. Between 2014 and 2015 a test excavation carried out in Aranbaltza III revealed the presence of a sand and clay sedimentary sequence formed in floodplain environments, within which six sedimentary units have been identified. This sequence was formed between 137–50 ka, and includes several archaeological horizons, attesting to the long-term presence of Neanderthal communities in this area. One of these horizons, corresponding with Unit 4, yielded two wooden tools. One of these tools is a beveled pointed tool that was shaped through a complex operational sequence involving branch shaping, bark peeling, twig removal, shaping, polishing, thermal exposition and chopping. A use-wear analysis of the tool shows it to have traces related with digging soil so it has been interpreted as representing a digging stick. This is the first time such a tool has been identified in a European Late Middle Palaeolithic context; it also represents one of the first well-preserved Middle Palaeolithic wooden tool found in southern Europe. This artefact represents one of the few examples available of wooden tool preservation for the European Palaeolithic, allowing us to further explore the role wooden technologies played in Neanderthal communities.


Sample preparation
In 2015, at the archaeological site of Aranbaltza OSL samples were collected by hammering PVC tubes (5x20 cm) into the freshly cleaned profiles. The material at the ends of each tube (~3 cm deep) was removed under dark room conditions to prevent light contamination during sampling. The material not exposed to light was wet-sieved, treated with concentrated HCl and H 2 O 2 to remove carbonate and organic matter, respectively. Two steps of heavy liquid separation at 2.62 g·cm -3 and 2.72 g·cm -3 were carried out to remove the feldspar fraction and heavy minerals, respectively (Aitken, 1985;1998). A magnetic separator was used to remove magnetic minerals from the samples following the protocol described in Porat, 2006.
The quartz-rich fraction was then etched for 40 minutes using 48% HF to remove any remaining feldspars and etch away the outer alpha irradiated layer. Following etching, the quartz fraction was treated with concentrated HCl for 60 minutes to remove any possible precipitated fluorides. After that, the sample was treated with a sodium pyrophosphate solution in an ultrasonic bath during 30 minutes, and washed several times to remove micas from the quartz-rich fraction.
For this study, a total of four luminescence samples were collected from the Aranbaltza III section. All of them yielded a good amount of quartz-rich fraction (more than 1 g). For single grain measurements, the quartz grains were mounted on aluminum discs with 100 holes with a 0.3 mm diameter.
The environmental dose rate determination was performed using a field gamma spectrometer (Canberra InSpector 1000) equipped with a NaI(Tl) probe, except for sample AZ15OSL-03 where the dose rate was obtained by high resolution gamma spectrometry using a High Purity Germanium Detector (HPGD). This sample was ground and packed into a gamma cup (Ø = 75 mm) and stored for four weeks to allow 222 Rn to build up equilibrium with 226 Rn before the measurements. The concentration values obtained were converted to dry beta and gamma dose rates using the conversion factors given in Guérin et al. (2011). Water contents for age calculation were between 24.1 % and 26.1 % in all of the samples. An internal quartz dose rate of 0.02 Gy·ka -1 (Vandenberghe et al., 2007) and a contribution from cosmic rays (Prescott and Hutton, 1994) were also incorporated into the total dose rate. The total environmental dose rates are listed in S3 Table 1.

Equipment and methods
Optical stimulation was carried out on a Risø TL/OSL Reader Model DA20 with a single grain attachment (green laser 532 nm; 10 mW). Luminescence was recorded using a photomultiplier tube (9235QB15), equipped with a 7mm Hoya-U340 filter. Samples were irradiated using a calibrated 90 Sr/ 90 Y source incorporated in the reader, with an effective dose rate of 0.11 ± 0.01 Gy·s -1 .
The natural OSL signals from the four samples listed in S3 Table 1 have been measured using the protocol given in S3 Table 2. In this work, we used the single aliquot regenerative (SAR) dose protocol adapted for single grain (Murray and Wintle, 2003;Wintle and Murray, 2006 Wintle, A.G., Murray, A.S., 2006. A review of quartz optically stimulated luminescence characteristics and their relevance in single-aliquot regeneration dating protocols. Radiation Measurements 41, 369 -391. Table 1 Summary of the OSL samples from the Aranbaltza archaeological site used inthis study, listing the sample names, location within the stratigraphic unit, depth of the sample from the surface, assumed life-time average water content (10% error), total environmental dose rate, measured natural dose (D e ), number of accepted data in each sample, Mean Age of the different probability density areas (showed in S3 Figure 1) Table 2. Single-aliquot regenerative-dose protocol (SAR) measurement conditions used in this study. A series of regenerative-dose (L i ) and test dose (T i ) OSL measurements are performed on each individual quartz grain to obtain a sensitivity-corrected dose-response curve on to which the sensitivity-corrected natural (L n /T n ) signal is interpolated to obtain a D e value. In order to check for contamination of the quartz OSL by feldspar grain or inclusions, a repeat dose point is added (step 1b), which includes an infrared (IR) bleach performed for 20 s at 50°C prior to the main L i measurement, as described in Duller (2003).
Step 1a is omitted when measuring a natural signal.
Step 1b is added only when measuring the OSL IR depletion ratio (Duller, 2003).