The authors have declared that no competing interests exist.
Conceived and designed the experiments: MP AA. Performed the experiments: RC HG RJ AP AP RR CS AO. Analyzed the data: PB CC RC ND HG RJ RR AP AP CM CS MV. Wrote the paper: RC ND HG CM MP RR CS.
The Arabian Peninsula is a key region for understanding hominin dispersals and the effect of climate change on prehistoric demography, although little information on these topics is presently available owing to the poor preservation of archaeological sites in this desert environment. Here, we describe the discovery of three stratified and buried archaeological sites in the Nefud Desert, which includes the oldest dated occupation for the region. The stone tool assemblages are identified as a Middle Palaeolithic industry that includes Levallois manufacturing methods and the production of tools on flakes. Hominin occupations correspond with humid periods, particularly Marine Isotope Stages 7 and 5 of the Late Pleistocene. The Middle Palaeolithic occupations were situated along the Jubbah palaeolake-shores, in a grassland setting with some trees. Populations procured different raw materials across the lake region to manufacture stone tools, using the implements to process plants and animals. To reach the Jubbah palaeolake, Middle Palaeolithic populations travelled into the ameliorated Nefud Desert interior, possibly gaining access from multiple directions, either using routes from the north and west (the Levant and the Sinai), the north (the Mesopotamian plains and the Euphrates basin), or the east (the Persian Gulf). The Jubbah stone tool assemblages have their own suite of technological characters, but have types reminiscent of both African Middle Stone Age and Levantine Middle Palaeolithic industries. Comparative inter-regional analysis of core technology indicates morphological similarities with the Levantine Tabun C assemblage, associated with human fossils controversially identified as either Neanderthals or
Though climate change and its effect on people around the world today is receiving considerable attention from scholars, governments and the general public, we have little understanding about how past populations coped with and adjusted to marginal environments in many regions of the world. The vast desert regions of the Sahara and the Arabian Peninsula contain numerous archaeological sites, indicating that Pleistocene hominins penetrated these areas, living in more favourable habitats
With respect to the palaeoenvironments of Arabia, cave speleothem records and lacustrine and sand dune deposits indicate that the region experienced dramatic climatic oscillations between wet and dry periods
Most archaeological sites in the Arabian Peninsula cannot be correlated with past environments as they are known from surface contexts, where chronometric dates and ecological information is not recoverable. Fortunately, several stratified archaeological sites have recently been investigated, making considerable contributions to our knowledge of Middle Palaeolithic occupation history. Middle Palaeolithic industries in Arabia are now known to date to MIS 5e
(A) Shuttle Radar Topography Mission Digital Elevation Model of the Jubbah region, overlain with interpreted palaeolake extents in blue. (B) Landsat ETM false color composite (bands 7, 4 and 1 on red, green and blue) of the Jubbah region, with exposed paleolake sediments appearing as the lightest blue shades, the rock outcrops as black, and the sand dunes light yellow. Both figures show the locations of the Middle Paleolithic sites identified during reconnaissance and discussed within the text.
Much information is yet to be gathered about Arabian palaeoenvironments and the nature of human occupations. In an attempt to help remedy this situation, we report here the presence of three stratified Middle Palaeolithic sites associated with the Jubbah palaeolake, a significant lake basin in the Nefud Desert of northern Saudi Arabia. We describe the ecological settings of the sites and their stone tool assemblages. We examine prehistoric technology and behaviour and discuss the implications of our findings relative to climate change and dispersal models.
A: Large sand dunes (MIS 4?); B: Upper palaesol (MIS 5a); C: Lower palaesol (MIS 5); D: Lower area; E: Base of Jebel Qattar footslope.
Note distribution of artefacts on palaeosol and those on the slope immediately beneath it (in red) and those on lower surface (in blue). Sand dune to the left, Jebel Qattar to the right.
Stratigraphic log showing the locations of the palaeosols, calcretes and the OSL ages. The multiproxy record is also displayed, showing organic content (LOI550), magnetic susceptibility, δ13C and phytolith data. The bar to the left indicates that the OSL ages correspond to MIS 7 (211±16 ka), MIS 5c (95±7 ka), and MIS 5a (75±5 ka). The archaeological assemblage is associated with the upper palaeosol (Unit B, MIS 5a). The inter-stratifed sands between the palaeosols are interpreted as MIS 6, MIS 5d and MIS 5b deposits, and the overlying dune sands as MIS 4.
1–4: retouched tools on flakes; 5: preferential Levallois flake with centripetal preparation; 6: recurrent centripetal Levallois flake. Artefacts from MIS 7 deposit: 7–8: (trench 16): flakes; 9 (trench 3): flake; 10 (trench 6): (preferential?) Levallois flake with faceted platform and centripetal dorsal scar pattern.
Debitage | Cores | Retouched | Total | |
|
95 (81) | 9 (9) | 10 (10) | 114 |
|
28 (100) | 0 | 0 | 28 |
|
1113 (91) | 99 (8) | 10 (1) | 1222 |
|
74 (84) | 10 (11) | 4 (5) | 88 |
Quartzite+ | Quartz | Rhyolite | Other++ | |
|
88 (77) | 13 (12) | 5 (4) | 8 (7) |
|
14 (50) | 5 (18) | 2 (7) | 7 (25) |
|
930 (76) | 277 (22) | 9 (1) | 7 (1) |
|
81(92) |
4 (5) | 0 | 3 (3) |
+this calculation includes small numbers of quartzitic sandstone.
++primarily refers to fine grained siliceous materials (flint-like material).
yellow fine-grained quartzite.
Core type | JQ-1+ | JKF-1++ (quartz) | JKF-1 (quartzite) | JSM-1 |
Total (%) |
Single platform | 1 | 24 | 3 | – | 28 (24) |
Amorphous/other | – | 11 | 7 | – | 18 (15) |
Bidirectional | – | 5 | – | – | 5 (4) |
Radial | 1 | 5 | – | – | 6 (5) |
Discoidal | 2 | – | 2 | – | 4 (3) |
Fragment | – | 4 | 4 | 2 | 10 (9) |
Multiplatform | – | 4 | 3 | – | 7 (6) |
Recurrent centripetal Levallois | 2 | 4 | 1 | – | 7 (6) |
Recurrent unidirectional Levallois | – | – | 4 | – | 4 (3) |
Bidirectional Levallois | – | 2 | – | – | 2 (2) |
Unidirectional convergent Levallois | – | – | 2 | – | 2 (2) |
Preferential Levallois with centripetal preparation | 2 | 2 | 7 | 5 | 16 (14) |
Levallois preform | 1 | – | 4 | 3 | 8 (7) |
Total | 9 | 61 | 37 | 10 | 117 |
+all cores produced from ferruginous quartzite, but for a single platform core from quartz.
++a single rhyolite core is a radial core (not included in tabulation).
all cores produced from the yellow fine-grained quartzite.
Type | JQ-1 | JKF-1 | JSM-1 | Total (%) |
Flake | 76 | 744 | 49 | 869 (68) |
Blade | 3 | 21 | 1 | 25 (2) |
Core management elements | – | 50 | 1 | 51 (4) |
Chips/chunks/fragments | 16 | 298 | 23 | 337 (26) |
Total | 95 | 1113 | 74 | 1282 |
Type | JQ-1 | JKF-1 | JSM-1 |
Side retouched flake | 7 | 3 | 1 |
Side and end retouched flake | 2 | 7 | 1 |
End retouched flake | – | – | 1 |
Retouched point | 1 | – | – |
Bifacial pieces | – | – | 2 |
The Jubbah palaeolake is located in the Nefud Desert of northern Arabia (
The exposed part of the Jubbah palaeolake lies downwind of large sandstone outcrops, particularly Jebel Umm Sanman to the west, which has diverted the westerly flow of sand around the outcrop, leaving a sand-free depression in its lee (
Note the mound to the left. JKF-12 quartzite source is at the base of the jebel.
The location of trench 1 with buried artefacts (in red). Surface finds (in blue) mainly occur between the contour lines of 830–832 m, suggesting the presence of an extensively buried artefact surface. Refitted artefacts occur in buried and surface contexts, indicating a close relationship between surface and subsurface assemblages.
Stratigraphic levels A–J. Stratum H is the exclusive artefact horizon.
The multiproxy record is also displayed, showing organic content (LOI550), carbonate content (LOI950) and magnetic susceptibility.
Ferruginous quartzite artefacts. 1 (no. 504, surface), 2 (no. S-76, stratified in stratum H), 3 (no. 513, surface): Levallois sub-triangular preferential flakes (Levallois “points”) with faceted platform and unidirectional convergent dorsal scar pattern; 4 (no. 921, surface): Levallois sub-triangular preferential flake with faceted platform and unidirectional convergent scar pattern, with some retouch on lateral mesial edge; 5 (no. 502, surface): Levallois sub-triangular preferential flake with dihedral platform and unidirectional convergent scar pattern; 6 (no. 521, surface): flake with faceted platform, produced on a ventral face of a flake.
1 (no. 940, stratified in stratum H): Levallois core for preferential flake with centripetal preparation; 2 (units K9/L9, stratified in stratum H): recurrent centripetal Levallois core; 3 (no. S-20, Stratum H): preferential quadrangular Levallois flake with dihedral platform and unidirectional convergent dorsal scar pattern; 4 (no. 873, surface): Levallois elongated flake (preferential?) with faceted platform; 5 (no. 847, surface): Levallois flake with faceted platform and unidirectional crossed dorsal scar pattern; 6 (no. 209, surface): blade with plain platform and unidirectional convergent dorsal scar pattern.
Artefact No. | Type | Residue | Location | Organic | Anthropogenic | Residue Analysis Results | Contamination |
S-76 | Levallois point | Yes | Lateral edge anddistal tip | Yes | Yes | Abundant evidence of plant (fattyacids, hydrocarbons, sterols) andanimal (fatty acids) | Fatty acids |
504 | Levallois point | Yes | Faceted platformand lateral edge anddistal tip | Yes | Yes | Evidence of plant and animal(fatty acids) | Fatty acids |
S-20 | Levallois preferential flake | Yes | Portion of facetedplatform and lowerlateral edge | Yes | Yes | Abundant evidence of plant - Brassicaceae(fatty acids, hydrocarbons, sterols) | Fatty acids, hydrocarbons |
847 | Levallois flake | Yes | Faceted platform andlower lateral edge withdamage | Yes | Yes | Evidence of plant (Brassicaceae) | Fatty acids |
521 | Flake | Yes | Faceted platform andlateral edge with slightnibbling | Yes | No | No evidence of authenticresidue | Fatty acids |
209 | Blade | Yes | Plain platform and twolower lateral edges | Yes | Indeterminate | Little evidence (Urea - proteinbreakdown or urine) | Fatty acids |
873 | Levallois preferential flake | Yes | Faceted platform andtwo lower lateral edges | Yes | Yes | Evidence of animal (fatty acids) | Fatty acids |
The presence of Middle Palaeolithic hominins in northern Arabia is demonstrated by the occurrence of numerous surface archaeological localities
Note that in situ artefacts occur in the middle of the mound. Artefacts occur on a palaeodune surface, situated above a lower-lying shallow lake.
The seam is at the base of the jebel hillslope. The Jubbah palaeolake bed is in the background (where trees are located). Dune surround the lakebed (on right of photograph, view southeast). The JSM-1 site is off of the photograph, to the right.
The Jebel Umm Sanman is the largest jebel in the area. The modern town of Jubbah is developing on the former lake bed. Note the location of yellow quartzite seam relative to artefact finds.
Yellow fine-grained quartzite artefacts. 1 (Trench 1, spit 2): Levallois flake from centripetal recurrent debitage, with dihedral platform; 2 (no. 81, surface), 3 (no. 1, surface), 4 (no. 84, surface): Levallois core for preferential flake removal, with centripetal preparation; 5 (no. 82, surface): biface on a flat tabular slab.
Key wadis are numbered and international borders displayed by dashed lines. Drainage network data (blue) is superimposed upon SRTM V.4 elevation data, overlain upon Natural Earth 2 data for the oceanic regions. Interpreted channel connections potentially active during recent wet phases are marked in grey. A: Simplified principal drainage networks for the Arabian peninsula. B: Detailed palaeodrainage networks for the Nefud region around the Jubbah sites, with palaeolakes identified through remote sensing displayed in light blue. Numbered widyan/rivers: 1- Wadi as Sirhan, 2-Wadi al Hamd, 3- Euphrates and widyan draining into the Euphrates basin, 4-Wadi al Batin, 5-Wadi Sabha, 6- Wadi ad Dawasir, 7- Wadi Hadramawt.
Note that the Jubbah artefacts are most similar to the Levantine Tabun C assemblage.
Attribute | Function 1 | Function 2 |
Proportion of Blades | 0.399 | 0.625 |
Average Platform Angle | 0.301 | −0.159 |
Relative thickness | 0.3 | 0.039 |
Scar Pattern Index | −0.276 | −0.408 |
Intersection Height | 0.231 | −0.07 |
Elongation | 0.066 | 0.319 |
Contracting/Expanding | 0.06 | 0.088 |
Distal Curvature | 0.024 | −0.003 |
Largest Scar Proportion | 0.006 | −0.102 |
Jebel Qattar 1 (JQ-1) (28°00′53.20 N, 41°03′35.94 E) is located just south of a sandstone jebel and close to the southern edge of the exposed palaeolake basin (
Loss on ignition organic content (LOI550) and carbonate content (LOI950) were conducted following the standard procedures
The sedimentary sequence at JQ-1 is comprised of a stratified sequence (
The MIS 7 palaeosol is overlain by aeolian sands and an interstratified calcrete/palaeosol deposit which, due to its stratigraphic position, we attribute to MIS 5e. The occurrence of dramatically increased humidity across Arabia during this period is well documented within a variety of palaeoclimatic records
The MIS 5c paleosol (Unit D–95±7 ka) consists of poorly sorted silts and medium-grained sands, with calcareous rhizolith casts and phytoliths which indicate C3 Pooid-dominated grassland with some tree cover and C4 Panicoid, but little C4 Chloridoid grass cover; this interpretation is corroborated by δ13C values of 23%. The uppermost palaeosol (Unit B) was formed during MIS 5a (75±5 ka) with phytolith evidence indicating that at this time, the landscape was dominated by a mix of C3 Pooid and C4 Panicoid and Chloridoid grassland types, with some trees. A shift to drier conditions during MIS 5a is indicated by a slight increase in C4 Chloridoid grasses coupled with a 2% shift in δ13C values, however, the vegetation is never dominated by C4 flora. The MIS 5a palaeosol is overlain by massive sands, representing a major change in the depositional regime, which we interpret as evidence for climatic desiccation at the start of the last glacial period (MIS 4), which would be consistent with other dated records for the region
In arid/semi-arid regions such as the Jubbah basin, increased wetness during humid phases such as MIS 5e, coupled with increased vegetation cover and landscape stability, was crucial in promoting palaeosol and calcrete formation at JQ-1. The topographic position of the MIS 5e calcretes discussed here, cannot preclude either groundwater or palustrine influence as an agent for their formation, however, their presence is an important indicator for a substantial increase in regional humidity at this time. The evidence from JQ-1 for a significantly wetter climate during MIS 5e is consistent with other palaeoclimatic records from Arabia at this time, and indicates that lake formation within the Jubbah basin likely achieved its maximum extent during MIS 5e and covered a smaller area during other pluvials such as MIS 7, MIS 5a and c. These periods, however, would have been substantially more humid than today, as evidenced by extensive palaeosol formation. It is suggested that these palaeosols would have formed along lake margins, providing an attractive setting for Middle Palaeolithic populations. Additionally, tufa deposits 300 m east of JQ-1, at the base of the jebel, preserve abundant rhizoliths further indicating that freshwater sources were widely available during humid phases.
A total of 518 lithic artefacts were recovered in association with the raised MIS 5a palaeosol and the lower-lying surface extending northwards (
The 114 artefacts recovered in association with the MIS 5a palaeosol consisted of 95 pieces of debitage (81%), 9 cores (9%) and 10 retouched tools (10%) (
A total of 28 pieces of debitage was recovered from the MIS 7 upper paleosol (
Around the base of Jebel Qattar, seams of ferruginous quartzite are found in the sandstone, and quartzite also occurs as weathered clasts at the base of the Jebel’s slope. The weathered clasts are small and hard, with many natural inclusions. The generally small size of the JQ-1 cores therefore relates to the size and quality of this raw material. Within the sandstone are small quartz pebbles, which weather to form common clasts in spatially extensive gravel sheets in the lower area. Given the local occurrence of quartzite and quartz at JQ-1, the site to source distance for the transport is perhaps no greater than 100 m. On the other hand, the recovery of small numbers of rhyolite and flint artefacts implies long-distance transport, since no volcanic and siliceous sources have yet been identified at Jubbah. It is also possible that these materials were procured locally from unknown stream sources that included secondary materials.
Jebel Katefeh 1 (JKF-1) (27°55′18.64 N, 40°48′18.14 E) is situated in a basin ca. 800 m east of the Jebel. A remnant mesa of relict aeolian sediments overlain by lacustrine silts, and capped by a calcrete deposit, measures ca. 40 m in diameter, rising to a height of 3 m above the current ground surface (
A trench measuring 2 m in width and 12 m in length was excavated, beginning at the top of the mound and extending downslope to the west. Nine stratigraphic layers were differentiated in the trench, labeled as A–I (
The artefact-bearing unit is overlain by Units G and E, which display a fining up sequence from cemented silt-sand sediments, into lacustrine material. Increases in magnetic values throughout the units correspond to grain size changes, while organic content (LOI550) also increases, achieving substantially high values at ca. 30 cm. Similarly, carbonate content (LOI950) values are increased throughout Unit E, reflecting the precipitation of lake carbonate material. Unit D reflects an influx of aeolian sand and a return to arid conditions, with low organic content and carbonate values. This is overlain by Units C4 and C1, which represent the formation of calcrete and reflects a substantial increase in humidity within the region. Analysis of the grain size characteristics from these calcrete units reveal the sediments to be very poorly sorted silts with a near symmetrical skewness. These are overlain by aeolian sediment (Unit B), which reflects the onset of arid conditions within the region.
All excavated sediment was sieved through a 6 mm2 mesh. A total of 300 artefacts were recovered from Unit H, most of which were small (<2 cm). A total of 30 larger artefacts were piece-plotted in place. Unit H corresponds with the point at which artefacts are concentrated on the sloping mound surface, indicating that the surface artefacts derive from the deposit, as confirmed by refits.
We used OSL dating to obtain burial ages for sand-sized grains of quartz. Three samples were taken, one from Stratum I (JKF-OSL-4), the lowermost excavated level, and two from Stratum H (JKF-OSL1, JKF-OSL-2), the artefact-bearing level. In each of the three samples, only between 1–2% of the grains were suitable for equivalent dose (De) determination. The single-grain De values were widely spread, with overdispersion values ranging between 35±5% and 62±8% (see
The 1222 artefacts from Unit H and the surface are technologically homogeneous and are treated here as a single assemblage. The assemblage consists of 1113 pieces of debitage (91%), 99 cores (8%) and 10 lightly retouched tools (1%) (
The cores made from quartz (n = 61) and quartzite (n = 37) vary typologically, likely owing to raw material type and clast size (
The majority of the debitage are flakes (n = 744), with the recovery of only small numbers of blades (
Given the relatively good preservation conditions at JKF-1, a feasibility study of artefact residues was conducted (Supporting Information S3). The seven selected artefacts had generally fresh edges and included 6 quartzite and 1 quartz artefact. The selected types and forms (2 Levallois points, 2 preferential Levallois flakes, 1 Levallois flake, 1 blade, 1 flake) were considered to be potentially good candidates for use. Visual inspection and low powered microscopy was first employed to screen the artefacts for possible residues. Screening was followed by the removal of residues from particular edges of artefacts. The removed residues were subsequently assessed using absorbance spectroscopy and gas chromatography coupled mass spectrometry (GC/MS) at Lakehead University (Canada). Five of the seven artefacts produced positive results for compounds consistent with utilization of plants and animals (
Most of the artefacts from JKF-1 are ferruginous quartzite, originating from an outcrop (JKF-12) (27°55′15.1 N, 40°48′05.9 E) at the base of the jebel which is ∼800 m to the west. At JKF-12, the quartzite occurs as spatially extensive beds of relatively fine-grained material available as large clasts. The quartzite was reduced at the base of the jebel, which served as a location for raw material procurement based on the identification of large Levallois cores measuring up to 25×20 cm (thereby contrasting with the typically much smaller cores at JKF-1). Within the sandstone are naturally occurring quartz cobbles and pebbles, which are also widely distributed across the basin floor. The quartz occurs within a few hundred meters of the site, while site-to-source distance for the transport of ferruginous quartzite is no greater than 800 m. The rare rhyolite artefacts suggest long distance transport.
In sum, we interpret the JKF-1 site to be a relatively short-term occupation given the presence of lithic assemblages from a single and thin deposit, the homogeneity of the lithic assemblage, and the presence of refitted cores and flakes. Based upon mapping and stratigraphic analyses, occupations appear to be on an aeolian dune, overlooking a low-lying area that may have had standing water (
Jebel Umm Sanman is the largest jebel in the area, measuring 7 km north-south and 3 km east-west, and reaching a height of 1264 m. Jebel Umm Sanman 1 (JSM-1) (27°58′34.83 N, 40°55′28.28 E) is located in the southeastern margin of the jebel, overlooking the western margin of the northern palaeolake (
Low densities of artefacts were found over a surface measuring ca. 160×75 m, and across a slight slope going west to east (
Two OSL samples were taken from the artefact-bearing levels, one from Stratum B (JSM1-OSL1), 20 cm below surface, and one from Stratum C (JSM1-OSL2), 42 cm below surface. In each of the two samples, only ∼1% of the grains were suitable for De determination. As at JKF-1, the single-grain De values were widely spread, with overdispersion values of 93±11% and 67±10% (see
The 11 surface and 77 buried artefacts were technologically homogeneous. The 88 artefacts included 74 pieces of debitage (84%), 10 cores (11%) and 4 retouched pieces (5%) (
We have discovered and investigated three stratified and buried lithic assemblages in the Nefud Desert, thereby nearly doubling the count of well preserved Middle Palaeolithic sites in the Arabian Peninsula. The geographic location of the Jubbah sites, in the northern interior region of Arabia, adds important new information about the distribution of Middle Palaeolithic hominins, as the other published site excavations are located in the extreme southern parts of the peninsula. Out of Africa dispersal models have mostly emphasized the Bab al Mandab route on the basis of genetic data
The environmental record from JQ-1 and JKF-1 reveals the presence of a series of palaeosol, calcrete and lacustrine deposits that formed under a substantially more humid and ameliorated climate. Multiproxy, sedimentological and stratigraphic evidence indicates that lakes within the Jubbah and Katefeh basins fluctuated in size, in response to changing climatic conditions between MIS 7 and MIS 4. During peak pluvials such as MIS 5e, lakes within these basins may have covered an area of up to 78 km2. Although lake volumes may have been reduced during less humid conditions, soil formation occurred along lake margins while the surrounding landscape remained dominated by grassland, with some trees. This is an ideal setting for migrating animals, which in turn, would have been attractive to mobile hunter-gatherers. We note that this humid period migratory scenario contrasts with genetic data suggesting dispersals at the outset of an arid phase, or MIS 4 (between 70–60 ka)
Having reached Jubbah, Middle Palaeolithic hominins would have resided on the margins of the palaeolake, which would have provided an attractive habitat with key resources. Sites were situated on the western and southern parts of the northern lake (JSM-1, JQ-1) and on the eastern fringes of the southern lake (JKF-1). In the case of JKF-1, geomorphological reconstruction reveals that the occupation occurred on an aeolian sand dune, overlooking the lower-lying lake (
Though so far a small excavated stone tool assemblage, the recovery of 28 artefacts in a deposit dated to 211±16 ka represents the oldest reliably dated occurrence in the Arabian Peninsula. We tentatively associate this assemblage with the Middle Palaeolithic on the basis of the age of the technology and the recovery of two Levallois flakes. Although we cannot be certain of the species that manufactured the artefacts, we note that the lithic assemblages were produced at a time corresponding with the origin of
An entirely different scenario presents itself 130,000 years later, as the date of 75±5 at JQ-1 generally corresponds with the occupation of the Levant by
The Jubbah sites demonstrate stone tool reduction methods that are characteristic of Middle Palaeolithic industries. The lithic assemblages contain a variety of reduction methods, including centripetal, bidirectional, unidirectional and convergent techniques. The JQ-1 and JSM-1 assemblages are primarily centripetal in character, with centripetal preparation of preferential Levallois cores at the latter. At JKF-1, it appears that the earlier phase of reduction of quartzite was primarily unidirectional and unidirectional convergent, with a move towards centripetal reduction as size decreased. At JKF-1, quartz was usually reduced in a simple flaking manner, but some examples are rather more sophisticated and are Levallois-like. The method of manufacture of the Jubbah industries differs from the more distinctive techniques described in the extreme southern zones of Arabia, including the Nubian technocomplex in Oman
To situate the Jubbah technology in a wider geographic context, 55 cores from JQ-1 (upper assemblage) and JKF-1 were compared to those from neighbouring regions. The comparative sample included Middle Stone Age sites from the Horn of Africa; the Levantine Mousterian sites of Tabun Cave, Layer C, and El Wad; the early
In conclusion, this article presents the first systematic research on archaeological sites in the Nefud Desert of Saudi Arabia. Our analysis indicates that Middle Palaeolithic hominins were able to penetrate deep into the interior of northern Arabia during ameliorated humid periods. Middle Palaeolithic archaeological evidence for terrestrial and inland movements contrasts with, or at least complicates, the coastal model for rapid human expansion
All necessary permits for the Jubbah fieldwork and analyses were obtained from the Saudi Commission for Tourism and Antiquities, Kingdom of Saudi Arabia.
Mapping and palaeohydrological methods report.
(DOCX)
Supporting information for OSL dating.
(DOC)
Residue analysis report.
(PDF)
Core analysis report.
(DOCX)
We thank HRH Prince Sultan bin Salman, President of the General Commission for Tourism and Antiquities, and Ali Ghabban, Vice President for Antiquities and Museums, for permission to carry out this study. We also thank Hussain Abu AlHassan and Habeeb Turki and the people of Jubbah for their support and assistance with the field investigations. We thank Gauthier Devilder for lithic illustrations and Yasaman Jafari for OSL sample preparation, and Zenobia Jacobs for OSL data analysis. We also thank the staff of the Lakehead University Instrumentation Laboratory, particularly Grzegorz Kepka and Ain Raitsakas, for their assistance.