R.J. Birtwistle & G. Yeritsyan

5

LATE MIDDLE PALAEOLITHIC NEANDERTHAL NETWORKS AND ASSEMBLAGE VARIABILITY IN ARMENIA: LITHIC EVIDENCE FROM LUSARKET I ROCKSHELTER

R.J. Birtwistle1 and B.G. Yeritsyan2

ABSTRACT

Until recently the region presently encompassed by the Republic of Armenia has been overlooked by researchers investigating Middle Palaeolithic Neanderthal activity in favour of surrounding areas. Exploration over the last decade has identified a number of new sites bolstering the previously limited data. In support of the new fieldwork re-analysis of previously studied Palaeolithic artefact assemblages is aiding development of the Middle Palaeolithic record within Armenia and wider Southern Caucasus. Recent reassessment of a stone tool industry recovered in the mid-1970s from Unit D at Lusarket I rockshelter in Central Armenia has examined a Middle Palaeolithic industry with a complex taphonomic history. Although there remains a question mark over the contextual validity of the material, typological analysis reveals affinities with Mousterian assemblages from the Tsutskhvati Cave System in Georgia, reinforcing notions that late Middle Palaeolithic Neanderthals were itinerant. In addition to the new work on Lusarket I the authors comment on the available artefactual evidence from other Mousterian assemblages in the Caucasus while highlighting typological relationships between late Middle Palaeolithic sites in Armenia and those in neighbouring regions.

Full reference: Birtwistle, R.J. and Yeritsyan, B.G. 2012. Late Middle Palaeolithic Neanderthal networks and assemblage variability in Armenia: lithic evidence from Lusarket I rockshelter. Lithics: the Journal of the Lithic Studies Society 33: 5–16.

Keywords: Late Middle Palaeolithic, Southern Caucasian Neanderthals, Lusarket, non-Levallois denticulated Mousterian, truncated facetted technique, retouch patterns, stone tool variability.

INTRODUCTION

Middle Palaeolithic models of Neanderthal movement patterns and assemblage variability in the Southern Caucasus highlight extensive cultural networks encompassing neighbouring regions such as the Zagros/Taurus and Northern Caucasus mountain ranges (Beliaeva & Liubin 1998; Golovanova & Doronichev 2003). Until recently the Armenian Middle Palaeolithic had been overlooked, and contributed little towards the chronological framework of the Middle Palaeolithic in these areas. Early reports identified a mixture of Levallois and non-Levallois industries, produced on a variety of raw materials, displaying variations in their typology (Baiburtian 1937 & 1938; Dzhafarov 1983; Formozov 1977; Liubin 1977 & 1984; Mansurov 1990; Yeritsyan 1970a; 1970b & 1975; Yeritsyan & Ghazarian 1978). However the typological variability has yet to be fully understood in terms of the broader context of Neanderthal mobility across the region.

The Middle Palaeolithic record in Armenia is limited to a handful of sites and find spots (Figure 1), including; Yerevan Cave

(Yeritsyan 1970a & 1970b), Lusarket I and II (Yeritsyan 1975; Yeritsyan & Ghazarian 1978; Fourloubey et al. 2003), Hovk in Northern Armenia (Pinhasi et al. 2006 & 2008), and a series of open air locales on the periphery of ancient paleo-lakes in the Aparan region, identified during survey work in 2003 (Gasparyan et al. 2004; Jaubert & Ollivier 2003). Recent work has provided stimulus for long-term research projects in the area to investigate behaviour and mobility patterns among late Middle Palaeolithic (LMP) Neanderthal communities. Through the continuing discovery of new sites (Liagre et al. 2007) and reappraisal of sediments and lithics from previously known archaeological sequences (Birtwistle 2008; Wilkinson 2008 & 2009), the Middle Palaeolithic record in Armenia is expanding.

This paper reports on lithic assemblages recovered from excavations of Unit D at Lusarket I in the Hrazdan gorge, Central Armenia, between 1970 and 1981. Poor storage of archive material from these excavations rendered the extant archive difficult to interpret. Re-excavation of the site was therefore initiated during the 2008 and

1 20 Aitchison Avenue, Hucknall, Nottingham, NG15 6LP, England, U.K., r.birtwistle@hotmail.com 2 Institute of Archaeology and Ethnology, Yerevan, Armenia

Lithics 33

6

2009 field seasons. New stratigraphic analysis reveals that the archaeology has been profoundly affected by a number of taphonomic processes, compromising its analytical value. Following a summary of the Middle Palaeolithic record of Armenia, we analyse the assemblage from Unit D, making typological comparisons with dated examples in Georgia. On this basis we advance some inferences concerning the cultural networks of local Neanderthal groups. We conclude with an overview exploring the range of variability between Mousterian assemblages within Armenia and those situated across the Southern Caucasus and beyond, and in particular, identifying notable typological relationships.

THE MIDDLE PALAEOLITHIC RECORD IN ARMENIA

Armenia’s landscape presents a rich archaeological environment that has been exploited by hominin species for the last 1.5my (Gabunia et al. 2000). In some locations stone artefacts are observed eroding out of ancient deposits, and although surface finds have been frequently recorded since the 1930s

(Baiburtian 1937 & 1938), no comprehensive examination of the archaeological sediments has yet been undertaken. Regional Palaeolithic sequences thus remain poorly understood.

The earliest Middle Palaeolithic activity probably occurred c.200kya at Metsavan (Figure 1) where Acheulean and Levallois industries co-existed for some time (Dolukhanov et al. 2004; Golovanova & Doronichev 2003). This pattern is recurrent at Jraber (Figure 1), where both bifaces and Levallois levels are interstratified (Fourloubey et al. 2003; Dolukhanov et al. 2004). These examples would appear to represent the earliest Middle Palaeolithic sites in Armenia. The Mousterian evidence begins at the Pechka rockshelter (Figure 1) where an exhausted core, two blades, a single flake, four points, a side and end scraper made on a combination of andesite and dacite blanks reflect hominins working a range of raw materials. Furthermore, at Pagakhpyur and Atkalich, cores producing Mousterian flakes and flaked tools were collected from the banks of the Atka Lake, located west of the Khurda-Jalal Mountain (Dolukhanov et al. 2004).

Figure 1. Middle Palaeolithic sites of Armenia.

R.J. Birtwistle & G. Yeritsyan

7

In northwest Armenia investigations around a series of cave, open-air and rockshelter sites represent the first modern archaeological excavations of stratified Middle-Upper Palaeolithic deposits in the country (Pinhasi et al. 2008). Situated in a forested area of the Lesser Caucasus mountain range, the Hovk sites (Figure 1) contain chronological evidence spanning the Middle, and possibly Upper, Palaeolithic (Hovk 1, Unit 5; Hovk 2 and 3). The sequence at the Hovk 1 cave consists of 11 archaeological horizons (Pinhasi et al. 2008); Unit 8 interestingly produced several Levallois points and blades, while similar elongated forms were recovered from Unit 7 (Pinhasi et al. 2006 & 2008). Pinhasi et al. postulate that the laminar material from Units 7 and 8 corresponds with the Levantine Mousterian from Tabun level D, which features numerous elongated Levallois points and blades. South of Hovk, the site of Kalavan 2 (Figure 1) (currently being examined by Christine Chataigner) has recently been dated to c.34,200 ± 360 14C BP (POZ-20366) or 39,643 ± 886 cal BPHuLu based on Accelerator Mass Spectrometry (AMS) 14C analysis (Pinhasi et al. 2008) and has yielded hundreds of tools including typical Mousterian

points, side scrapers, borers and burins produced by radial Levallois techniques on obsidian, flint and local limestone blanks.

Following surveys at Angeghakot (Figure 1) in the valley of Vorotan in 2003, numerous obsidian “Yerevan” points were identified displaying presence of the cultural phenomenon first recognised at Yerevan Cave (Figure 1) by Yeritsyan (1970a) as the truncated-facetted technique (Liagre et al. 2007). The Mousterian assemblage recovered from Yerevan Cave in the Hrazdan Valley is characterised by Levallois flakes, side scrapers and retouched “Yerevan” points (Golovanova & Doronichev 2003). Liubin (1984 & 1989) originally assigned the lower levels of Yerevan Cave to Würm II and the upper levels to the Middle Würm based on the faunal record. However since then layers 3, 4 and 7 have been repeatedly dated using AMS 14C analysis based on various charcoal and bone fragments providing a range of dates from 49–27 ka 14C BP (Pinhasi et al. 2008; Table 3). If these dates prove robust, these deposits would tentatively represent some of the latest Neanderthal phases in Europe.

Figure 2. Archaeologists working at Lusarket I during summer 2008 with Lusarket II rockshelter on the right.

Lithics 33

8

Lusarket 1

The Caucasus Mountains were formed by extensive folding of the pre-existing layered sedimentary rocks sometime between 28–23mya (Mikhailov et al. 1999). With Armenia geographically located within an active and volatile tectonic zone many varieties of igneous, as well as sedimentary rocks (namely basalt, flint and limestone) were widely exploited (Chataigner et al. 2003). Lusarket I is a basalt rockshelter (Figure 2) positioned along the Hrazdan valley approximately 25km north of Yerevan, at an altitude of 1433m above sea level. From its origin in Lake Sevan, the valley runs along a north-east to south-west course and is predominantly layered with basalts and a number of obsidian outcrops, and contains the sites of Lusarket I and II and Yerevan Cave. The post-excavation problems at Lusarket I prior to 2000 limited detailed analysis beyond that of basic description. As a result, the sample is unlikely to accurately reflect activity at the site. No micro-debitage or knapping debris has survived, suggesting that sieving was not undertaken, and since excavation, the unit D material has been stored in small containers. This suggests that larger pieces, such as cores, were simply ignored. No records exist documenting where the artefacts

were recovered from within the stratigraphy, and thus the contextual validity of the sample is questionable.

Recent investigations has established the full stratigraphic sequence, and reappraised the deposits formerly described as Units A to H (Figure 3) (Fourloubey et al. 2003). The sequence has been grouped according to mode of formation into two major stratigraphic units (Table 1); an upper stratum comprised of Units A and B and a lower stratum represented by Units C, D1 and D2 (Wilkinson pers. comm. 2009). Additional work inside the cave has identified a slightly different set of deposits which contain in situ lithics, fauna, macrobotanical remains and even evidence of hearth activity (Adler pers. comm. 2011), whilst the sediments outside the cave have experienced a completely different taphonomic history. Taphonomic and environmental analyses suggests that Unit D accumulated in a floodplain setting, resulting in the mixed condition of the assemblage (rolled and fresh artefacts co-occur). The majority of the assemblage exhibits damage resulting from movement (Adler pers. comm. 2011). The rearrangement of the material clearly impacts upon its integrity and limits the conclusions that can be drawn from it.

Figure 3. Section through Lusarket I following Yeritsyan’s early excavations (Fourloubey et al. 2003).

R.J. Birtwistle & G. Yeritsyan

9

Stratigraphy Unit Matrix Assemblage Dating

UPPER

STRATUM

A

Dark greyish brown clay/silt containing frequent herbaceous

plant roots and limestone boulders. Fine sub-angular basalt clasts. Affected by

pedogenesis.

Floral remains

B

Poorly sorted yellowish brown silty clay containing randomly disturbed moderate sub-angular

basalt clasts

Upper Palaeolithic style assemblage containing large

numbers of denticulates, scrapers and burins

UNCONFORMITY

LOWER

STRATUM

C

Greater quantity of basalt gravel in finer matrix.

Yellowish brown clay. Gravels poorly sorted. Top of Unit modified by pedogenesis

Lower industry = non-Levallois denticulates, points

and backed knifes. Upper industry = facetted Levallois.

50% scrapers, 25% points

26,920 ± 220 14C

BP

D1

(Same as C) With rare sub-angular basalt clasts. Carbonate

probably a post-depositional feature

Denticulated Mousterian produced by facetted non-

Levallois techniques

D2

Light yellowish brown fine sandy silt, containing frequent sub-angular basalt pebbles and

fine cobbles. Granular-sized carbonate present throughout

the stratum

Table 1. Revised stratigraphy and lithic data from Lusarket I.

The latest Middle Palaeolithic deposits are found in the upper part of Unit C (formerly C1) where Yeritsyan (1975) identified a laminar Levallois blade industry which displayed the truncated facetted technique; including numerous Levallois “Yerevan” points (Figure 4), side-scrapers and backed knifes. The lower level of Unit C (formerly C2) displays none of these characteristics; the assemblage is rich in denticulates (notches and saws), points and backed knifes produced by non-Levallois techniques (Fourloubey et al. 2003). The Unit D material was recovered from beneath this unit, and was reported to show the same characteristics, supposedly containing high numbers of denticulates and points (Yeritsyan 1975; Fourloubey et al. 2003). The reliability of the current dating evidence is suspect as bone samples from the

lower part of Unit C (formerly C2) were tested by AMS, producing a very late date for suspected Neanderthal occupation of 26,920 ± 220 14C BP (Fourloubey et al. 2003). Moreover, seasonal silting would have quickly and repeatedly reworked deposits, including the bone samples contained therein.

THE UNIT D MOUSTERIAN FROM LUSARKET I

Recent analysis of the limited Unit D material recovered in the mid-1970s has identified an obsidian assemblage totalling 2132 pieces collected between depths of 130–216cm (Unit D1 and D2) (Birtwistle 2008). The sample selected for study is thought to only represent a fraction of the recovered material, as the survival of the entire assemblage has been heavily affected during post-excavation.

Lithics 33

10

Flaked Tool D1 D2 Unit D D%

Notch 12 187 199 54.4

Scraper 6 85 91 24.9

Denticulate (saw) 0 7 7 1.9

Backed Knife 4 20 24 6.6

Burin 1 1 2 0.5

Pseudo-Lev. Point 0 4 4 1.1

Multiple Tool 1 6 7 1.9

Piercer 0 3 3 0.8

Non-Levallois Blade 2 10 12 3.3

Levallois Flake 0 4 4 1.1

Levallois Point 0 2 2 0.5

Levallois Blade 0 11 11 3.0

Total 26 340 366 100

Table 2. Flaked tool and percentage frequencies recorded from the Unit D excavations between 1970–1981

Therefore the sample presented in this study is all that remains from the original excavations (Birtwistle 2008). Consequently only 366 flaked tools exist within the sample, although the fresh and rolled appearance of the material supports the sedimentary evidence that Unit D formed under the changeable conditions of a floodplain environment.

Table 2 provides a summary of the findings indicating higher frequencies of notched and scraper-like pieces than any other tool. The occurrence of 24 backed knives (probably fashioned on rounded obsidian blanks retrieved from the banks of the Hrazdan River) challenges previous observations by Fourloubey et al. (2003) who noted their absence. Broken flakes and debitage account for 61% of the material; however, only limited evidence of reduction can be read from the small, shapeless and exhausted cores. The dorsal scar patterns among the flake and flaked tools indicate non-Levallois discoidal and conical forms of flaking, (see Birtwistle 2008), as flakes appear to have been removed from cores by unipolar parallel techniques from a fixed platform. The presence of pseudo-Levallois points would support this. An examination of the flake butts identifies preparation of the striking platform as 35% of the flake and flaked tool butts display

facetting. In addition two radial-Levallois cores and a small selection of Levallois preparation flakes identified through refitting, suggest a limited presence of Levallois technology which produced at least 17 Levallois pieces (Table 2). Interestingly, one partially retouched Levallois point presents fascinating evidence for impact damage presumably sustained during hunting. Finally large flakes are under-represented: only 3.1% of the flakes exceeded 100mm in length, whereas 51.4% ranged between 50–99mm and 45.5% less than 49mm long (Birtwistle 2008).

Scraper n %

Straight Single Side 41 45.0

Double Side 5 5.5

Concave 14 15.4

Convex 22 24.2

Convergent 1 1.1

Core 6 6.6

Déjeté 2 2.2

Total 91 100

Table 3. Typology and frequency of scrapers identified in Unit D assemblage.

R.J. Birtwistle & G. Yeritsyan

11

The sample (Table 2) contains numerous “processing implements” (e.g. notches and scrapers) rather than elements related to hunting. The notches in particular vary in size and may have been used in a variety of ways. This variation in size could relate to function, level of curation, or the quality of available obsidian; in the absence of microwear it is difficult to distinguish between these possibilities. A lack of heavy duty chopping tools and the recovery of only two small bifaces was noted by Yeritsyan (1975) and Yeritsyan & Ghazarian (1978) who labelled the material as Mousterian of Acheulean Tradition (MTA). However, Unit D, although a heavily selected sample, appears to be characterised entirely by flaked tools. The low number of pseudo-Levallois points and true Levallois points could reflect behavioural differences between the Unit D hominins and those associated with Yerevan Cave, Angeghakot and later periods at Lusarket I although such conclusions cannot currently be sustained, as the assemblage is mixed and incomplete.

The scraper assemblage is dominated by single side-scrapers, together with basic convex and concave scrapers (Table 3). Interestingly none of the 22 convex scrapers were transversely retouched, and, with exception of both déjeté examples and a single convergent piece, most scrapers were only lightly and irregularly retouched. The retouched implements are grouped together based on four typo-technological observations: the morphology, extent, distribution and position of retouch (Birtwistle 2008). The retouch patterns for scrapers falls within the range apparent for the retouched assemblage as a whole (Birtwistle 2008). Most retouch is irregular; 85.6% of pieces were lightly retouched, although the distribution of retouch varied throughout the assemblage. This variability probably relates to function. Most artefacts were retouched on the dorsal face, with other positions (inverse, alternative, bifacial and crossed) accounting for the remaining 33.3% (Birtwistle 2008). On the whole, the finely retouched examples common across Middle Palaeolithic Europe are lacking, perhaps representing a cultural facies or the expedient use of technology (Binford 1973; Ebert 1979).

DISCUSSION

Re-investigation of sedimentary and lithic data has refined the previous work by Fourloubey et al (2003) but is disappointing, as the material seems contextually insecure. Irrespective of this, the assemblage still has merit; certain elements can still contribute toward the developing Middle Palaeolithic record of Armenia; although full comprehension of hominin behaviour at the site during this period will never be forthcoming.

Both the Levallois and non-Levallois elements of the assemblage from the overlying archaeological deposit (Unit C) differs typologically from Unit D; Unit C also produced greater quantities of material (Gasparyan pers. comm. 2008). The Unit D industry is comprised entirely of obsidian, in contrast to Unit C where a variety of raw materials were exploited, behaviour that is recognised at other Middle Palaeolithic sites in Armenia. This implies a shift in the economic preference of the hominins towards collecting and transporting materials from a variety of sources. Binford (1980) has suggested that typological differences between deposits occur when stable conditions are disturbed by environmental and biological factors. The Unit D material appears to remain typologically consistent, which might indicate relatively stable conditions preceding the accumulation of Unit C.

It has been suggested that in the Caucasus higher elevations were occupied during the summer months while populations migrated to lower altitudes during colder periods (Adler et al. 2006). At Lusarket resting 1433m above sea level, occupation was probably determined by annual temperature fluctuations and resource availability. Studies conducted elsewhere emphasise the importance of hunting as the main source of meat procurement for Middle Palaeolithic hunters (Bocherens 2011; Speth 2010); occupation may thus have been dependent on migratory behaviour of animals like the native Caucasian Tur (Capra caucasica) (Adler et al. 2006, 2008). However recent work has highlighted the potential versatility of Neanderthal subsistence (Hardy & Moncel 2011). The Levallois point displaying impact damage

Lithics 33

12

recovered from Unit D2 could have been the result of hunting; although Shea (1988a & 1988b) suggests that not all points may have had the same function.

The Lusarket material appears to share techno-typological similarities with the Tsutskhvati-type Mousterian found at the Bronze and Tsutskhvati cave systems in Georgia. Therefore we can infer that Neanderthals were highly mobile across this part of the region during the late Middle Palaeolithic. Both sites located around 12km north of Kutaisi in Western Georgia; typify classic non-Levallois flaking methods producing large quantities of side-scrapers and notches. Moreover retouch is only lightly applied as pieces exclusively display irregular and stepped morphologies with short applications (Tushabramishvili 1978). These similarities extend to the two déjeté scrapers which are alike to examples from Il’Skaya 1 and Mezmaiskaya Cave in Russia (Stepanchuk 1998) and many European specimens.

Late Middle Palaeolithic assemblages in Armenia exhibit a range of flaked industries produced by distinctive Levallois and non-Levallois techniques. Typological studies underline cultural relationships linking the material from these sites with examples from the Levant, Zagros and wider Caucasus. For instance, Liagre et al. (2007) recognise that the

Mousterian found at Gazma and Zar in Karabakh (Mansurov 1990), Lusarket II and in Unit C at Lusarket I are typologically comparable to the Zagros-Taurus Mousterian. Similarly, at Taglar Cave, Karabakh, the truncated facetted technique has been recognised on various Levallois pieces, and a link suggested between this material and the use of the technique in Armenia (Dzhafarov 1983).

Techno-typological assessment of lithic data from the Georgian Republic has identified similarities between numerous assemblages (Tushabramishvili et al. 1999), such as the Kudaro group in South Ossetia which includes Kudaro (1 and 3) and the Tsona caves (Golovanova & Doronichev 2003), and the cave at Djruchula (Meignen & Tushabramishvili 2006). The assemblages all, with the exception of Kudaro 1, consist of numerous laminar flakes and elongated points, a technology which is considered to be comparable with material from Hovk 1 (Units 7 and 8), Taglar Cave, Dashsalakhli Cave in Western Azerbaijan and the Tabun D Levantine assemblages (Golovanova & Doronichev 2003). Deposits in the Tskhinvali group (located along the Bolshaya River valley) display Levallois technology produced on local andesite with high quantities of blades, denticulates and side-scrapers similar to examples from the Tabun B Levantine Mousterian (Beliaeva & Liubin 1998).

Figure 4. Middle Palaeolithic artefacts from Lusarket I: a) notch, b) denticulate, c) single side-scraper, d)

convergent scraper, e) true Levallois point, f) Yerevan point (drawings by Rob Sherwood).

R.J. Birtwistle & G. Yeritsyan

13

Figure 5. Middle Palaeolithic artefacts from Lusarket I: a) elongated point, b) retouched elongated point, c)

Levallois blank (drawings by Rob Sherwood).

Moreover, at Karkustakau and Tamarasheni the Mousterian end-scrapers, leaf-like bifaces and Quina-type side scrapers are characteristic of invasively and bifacially worked pieces from the Northern Caucasus (Golovanova and Doronichev 2003). Given these strong typo-technological differences in lithic toolkits, it seems likely that Armenia was occupied at various times by culturally distinct groups of Neanderthals.

Furthermore, excavations at the Ortvale Klde rockshelter have shown that laminar-Levallois techniques also dominated this industry. The faunal assemblage reflects reliance upon Caucasian Tur, a pattern which may indicate seasonal migration (Adler et al. 2006 & 2008; Bar-Oz & Adler 2005; Bar-Yosef et al. 2006). The frequency of Caucasian Tur has parallels with the faunal recorded at Hovk 1 (Unit 5) and Kalavan 2 (Pinhasi et al. 2008), suggesting that this species was consistently targeted by LMP Neanderthal communities.

At present, AMS 14C measurements from relatively few sites suggest that some of the latest Middle Palaeolithic sequences in Europe are located in Armenia. Material from Unit 5 and Hovk 1 has been assigned to the Middle Palaeolithic on typological grounds, and is currently dated to 33,800 ± 500 14C BP; similarly, dates from nearby Kalavan 2 suggest that the Middle Palaeolithic ended around 34 kya in this corner of Armenia. The earliest Aurigacian dates from the region suggest that

it began around 38–34 k BP across north and south Armenia, as evidenced by dates from Warwasi Cave (Olszewski & Dibble 1994) and Yafteh Cave (Otte et al. 2007) in the Zagros, as well as Ortvale Klde (layers 4d and 4c) (Adler et al. 2008). If AMS 14C determinations from Lusarket I (Units C & D) and Yeravan Cave (Layer 7) are reliable, then the Hrazdan valley may have offered one of the final refugia for late Neanderthal groups in the region.

CONCLUSION

Analysis of the lithic and sedimentary data recovered from Unit D four decades ago has yielded disappointing results. The substandard excavating and recording techniques employed, combined with extensive rearrangement and disturbance of the assemblage, have limited the inferences that can now be made based on this assemblage. Nevertheless, some useful observations can be made based on the extant material; the surviving assemblage is discoidal in character; faceting and denticulation were commonly employed, and Levallois flaking was occasionally practiced. The retouched tools are dominated by scrapers and denticulates, predominantly worked on one edge of the dorsal face using short, irregular removals. On the whole, the Unit D assemblage is similar to industries from the Tsutskhvati Mousterian in Georgia. However, the assemblage from the overlying unit (Unit C) shares similarities with other Armenian sites, including Angeghakot

Lithics 33

14

and Yerevan Cave. Particular aspects of the Unit D assemblage are also typical of other Armenian sites, notably the occasional use of the truncated-facetted techniques and the presence of some Levallois elements. This may, however, reflect the fact that the boundary between Units C and D was subject to seasonal reworking, resulting in the intermixing of the assemblages. Although lithic assemblages from sites in the Caucasus are immensely varied, many share close typological affinities and can be considered as discrete entities. The existence of many discrete typo-technological entities within the same geographic region is interpreted here as reflecting repeated movement of Neanderthal populations across the region during the climatic instability of MIS 3.

Earlier Middle Palaeolithic (EMP) assemblages in Armenia share similarities with assemblages of “Tabun D-type” in the Levant. This might suggest an exogamous source for the EMP of the region between c.200–100 kys, rather than indigenous local development of similar techniques. The Caucasus acted as a geographic and cultural crossroads in the Middle Palaeolithic, movement through and within the region resulting in notable variability. New dating evidence suggests rapid replacement of Neanderthal communities as Aurignacian groups moved in from neighbouring regions. The possibility remains that Neanderthals may have survived into MIS 2 along the Hrazdan gorge, although this possibility remains unlikely. Further research, and especially tighter chronological resolution, may shed light upon late Neanderthal social and economic behaviour in the region before they were finally replaced by Homo sapiens c.28–39ka 14C BP.

ACKNOWLEDGEMENTS

The work reported in this paper was carried out as part of a Masters dissertation while the author was taking a MA in Palaeolithic Archaeology and Human Origins at the University of Southampton. The following individuals are thanked for the help they have provided. Dr Benik Yeritsyan (Institute of Archaeology and Ethnology, Yerevan) for permitting access to the Lusarket material; Boris Gasparyan (Institute of Archaeology and Ethnology, Yerevan) for his comments on core reduction techniques employed at Lusarket; Dr John McNabb (University of Southampton) for

comments on early drafts of the paper; Dr Keith Wilkinson (University of Winchester) for helping me organise my visit to Armenia in summer 2008, for his preparation of Figure 1, and for sharing his comments on the stratigraphy of Lusarket I; Dr Dan Adler (University of Connecticut) and Dr Ron Pinhasi (University College Cork) for their help with logistics and discussions in the field; Anna Mirrikyan for her translation of French and Russian archaeological papers and Rob Sherwood for his drawings of the lithics (Figures 4 and 5). Finally I would like to express particular thanks to my sister Emma Green for financially supporting the project and providing support when I was writing this paper.

REFERENCES

Adler, D.S., Bar-Oz, G., Belfer-Cohen, A. & Bar-Yosef, O. 2006. Ahead of the game: Middle and Upper Palaeolithic hunting behaviors in the Southern Caucasus. Current Anthropology 47: 89–118.

Adler, D.S., Bar-Yosef, O., Belfer-Cohen, A., Tushabramishvili, N., Boaretto, E., Mercier, N., Valladas, H. & Rink, W.J. 2008. Dating the demise: Neanderthal extinction and the establishment of modern humans in the Southern Caucasus. Journal of Human Evolution 55: 817–833.

Baiburtian, Y. 1937. New finds of Stone Age in Armenia. Sovetskaya Arkeologiya 3: (in Russian).

Baiburtian, Y. 1938. Tools in Ancient Armenia. Reports of the Armenian of the Academy of Sciences of the USSR1: (in Russian).

Bar-Oz, G. & Adler, D.S. 2005. Taphonomic history of the Middle and Upper Palaeolithic faunal assemblage from Ortvale Klde, Georgian Republic. Journal of Taphonomy 3: 185–212.

Bar-Yosef, O., Belfer-Cohen, A. & Adler, D.S. 2006. The implications of the Middle-Upper Paleolithic chronological boundary in the Caucasus to Eurasian Prehistory. Anthro-pologie XLIV: 49–60.

Beliaeva, E.V. & Liubin, V.P. 1998. The Caucasus-Levant-Zagros: possible relations in the Middle Palaeolithic. In M. Otte (ed.) Anatolian Prehistory: at the crossroads of two worlds: 39–55. Université de Liège, Service Préhistoire (Eraul 85), Liège.

Binford, L.R. 1973. Interassemblage variability: The Mousterian and the ‘functional’ argument. In C. Renfrew (ed.) The explanation of culture change: 227–254. Duckworth, London.

Binford, L.R. 1980. Willow smoke and dogs’ tails: hunter gatherer settlement systems and archaeological site formation. American Antiquity 45: 4–20.

R.J. Birtwistle & G. Yeritsyan

15

Birtwistle, R.J. 2008. Neanderthal behavioural patterns in the Hrasdan Valley, Armenia, seen through there core reduction strategies. Unpublished M.A. Thesis, Southampton University.

Bocherens, H. 2011. Diet and ecology of Neanderthals: Implications from C and N isotopes. In N.J Conard & J. Richter (eds.) Neanderthal lifeways: subsistence and technology: 73–85. Springer, Tubingen.

Chataigner, C., Badalian, R., Bigazzi, G., Cauvin, M.C., Jrbashian, R., Karapetian, S.G., Norelli, P., Oddone, M. & Poidevin, J.L. 2003. Provenance studies of obsidian artefacts from Armenian archaeological sites using the fission-track dating method. Journal of Non-Crystalline Solids 323: 16–171.

Dzhafarov, A.K. 1983. Mustierskaya kultura Azerbaidjana (po materialam Taglarskoi pesheri). Elm, Baku. (in Russian).

Dolukhanov, P., Aslanian, S., Kolpakov, E. & Belyaeva, E. 2004. Prehistoric sites in Northern Armenia. Antiquity Project Gallery: http://62.189.20.34/projgall/dolukhanov/index.html.

Ebert, J. I. 1979. An ethnoarchaeological approach to reassessing the meaning of variability in stone tool assemblages. In Kramer (ed) Ethnoarchaeology: Implications of Ethno-graphy for archaeology:61–74. Columbia university press, New York.

Formozov, A.A. 1977. Problemi etno-kul’turnoi istorii kamennogo veka na territorii evropeiskoi chasti SSSR. Nauka Press, Moscow. (in Russian).

Fourloubey, C., Beauval, C., Cologne, D., Liagre, J., Ollivier, V. & Chataigner, C. 2003. Le Paléolithique en Arménie: État des connaissances acquises et données récentes. Paléorient 29: 5–18.

Gabunia, L., Vekua, A., Lordkipanidze, D., Swisher III, C. C., Ferring, R., Justus, A., Nioradze, M., Tvalchrelidze, M., Antón, S.C., Bosinski, G., Jöris, O., de Lumley, M.A., Majsuradze, G. & Mouskhelishvili, A. 2000. Earliest Pleistocene hominid cranial remains from Dmanisi, Republic of Georgia: taxonomy, geological setting and age. Science 288: 1019–1025.

Gasparyan, B., Nahapetyan, S., Arakelyan, D., Cologne, D. & Chataigner, C. 2004. The Stone One age sites of the Aparan depression (Republic of Armenia). In G. Gambashidze (ed.) Collection of Short Reports of the International Conference “Archaeology, Ethnology and folklore of the Caucasus”: 49–50. (in Russian).

Golovanova, L.V. & Doronichev, V.B 2003. The Middle Palaeolithic of the Caucasus. Journal of World Prehistory 17: 71–140.

Hardy, B.L. & Moncel, M-H. 2011. Neanderthal Use of Fish, Mammals, Birds, Starchy Plants and Wood 125-250,000 Years Ago. PLoS ONE 6(8): e23768. doi:10.1371/journal.pone.0023768

Jaubert, J. & Ollivier, V. 2003. Sites Paléolithiques de la region d’Aparan. In. Rapport sur les Opérations effectuées en 2003 par la mission “Caucasus”. Unpublished CNRS Report: 2 – 31.

Liagre, J., Gasparyan, B., Ollivier, V. & Nahapetyan, S. 2007. Angeghakot-1 and the identification of the Mousterian cultural facies of “Yerevan points”, type in the Southern Caucasus. Paléorient 32: 5–18.

Liubin, V.P. 1977. Mustierskie Kul’turi Kavkaza. Nauka Press, Leningrad.

Liubin, V.P. 1984. Early Paleolithic of Caucasus. In P.I. Boriskovski (ed.) The Paleolithic of USSR: 45–93. Nauka Press, Moscow. (in Russian).

Liubin, V.P. 1989. Paleolit Kavkaza. In P. I. Boriskovski (ed.) Paleolit Kavkaza I Severnoi Azii: 7 – 142. Nauka Press, Leningrad.

Mansurov, M.M. 1990. Novye paleoliticheskie pamjatniki v srednem basseine r. kuri. In D. Tushabramishvili (ed.) Paleolit kavkaza I sopre-del’niki territorij: 40–42. Metsniereba, Tbilisi.

Meignen, L. & Tushabramishvili, N. 2006. Paléolithique moyen laminaire sure les flancs sud du Caucasus: productions lithiques et fonctionnement du site de Djruchula (Géorgie). Paléorient 32: 81–104.

Mikhailov, V.O., Panina, L.V., Polino, R., Koronovsky, N.V., Kiseleva, E A., Klavidieva, N.V. & Smolyaninova, E.I 1999. Evolution of the North Caucasus foredeep: constraints based on the analysis of subsidence curves. Tectonophysics 307: 361– 79.

Olszewski, D.I. & Dibble, H. 1994. The Zagros Aurignacian. Current Anthropology 35: 68–75.

Otte, M., Biglari, F., Flas, D., Shidrang, S., Zwyns, N., Mashkour, M., Naderi, R., Mohaseb, A., Hashemi, N., Darvish, J. & Radu, V. 2007. The Aurignacian in the Zagros region: new research at Yafteh Cave, Lorestan, Iran. Antiquity 81: 82–96.

Pinhasi, R., Gasparian, B., Wilkinson, K., Schreve, D., Branch, N. & Nahapetyan, S. 2006. The archaeology of Hovk, north-east Armenia: a preliminary report. Antiquity Project Gallery: http://antiquity.ac.uk/ ProjGall/Pinhasi/index.html.

Lithics 33

16

Pinhasi, R., Gasparian, B., Wilkinson, K., Bailey, R., Bar-Oz, G., Bruch, A., Chataigner, C., Hoffmann, D., Hovsepyan, R., Nahapetyan, S., Pike, A.W.G., Schreve, D. & Stephens, M. 2008. Hovk 1 and the Middle and Upper Palaeolithic of Armenia: a preliminary framework. Journal of Human Evolution 55: 803–816.

Pinhasi, R., Higham, T.F.G., Golovanova, L.V. & Doronichev, V.B. 2011. Revised age of late Neanderthal occupation and the end of the Middle Paleolithic in the northern Caucasus. Proceedings of the National Academy of Sciences of the United States of America 108: 8611–8616.

Shea, J.J. 1988a. Spear points from the Middle Paleolithic of the Levant. Journal of Field Archaeology 15: 441–450.

Shea, J.J. 1988b. Tool use in the Levantine Mousterian of Kebara Cave, Mount Carmel. Mitekufat Ha’even: Journal of the Israel Prehistoric Society 22: 15 – 30.

Speth, J.D. 2010. The paleoanthropology and archaeology of big game hunting: Protein, fat, or politics. Springer, New York.

Stepanchuk, V. 1998. The Crimean Palaeolithic: Genesis and Evolution between 130–30kyr BP. In M. Otte (ed.) Anatolian Prehistory at the Crossroads of Two Worlds: 261 – 300. Université de Liège, Service Préhistoire (Eraul 85), Liège.

Tushabramishvili, D.M. 1978. Arkheologicheskie panyatniki Tsutskhvatskogo Mnogoetajnogo

peshernogo kompleska. Metsniereba, Tbilisi..

Tushabramishvili, N., Lordkipanidze, D., Vekua, A., Tvalcherlidze, M., Muskelishvili, A. & Adler, D. 1999. The Palaeolithic rockshelter of Ortvale Klde, Imereti Region, The Georgian Republic. Préhistoire Européenne 15: 65–77.

Wilkinson, K. 2008. A summary of the stratigraphy encountered at Lusarket and Nur Gehry during excavation in July–August 2008. Unpublished Report.

Wilkinson, K. 2009. Geoarchaeological fieldwork in the Razdan Gorge 2009. Unpublished Report

Yeritsyan, B.G. 1970a. Paleolithic cave site of Yerevan and its place among the earliest sites of the Caucasus. Unpublished PhD Thesis, University of Moscow. (in Russian).

Yeritsyan, B.G. 1970b. New Paleolithic site Yerevan II: selected papers of Archaeological openings in 1969. Nauka Publishers, Moscow. (in Russian).

Yeritsyan, B.G. 1975. The New Lower Palaeolithic cave site Lusarket 1 (Armenia). Kratkiye Soobshcheniya o Dokladakh I Polevykh Issledovaniyakh 141: 54 – 67. (in Russian).

Yeritsyan, B.G. & Ghazarian, H.P. 1978. The work of the Lusarket expedition. In. Selected Papers in Archaeology: 498–499. Nauka Publishers, Moscow. (in Russian).