Keywords

Introduction

Cave systems are a typical feature of karst landscapes where they develop through the dissolution of soluble bedrock, leading to the formation of a variety of open-space cavities and passageways. They represent a geomorphologic link between the surface and sub-surface environments, provide a conduit for the flow of groundwater and act as a natural repository for the accumulation of sediment (e.g., Bogli 1980; White 1988; Sasowsky and Mylroie 2004). Caves also act as natural shelters for animals and have been similarly exploited for thousands of years by humans. Consequently, caves and rock shelters represent a habitat that may have assisted in the intellectual development of human kind (e.g., Chauvet et al. 1996).

In recent years, investigations into the geological, hydrogeological and biogeochemical properties of cave systems have transformed our understanding of cave genesis (Moore and Sullivan 1997; Engel et al. 2004; Ford and Williams 2007). In particular, studies of cave-fill sediments have provided new insights into the timing of cave formation, subterranean environmental processes, karst landscape evolution, groundwater dynamics and paleoclimatology (e.g., Bretz 1942; Polyak et al. 1998; Musgrave and Webb 2004; Pickering et al. 2007; White 2007). Thus, the study of cave sediments can increase our understanding of the sedimentological and environmental conditions that existed during the formation and evolution of a cave system. In addition, cave sediment deposits are frequently associated with fossil preservation and have proven to be important sites for the discovery of archaeological artifacts and hominin remains (Jelinek 1982; Torres et al. 2003; Pinhasi et al. 2008; Dirks et al. 2010; Moldovan et al. 2011; Pickering et al. 2011).

This chapter reviews the stratigraphic and sedimentary characteristics of Azokh Cave, with particular emphasis on its three main entrance passages: Azokh 1, Azokh 2 and Azokh 5. We summarize the nature of cave-fill sediments within Azokh 1 passage, as previously reported by Murray et al. (2010), and provide additional data (sedimentological, textural, mineralogical) that further constrain the stratigraphy. Azokh 1 has been the principal site of archaeological excavation at the cave since 2002 (Fernández-Jalvo et al. 2009, 2016). In addition, an assessment of the stratigraphy and sedimentology of Azokh 2 and Azokh 5 entrance passages is also made. Knowledge of the sedimentary infill within these smaller passageways is presently at a reconnaissance level and the data presented herein should thus be viewed as preliminary.

Geological Setting and Overview of the Azokh Cave System

Azokh Cave (located at 39º 37.15′ north; 46º 59.32′ east; Fig. 2.1) is hosted in a thickly-bedded sequence of Mesozoic (possibly Jurassic) limestone that has experienced variable levels of uplift and karstification since the Pleistocene (Lioubine 2002). The carbonate bedrock forms part of a limestone massif that is developed on a regional-scale across the Southern Caucasus (Khain 1997). In the vicinity of the cave, the host limestone is a fossiliferous grainstone which has undergone partial silicification, possibly due to the combined input of volcanogenic siliciclastics and siliceous fossil material. The cave system comprises a series of dissolution cavities that may have developed partly in response to vadose zone fluviokarst processes (e.g., White 1988; Domínguez-Alonso et al. 2016). The cave consists of a NNW- to SSE-aligned internal zone that is composed of several interconnected, sub-rounded chambers extending over approximately 130 m (Fig. 2.2). This main body of the cave is transected on its western flank by several WSW- to ENE-trending entrance passageways that connect the internal zone to the exterior (Figs. 2.2 and 2.3). The orientation of the main chamber and entrance passages broadly corresponds with the alignment of conjugate joint sets and fractures that are pervasively developed in the limestone bedrock (Domínguez-Alonso et al. 2016).

Fig. 2.1
figure 1

Geographic location of Azokh Cave in the South Caucasus. a General view of region relative to the Eastern Mediterranean, Black and Caspian seas. b More detailed view of location, with an indication of topographic elevation shown. Position of (b) is indicated by inset box in (a)

Full size image
Fig. 2.2
figure 2

a Field photograph of the west-facing hillside containing the Azokh Cave system. Locations of Azokh 1, 2 and 5 entrance passages are indicated. b Simplified plan-view sketch map of the cave system showing the location of the main entrance passages. Internal cave chambers are labeled with roman numerals (I–IV). Reproduced from Murray et al. (2010)

Full size image
Fig. 2.3
figure 3

Field photographs of entrance passages to Azokh Cave. a Azokh 1. The distinctive sediment pedestal marking the entranceway is visible towards the bottom of the image. The white asterisk indicates the approximate position of the original sediment infill of the passage, prior to excavation in the 1960s and 70s. b Azokh 2 and c Azokh 5. Both (b) and (c) were photographed in 2004. The hammer for scale (highlighted with a white arrow in both images) is 35 cm long

Full size image

The initial investigations of Azokh Cave by the Huseinov team in the 1960s did not establish a clear and detailed record of the Pleistocene and Holocene sedimentary infill of the cave system (Mustafayev 1996; Lioubine 2002). In particular, early excavations lacked any rigorous stratigraphic control. Sedimentary units were more commonly distinguished based on their archaeological content, rather than their sedimentological properties (Lioubine 2002). Definitive thickness estimates for the sedimentary units identified at that time appear not to have been unequivocally established and any lateral shifts in sedimentary facies were not made apparent. In addition, systematic archaeological excavation methods, utilizing aerial grids and three-dimensional spatial recording of finds, were not employed. These factors have combined to make the understanding of the context and significance of the large volume of fossil and lithic artifacts recovered prior to 2002 a challenging prospect.

Azokh 1

The Azokh 1 passageway is a broadly linear chamber measuring 40 m long by 11.5 m high with a WSW-ENE alignment (Fig. 2.4). This orientation results in the entranceway being well illuminated, particularly by the afternoon sun. Towards the interior of the passage, in an ENE direction (“Uppermost Platform” in Fig. 2.4), the light is not as good and artificial illumination has been employed there during excavation work. The floor of the chamber drops (slopes) down at approximately the midway point in the passage, which increases the height of the chamber to approximately 14 m towards the entrance. The sedimentary infill of Azokh 1 passage yielded a human jaw fragment in 1968 that was later assessed as Middle Pleistocene in age (Kasimova 2001). The nature of this discovery in the southern Caucasus (Fig. 2.1), coupled with additional archaeological and paleontological finds, has established Azokh 1 as a site of significant archaeological and paleoanthropological interest (e.g., Ljubin and Bosinski 1995; Bridgland et al. 2006; Fernández-Jalvo et al. 2010; Pinhasi et al. 2011).

Fig. 2.4
figure 4

a Sketch cross-section through Azokh 1 cave passage (drawn facing NW). The estimated amount of cave-fill sediment removed by previous excavation teams is indicated by the upper dashed line. The floor of the passage is illustrated insofar as its extent is currently known and the height of the roof was measured at various points along the section using a telemeter with an accuracy of 1 cm. b Cross-section (A–A′) across the axis of the passage [orthogonal to (a) and drawn to the same scale] indicates levels of sediment infill. Reproduced from Murray et al. (2010)

Full size image

During these early phases of excavation, a considerable amount of sediment was removed from the passageway (Fig. 2.3a), and Lioubine (2002) noted that before the first excavations in the 1960s, the chamber was filled to within 2–3 m of the roof. A graphical estimate of the original sediment thickness is provided in Figs. 2.3a and 2.4 and it is apparent that a considerable amount of the stratigraphic section is now gone. Huseinov initially identified 10 stratigraphic horizons infilling the chamber during the 1960s. This was increased to 17 by Veilicko in 1979 and then to 25 by the Gadzhiev team in 1980 (Huseinov 1985; Lioubine 2002). Detailed records of the extent of these excavations and the amount of sediment removed are no longer readily available. Therefore, an appraisal of the sedimentology incorporating the full pre-2002 stratigraphic sequence is extremely difficult to ascertain. In this regard, what is presented below is a description based on the stratigraphic remnants that we found remaining in the passage.

When excavation work restarted in 2002, the Azokh 1 passageway was an obvious priority for renewed investigation and was initially termed Azokh Main. Subsequently, the passage was renamed Azokh 1 following reconnaissance geophysics and geological work that identified appreciable thicknesses of sediment fill in two other entrance passageways – Azokh 2 and Azokh 5 (Fig. 2.3b, c; see also Fig. 2.2 for general location). Post-2002, systematic and detailed archaeological investigations have been conducted in the upper half of the sedimentary sequence remaining in Azokh 1 (Asryan et al. 2016; Fernández-Jalvo et al. 2016; King et al. 2016).

The most recent assessment of the stratigraphy of Azokh 1 was provided by Murray et al. (2010) and their proposed lithostratigraphic framework is retained here. Nine sedimentary units, occurring within two physically separated stratigraphic remnants (termed Sediment Sequences – see Fig. 2.4), are recognized based on their sedimentological properties. Sediment Sequence 1 is located at the ENE end of a basal trench at the cave entrance and accounts for 4.5 m of stratigraphy (Fig. 2.5). Sediment Sequence 2 is located towards the rear of Azokh 1 passage and is estimated to be at least 8.5 m thick (Fig. 2.6). This latter sequence is interpreted to have overlain the former, although since no physical connection remains between the two sequences, and practically no sediments remain along the sides of the cave walls, this inference is equivocal.

Fig. 2.5
figure 5

Stratigraphic column for Sediment Sequence 1 in Azokh 1 passage. Unit numbers are indicated in the central column with roman numerals. The photograph of the actual section to the left of the column is for reference and indicates precisely where the boundaries of the units have been set. Much of this section is exposed in the basal trench in the entrance to the cave (see Fig. 2.4). The key to the various sedimentological features is also applicable to Fig. 2.6. Reproduced from Murray et al. (2010)

Full size image
Fig. 2.6
figure 6

Composite stratigraphic column for Sediment Sequence 2 in Azokh 1 passage. The height of the wooden ruler in the lower and middle photograph is 2 m whilst in the upper photograph the length of the tape is 88 cm. See Fig. 2.5 for a general sedimentological key. Modified from Murray et al. (2010)

Full size image

Table 2.1 summarizes the main stratigraphic subdivisions of the infill of Azokh 1. It provides average estimates of the color, texture and sedimentary characteristics of the various lithostratigraphic units and is built upon the descriptions presented in Murray et al. (2010). Detailed excavation work, particularly towards the top of the stratigraphic succession, has revealed much intra-unit variation. This has become particularly evident as horizons have been tracked laterally from the center of the passageway, where most of the lithostratigraphic units were originally diagnosed, and out towards the cave walls.

Table 2.1 Stratigraphic divisions and character of the remaining sediments in Azokh 1
Full size table

Sediment Sequence 1

This sequence occupies a lower sub-level within Azokh 1 passage (Fig. 2.4) and it contains Units IX to VI (Fig. 2.5). Given that it rests on a down-sloping cave floor surface (Fig. 2.7; see also Fig. 2.4), the section effectively wedges out, so the amount of remaining stratigraphy becomes progressively more limited moving downwards. The base of the sequence (Units IX and VIII) is largely composed of non-calcareous sandy loam/loamy sand (Table 2.1). A gradual and pronounced development of a granular structure midway through Unit IX (Fig. 2.8) marks the contact between its two constituent subunits. The overlying Unit VIII is characterized by a higher concentration of limestone and chert clasts. The contacts of Unit VIII with its bounding (enclosing) units are not sharply defined and it is possible that it may represent a localized accumulation of larger clast types within a loamy sand matrix.

Fig. 2.7
figure 7

Thinning of the base of what remains of Sediment Sequence 1 (Unit IX) which rests on the sloping cave floor. The tape measure (for scale) is showing 1 m

Full size image
Fig. 2.8
figure 8

Detail of the transition seen in the middle of Unit IX. The base of the unit (IXb) is more massive in character whilst the upper half (IXa) becomes progressively more granular in appearance towards the top of the photograph. The visible length of the scalebar is 86 cm

Full size image

Units VII and VI above see a shift to clay-loam textures, with the exception of subunit VIc which is a conspicuous clast-supported pebble to cobble conglomerate (Fig. 2.9). The two subunits (VIb and VIa) overlying the conglomerate are calcareous, and this contrasts with the non-calcareous units beneath. Sediment Sequence 1 is capped by a large limestone boulder, which has presumably fallen into position from the cave roof. Excavation by previous teams around this collapse feature has resulted in the characteristic “mushroom” shaped pedestal close to the entrance to the passage (Figs. 2.4 and 2.5).

Fig. 2.9
figure 9

Clast-supported subunit VIc conglomerate. Clasts are sub- to well-rounded. Scale bar is 21.5 cm in length

Full size image

Fossils and lithic artifacts have not been observed or recorded so far in Units VII, VIII and IX. Given the limited extent of the remaining stratigraphy this is perhaps unsurprising and it partly explains why this portion of the succession remains largely undated (Table 2.1). Huseinov (1985) reported the recovery of very fragmentary fossils from this lower part of the stratigraphy, along with pollen. Clearly identifiable fossil fragments and charcoal are present in Unit VI towards the very top of Sediment Sequence 1. Murray et al. (2010) speculated that this divide between (largely) unfossiliferous and fossiliferous strata might be a reflection of a shift between the cave being closed during accumulation of most of Sediment Sequence 1 to a more open system towards the top. In particular, conglomeratic subunit VIc (Fig. 2.9) is unequivocally the product of energetic water flow through the passage (probably a small river) and the coincidence of this horizon with the first appearance of fossils supports this contention.

According to M.M. Huseinov (reported in Lioubine 2002), Sediment Sequence 1 equates to “cultural” layers V–VI (at the top) and VII down to X (below) [compare Fig. 2.5 herein to Fig. 8 of Lioubine (2002)]. Huseinov (1985) recorded over 200 lithic artifacts from layers VII–X (as he had interpreted the strata); however, the validity and stratigraphic integrity of these finds has subsequently been questioned and claims that they are “Lower Paleolithic” in character have been largely dismissed (Doronichev 2008; Doronichev and Golovanova 2010).

Sediment Sequence 2

This is a composite sequence that has been reconstructed from a series of vertical sections or “steps” (largely a by-product of the pre-2002 excavations) in the cave filling strata (Fig. 2.4). Sediment Sequence 2 can be subdivided into five constituent units (I–V; Fig. 2.6) totaling about 8.5 m in thickness. Over half of this thickness is accounted for by Unit V (approximately 4.5 m), which is located at the base. All five units of Sediment Sequence 2 have proven to be fossiliferous and much of the excavation work by the current team has been focused in this part of the succession.

Unit V is predominantly fine-grained in character and is divisible into two subunits: Vb (located at the base and largely non-calcareous) and Va (located directly above and calcareous in nature; see Table 2.1). It is likely that Unit V can be further subdivided beyond this two-part scheme; however, subunit Va presents a steep vertical face (just over 2 m; Fig. 2.6) in the section and, for safety reasons, it has not been possible to thoroughly examine the stratigraphy in detail.

Subunit Vb is best exposed in a small trench that was initially excavated through the Middle Platform in 2002 (Fig. 2.10; see also Fig. 2.4a for general location in Azokh 1 passage). Murray et al. (2010) described five horizons within this trench section and a refinement of some of their sedimentological details is outlined in Table 2.2.

Fig. 2.10
figure 10

The base of Sediment Sequence 2. a General location of the geological trench in the base of Unit V within Azokh 1 passage. This photo faces ENE and was taken in 2008, when the bedrock floor to the cave was found only a short distance below the (present) sediment level on the “Lower Platform”; b 1 Photo and b 2 corresponding scaled panel diagram of the sedimentary succession in the geological trench. A description of horizons (a) to (e) is provided in Table 2.2. Tape measure in b 1 is showing 1 m. b 2 is modified from Murray et al. (2010)

Full size image
Table 2.2 Subunit Vb succession evident in geological trench, Middle Platform, Azokh 1
Full size table

The uppermost horizon (e) of subunit Vb can be traced laterally across the excavation surface of the Middle Platform and is seen to continue stratigraphically upwards for a further 35–40 cm. It is capped in places by a distinctive 1 cm-thick cream-white to white, non-calcareous phosphatic crust (see lowest photo correlation line at top of subunit Vb in Fig. 2.6), which forms a useful marker horizon.

Subunit Va is 220–230 cm thick and is predominantly composed of friable calcareous silty clay. The basal 55 cm is granular with common angular limestone clasts (2–10 mm), which are typically flattened parallel with bedding. The overlying 105 cm is more massive in structure and contains a distinctive horizon of flattened (cm-scale) clasts in the top third (see photo correlation line in Fig. 2.6). Charcoal was noted in this zone also. The uppermost 70–80 cm of subunit Va comprises friable calcareous silty clay. Its base is finely granular, however, its top is predominantly massive, lacks limestone clasts, and has a more reddish hue (resulting in 7.5YR rather than 10YR color designation; see Table 2.1). This subtle color transition is generally gradual in nature.

The contact between the top of Unit V and overlying Unit IV is diffuse; and is irregular and undulose when tracked laterally from the centre of the passage towards the cave walls. Where it is more clearly displayed it presents a subtle shift in texture (moving upward from predominantly massive to fine granular) and color (the “reddish” 7.5YR top of Va is overlain by 10YR Unit IV; see Table 2.1). A characteristic feature of Unit IV is a progressive increase in flattened sub-angular to rounded (cave-wall) pebbles and cobbles towards the top of the unit, along with fragments of bone and charcoal.

When examined in the centre of the passage, the contact between Unit IV and (overlying) Unit III is quite obvious and sharp (see relevant photo correlation line in Fig. 2.6) and is marked by a shift in structure and a noticeable decrease (in Unit III) in the limestone clast content of the sediments. However, the contact has proven difficult to trace laterally when moving away from the centrally positioned reference section. At the time of writing, detailed excavation has begun to reveal more (from a lateral perspective) of this transition and it is likely that a reassessment of this particular contact may have to be made with new exposure. A possible two-part subdivision of Unit III into a lower (largely) massive subunit and an upper weak to moderate granular subunit is also becoming apparent.

The contact between Unit III and (overlying) Unit II is conspicuous and is defined by a marked increase in the granularity of the sediments (Fig. 2.11). Murray et al. (2010) noted reddish-brown staining along this contact close to the northwestern wall of the chamber. Analysis of red- and orange-stained sediment from several units in Azokh 1 using Raman spectroscopy indicates the presence of fine-grained hematite and magnetite within the sediment (see below for further discussion). Subsequent excavation of the Unit III/II boundary has shown the hematitic staining to be more laterally widespread and the irregular nature of the contact to be more pronounced than initially thought.

Fig. 2.11
figure 11

Lateral view of contact between Units III and II, Azokh 1. Hammer (arrowed) for scale

Full size image

Unit II rapidly (and somewhat irregularly) becomes non-calcareous upwards and also contains an elevated amount of limestone clasts (0.5–5 cm). These clasts, along with fossil bone fragments, are often strongly degraded, particularly in the non-calcareous zones. The deterioration of bone material within Unit II has been linked to accumulations of bat guano during its deposition, resulting in a non-calcareous, more acidic sediment (Murray et al. 2010). These authors reported the detection of tinsleyite (K and Al-rich hydrated phosphate) in the sediment. This particular mineral phase likely reflects syn-diagenetic processes where phosphatic mineralisation can precipitate due to the presence of bat guano (Magela da Costa and Rúbia Ribeiro 2001; Marincea et al. 2002; Shahack-Gross et al. 2004). It is evident that there is considerable lateral heterogeneity within Unit II in terms of its consistency, texture, geochemistry and the quality of taphonomic preservation (personal observations; see also Smith et al. 2016 and Marin-Monfort et al. 2016). Unit II was initially examined in a small cut section, near the Upper Platform and adjacent to the northern wall of the cave passage, where it measured c. 120–140 cm in thickness. More recent investigations of newly exposed surfaces of the Unit III/II contact in a more central position within the chamber, and also the overlying Unit II/I contact located approximately five meters deeper within the cave passage, have suggested potential thickness variation for Unit II of 150–200 cm. However, since these contacts are exposed in different positions within the passage, and neither section reveals Unit II in its entirety, it is unclear whether the thickness disparities inferred (from the differences in the elevations of the contacts) reflect real lateral thickness variation, or simply a slope in either/both of the unit boundaries towards the cave entrance.

The contact between Units II and I is sharp and irregular when traced out in detail, and the latter appears to infill the uneven topographic surface of the former. Unit I, which caps the entire cave-fill succession, is non-calcareous and predominantly a friable to loose clay loam. Excavation work on the Uppermost Platform (Fig. 2.4a) has shown this unit thins from more than 135 cm to between 80 and 90 cm towards the interior of the cave. A reference section for Unit I has been preserved in the rear of Azokh 1 passage (Fig. 2.12). Murray et al. (2010) noted that this particular section demonstrated two key features:

Fig. 2.12
figure 12

a Photograph of the top of Unit I after it was exposed and cleaned during the 2007 field season. The visible length of the tape measure is 88 cm. b Sketch interpretation of the photograph in (a) showing hearth layer (fumier), disturbance by burrows and eroded top of Unit I. Reproduced from Murray et al. (2010)

Full size image

  • Considerable disturbance and reworking of the sediment by recent mammal burrowing activity. Fossils of Ursus spelaeus and coprolites, as well as Paleolithic stone tools have been recovered from these burrows (Marin-Monfort et al. 2016). This large-scale bioturbation has served to greatly complicate the internal stratigraphic details of Unit I.

  • Close to the top of the unit a conspicuous c. 30–40 cm thick fumier (manure hearth) occurs (Fig. 2.12; see also Fig. 2.6). This feature consists of a series of black, carbon-rich bands with greyish-white ash-rich interlayers. Dispersed, but common, soft white carbonate granules (occasionally these are decalcified) in the top 35 cm of Unit I may possibly be related to the heating effect of this large hearth structure on the surrounding sediment.

A conspicuous component of several of the stratigraphic units within Azokh 1 is the presence of disseminated clay-like pedofeatures within the sediment groundmass (Fig. 2.13). These features typically occur as millimeter- to centimeter-scale, sub-circular to lensoidal nodules and disseminated specks, as well as thin (c. 1–3 mm) discontinuous sub-horizontal seams (typically 1–5 cm long). They are composed of fine-grained (<0.05 mm), white to buff, powdery, clay-like material and generally do not display any internal banding or lamination. Similar clay-like material also forms partially developed concentric laminae within and around decomposing bone fragments in the sediment (Fig. 2.13d). The occurrence of these nodules appears to begin within Unit VI at the top of Sedimentary Sequence 1 and remains variably developed, moving up the stratigraphy, throughout Sedimentary Sequence 2 (Units V–I). This distribution appears to broadly correlate with marked increases in numbers of fossils and the calcareousness of the sediment (see Table 2.1).

Fig. 2.13
figure 13

a Phosphate nodules in Unit VIa. Nodules occur as cream to white, clay-like lenticular concretions and disseminated specks, or as partially developed concentric laminae within decomposing bone material (arrowed). b, c Detailed views of phosphate nodules disseminated in Unit VIa. In (c) the lenticular form of the nodule in the center of the image is similar to the shape of the bone material shown in (a) and (d). d Partially decomposed bone fragment in Unit Vb displaying concentric phosphate (white) and hematite (dark grey) banding. e Contact between Unit Vb and Va partially defined by a 1 cm nodular seam of phosphatic material (arrowed). The dark grey, sub-vertical patches seen in Unit Vb represent hematitic-stained sediment. View looking approximately east. f Rounded hematite nodule in Unit IXa (arrowed)

Full size image

The analysis of nodule material was performed using Raman spectroscopy in an attempt to characterize its mineralogy/composition and help identify a likely formational mechanism. Representative material was sampled from several stratigraphic horizons (e.g., Units VIa, VIb, Va, Vb and IV; Fig. 2.13) and analyzed following the procedure outlined in the Spectroscopy Methodology section (preceding the references). Preliminary results indicate that the nodules are predominantly composed of fine-grained phosphatic material including apatite and/or hydroxylapatite (Fig. 2.14). Representative Raman spectra display strong peaks shifts in the range 967–1020 cm−1, diagnostic of phosphatic minerals (e.g., Sinyayev et al. 2005; Kizewski et al. 2011). The variation in the width of these Raman peaks for phosphates likely reflects a spectroscopic response between crystalline phases (narrow peak) and more amorphous mineral forms giving broader peaks (cf. Fig. 2.14).

Fig. 2.14
figure 14

Representative Raman spectra of mineral phases present within several Azokh 1 sedimentary units. Diagnostic peak positions are labeled using Raman shift values (cm−1). Raman intensity is in arbitrary units (a.u.)

Full size image

The likely provenance of the phosphatic nodules includes the weathering of bone material, as evident by its association with partially decomposed bone fragments (Fig. 2.13a, d), and/or diagenetic formation following the syn-sedimentary accumulation of bat guano and a subsequent increase in the concentration of dissolved phosphate within infiltrating aqueous fluids (e.g., Karkanas et al. 2000, 2002; Shahack-Gross et al. 2004). Thus, the nodules appear to be autochthonous and formed as a result of post-sedimentation weathering/alteration and diagenetic processes resulting in the formation of authigenic phosphate. This hypothesis is supported by the general disseminated, undeformed and granular appearance of the nodules that display little evidence of mobilization or re-working.

Additional diagenetic features of the sedimentary sequence within Azokh 1 include the occurrence of rust-red to orange-brown colored staining, coatings, nodules and grains throughout the succession (Fig. 2.13e, f). Raman analysis of representative orange-stained sediment and sub-rounded nodules and specks from Units IXa, IXb and Vb indicates that this material is primarily of iron oxide composition and is dominantly hematite with lesser magnetite (Fig. 2.14).

Dating and Correlating the Sediment Sequences

A range of radiometric dates for Sediment Sequence 2 is reported in the Appendix of this volume. Age determinations are included here and also summarized in Table 2.1. Moving from the base to the top of the sequence:

Units V and IV: Uranium series dating suggested an age of c. 200 ka for Unit V, whilst racemization (D/LAsp) indicated an age closer to 300 ka. However, the most up to date ESR estimate indicates an age of 293 ± 23 ka. An ESR date of 205 ± 16 ka has been calculated for the base of Unit IV, very close to the contact with underlying Unit V.

Unit III: No dates are available for this unit.

Unit II: Murray et al. (2010) noted an unsuccessful attempt to radiocarbon date this unit and suggested its age likely exceeded the lower radiocarbon range of 60 ka. Subsequent ESR dating has provided an age of 184 ± 13 ka for the base and 100 ± 7 ka for the top of Unit II (see Appendix, ESR).

Charcoal from the fumier in Unit I provided a radiocarbon age of 157 ± 26 14C BP (see Appendix, radiocarbon). Murray et al. (2010) noted that a Russian coin, from around the mid-1960s, was discovered in 2006 (although it had been moved by subsequent bioturbation). Below the hearth in Unit I, the sediments are highly disturbed (Fig. 2.12), so confident dating this unit remains problematic.

It is clear that Sediment Sequence 2 ranges in age from Middle to Late Pleistocene (Units V to II; Table 2.1). The Middle Pleistocene age for Unit V is significant as it is from this level in the succession that the hominin mandible was recovered in the 1960s (Kasimova 2001; see also King et al. 2016). The sharp, irregular contact between Units II and I at the top of Sequence 2 is disconformable and may represent a hiatus in sedimentation, with possible subsequent erosion, between Late Pleistocene and Holocene times (Table 2.1). This relationship suggests that the Pleistocene-Holocene boundary transition is not fully represented in the succession (Murray et al. 2010).

The details of the age of Sediment Sequence 1 remain unclear. Attempts to resolve the matter are hampered by two principal factors:

  1. 1.

    The limited extent of the remaining stratigraphy. This has already been discussed, but the lack of fossil remains and bona-fide lithic artifacts in Units IX to VII is also problematic.

  2. 2.

    Uncertainty in how the two sediment sequences precisely correlate. This is a function of the fact that no in-situ sediment connection now remains between the two sediment sequences in Azokh 1. The simplest view of the situation (Occam’s razor) would be to assume that Sediment Sequence 1 is positioned at a lower level in the passage (Fig. 2.4) and therefore must stratigraphically (directly) underlie Sediment Sequence 2. However, with the lack of information about the lateral connection of strata, it is impossible to establish with any degree of certainty if the cave-fill sequence is progradational. The presence of conglomeratic subunit VIc (Fig. 2.9) signifies a period of increased water flow through the passage. This may have eroded parts of any pre-existing strata, introducing a time gap of unknown duration into the sequence, casting an element of doubt into the assumption that Sediment Sequence 1 records a smooth, unbroken succession from Middle Pleistocene (near the top) to older times (moving stratigraphically downwards).

Correlation between the two sediment sequences in Azokh 1 is discussed in detail by Murray et al. (2010), who propose several possibilities:

  • Using the bedrock floor of the cave as a datum, the base of subunit Vb in Sequence 2 is equivalent to the base of subunit VIc or possibly even the upper portion of Unit VII in Sequence 1.

  • The highly conspicuous conglomeratic subunit VIc in Sediment Sequence 1 (Fig. 2.9) may correlate with the lenticular unit (horizon (c) in Fig. 2.10b) in the base of Vb. The latter displays an erosive, channel style geometry and exhibits an elevated gravel content. Although the sedimentological details of the two units are not identical, both could have been produced by fluvial processes and the differences between the two may be a reflection of lateral facies variation.

  • The increase in calcareousness in the units overlying conglomeratic VIc and horizon (c) in subunit Vb may provide grounds for a chemostratigraphic correlation. In Sequence 1, subunit VIb is mildly calcareous, whilst VIa at the very top of the preserved section is strongly calcareous (Table 2.1). A similar transition is seen towards the very top of subunit Vb in Sequence 2.

According to Huseinov (1985) paleomagnetic work on the sediments infilling Azokh 1 indicated that the bulk of the middle and upper part of the stratigraphy lies within the Brunhes Polarity Chron (i.e. dating back to 0.781 Ma). Huseinov (1985) noted though, that his “Layer VIII” (very broadly equivalent to the middle of Sediment Sequence 1, as defined herein) was reversely magnetized, suggesting possible placement within the Matuyama Polarity Chron. Ljubin and Bosinski (1995) also noted this possible magnetic reversal in the lower part of the succession. If this is indeed correct, it would imply that the very basal part of the stratigraphy of Azokh 1 is Early Pleistocene in age.

Discussion on the Stratigraphy of Azokh 1

Depending on the method of lateral correlation employed between the two sediment sequences, a total of between 11.2 and 12 m of stratigraphic infill can be accounted for in Azokh 1 passage. Much of this sediment has been removed by previous excavations (Fig. 2.4) and the lack of rigorous recording of this material compromises the information potential of the stratigraphic remnant described here. A graphic illustration of this is the confusion over the precise level within Unit V of the find of the partial Middle Pleistocene human mandible (see discussion in Murray et al. 2010 and references therein).

Lioubine (2002) noted that the “stepped back” appearance of the excavation in the passage (Fig. 2.4) severely hinders any potential study of paleoclimatic proxies, which are generally best preserved in the sediments close to the cave entrance. A similar argument can be made for evidence for human occupation and activity, which is usually best preserved near entranceways in cave settings. Uncertainty over the lateral connection of strata through the cave passage has already been discussed here. Lioubine (2002, p. 23) noted, for example, that Unit V apparently thinned dramatically from 5 to 2 m. This was based on a review of previous reports on the stratigraphy and, admittedly, more precise details were not available to him.

The distinction between the largely unfossiliferous Sediment Sequence 1 and fossiliferous Sediment Sequence 2 above is not easy to explain. It may be a taphonomic artifact; a result of the limited amount of stratigraphy remaining in Sediment Sequence 1 or it may simply be a function of accessibility, with the lower level of the passage (see “basal trench at cave entrance” in Fig. 2.4) not as easy to enter at the time it was originally infilling with sediment. Murray et al. (2010) also highlighted this fossiliferous distinction between the two sequences and tentatively suggested that this may reflect the degree to which the cave passage was open to the outside world. Conglomeratic subunit VIc (Fig. 2.9) is located at the top of this apparently unfossiliferous sequence. It is a particularly distinctive horizon that contrasts with the largely fine-grained units below and directly above. Sedimentologically, it represents a marked increase in the strength of water flow through the passage at this point and it may possibly be related to improved accessibility of the passage (discussed previously herein; see also commentary in Murray et al. 2010).

Sediment Sequence 2 dates from the Middle Pleistocene to the present (Table 2.1; see also Appendix of this volume); although the disconformable relationship between the top of Unit II and base of Unit I means the actual Pleistocene-Holocene transition is not represented. A rich and diverse Pleistocene fauna has been recovered from Units II–V and preliminary findings are listed in Fernández-Jalvo et al. (2010). Cave bears dominate the macro-mammal fraction (Van der Made et al. 2016), whilst bats are a common constituent of the micro-mammal component of the fauna (Sevilla 2016).

The sedimentological differences between the various units infilling Azokh 1 may reflect individual episodes of deposition and sedimentation in response to karst development and paleoenvironmental change, as opposed to a gradual evolution of the entire sedimentary sequence. Much of the sediment in Sequence 2 is quite fine-grained (see Table 2.1) suggesting generally low levels of depositional energy. However, two levels within this portion of the cave-fill (Units IV and II) contain elevated amounts of relatively coarse, angular, limestone debris (Fig. 2.6). Murray et al. (2010 [p. 87] and references therein) suggested that this could represent frost action during cooler climatic intervals but cautioned that other geomorphological processes, such as seismic activity, dissolution and hydration shattering, may produce similar results. These authors also noted that the slope of the various chambers and passages comprising the Azokh Cave system (Fig. 2.2) suggested water and fine-sediment flow from the interior towards the exterior. However; it is entirely possible that the patterns of sedimentation varied throughout the cave’s history, with alternation between the two flow directions occurring. As noted in the opening paragraphs of this section, due to the fact that lateral facies changes were undocumented during the original excavation phase and the fact that the sediment is now removed, this will have to remain a point of conjecture.

Azokh 2

The entrance to the passage we have named Azokh 2 (Fig. 2.3b) is located approximately 42 m NNW from the Azokh 1 entrance (Fig. 2.2). The present level of archaeological excavation has resulted in a chamber that is accessible for about 7.5 m (length) by 3.5 m wide (Fig. 2.15), while the unexcavated level of sediment within the chamber begins approximately 2 m below the roof of the passage. A large boulder collapse has choked the rear, or northeastern end, of the passage where it leads into the interior of the cave (Fig. 2.15; see also Fig. 2.2). This blockage has been a hindrance to further exploration and excavation work within this passage. Azokh 2 is sunlit during daylight hours.

Fig. 2.15
figure 15

Sketch cross-section through the Azokh 2 chamber (drawn facing SE). The locations of various trenches developed over several field-seasons (2002, 2003 and 2007) are indicated. Note that the SE walls of these trenches are out of the plane of section and are thus unornamented. Note also the original level of cave fill prior to the current phase of excavation and the boulder filled inner chamber in the rear of the cave passage

Full size image

Two geological test trenches (see Fig. 2.15) were dug in 2002 [Pit 1] and 2003 [Pit 2] to begin investigating the stratigraphy of the sedimentary infill of Azokh 2. This preliminary work sub-divided the sediments into two stratigraphic units which are readily distinguishable on the basis of color. The lower Unit 2 is light yellow-brown in color, which contrasts sharply with the dark greyish-brown appearance of Unit 1 above. Details of the findings of that work are summarized in Table 2.3. Measurement of the elevation of the contact between the two units (below the level of the cave datum) exposed in the two test pits suggested a possible slope of c. 10° towards the southwest (i.e. towards the cave exit; Fig. 2.15). This slope is less apparent when examined in detail near the entrance to the passage.

Table 2.3 Stratigraphy of Azokh 2 Passage
Full size table

Unit 2 is at least 90 cm thick (the base was not seen) and can be divided into a lower subunit with rare limestone clasts (2b) and an upper subunit with an increased proportion of limestone clasts (2a). Unit 1 is considerably more complex and heterogeneous in character. A subtle shift in granular structure of the sediment effectively marks the distinction between its two constituent subunits (Table 2.3), although this is not always apparent when traced laterally. Unit 1 contains numerous hearth layers, particularly in subunit 1b. These are commonly white to light grey and ashy in appearance, with an associated reddening of the surrounding sediment.

In 2007 the two test excavation pits in Azokh 2 were reopened in order to excavate the intervening sediment section (see details on Fig. 2.16) and during this work, modern human postcranial skeletal remains were discovered. This particular find was reported by Fernández-Jalvo et al. (2010), who noted an age estimate of 1265 ± 23 14C BP.

Fig. 2.16
figure 16

Sketch cross-section through the Azokh 5 chamber (drawn facing NW). The locations of Pit 3 and the in-situ sediments exposed in the inner chamber are indicated. The NW wall of Pit 3 is out of the plane of section and is thus unornamented. Note the original level of cave fill prior to the current phase of excavation. Three consecutive cross-sections (X, Y and Z) illustrating the profile and shape of the entrance tunnel are presented. These are drawn to the same scale as the main section. These sections were drawn before any systematic excavation took place in Azokh 5, and consequently the level of sediment fill corresponds with the dashed “original level of cave fill” line. For key to ornaments see Fig. 2.15

Full size image

Discussion on the Stratigraphy of Azokh 2

Knowledge of the stratigraphy of Azokh 2 is still very much at a preliminary stage. This is largely a function of the boulder choke at the rear of the passage (Fig. 2.15) that has imposed a physical restriction on the direction and degree of excavation work possible. The possible distinction between the two sedimentary units may possibly reflect the passage shifting from a closed (Unit 2) to a more open (Unit 1) system. The increase in the abundance of limestone clasts between subunits 2b and 2a (above) may represent a gradual shift towards wetter and/or cooler climatic conditions, but this remains debatable (see discussion on coarse limestone debris in sediments in Azokh 1 above).

The overlying sediments of Unit 1 are conspicuously more humic and organic rich. They are noticeably heterogeneous when tracked laterally and the presence of numerous hearth layers attest to a sustained period of past human activity in the passage. The first hearth layer appears directly on the contact between Units 1 and 2, which may possibly support the contention that humans only entered the passage when it became open to the outside world. The age of the underlying Unit 2 is uncertain since no artifacts or fossil remains have been recovered from this unit.

Azokh 5

The entrance to the Azokh 5 passage is located about 100 m NNW from Azokh 1 and it connects to one of the largest inner chambers of the Azokh Cave system (Chamber IV; see Fig. 2.2 and also Domínguez-Alonso et al. 2016). The roof of the entrance passageway bears morphological characteristics suggestive of formation as a phreatic tube (Fig. 2.3c). This passage continues northeastwards from the entrance for 5 m before rapidly opening upwards and outwards, creating an inner sub-chamber (Fig. 2.16) which has a significant sediment infill. Chert development is quite prominent in the limestone forming the roof to this connecting passage.

Excavation and investigation of the stratigraphy in Azokh 5 has been conducted intermittently since 2006. The in situ sediments were covered by a ramp of mixed sediments containing, amongst other things, human teeth and tools, as well as additional fauna (see Fig. 2.16 for an indication of the original level of cave fill). A 145 cm deep trench [Pit 3] in the entranceway to the chamber (Fig. 2.17; see also Fig. 2.16 for general location) revealed the following stratigraphy:

Fig. 2.17
figure 17

Photograph of the geological trench (Pit 3; see Fig. 2.16) excavated in the entranceway to Azokh 5. Two units (1 and 2) are clearly visible. The northeastern edge of the trench is indicated with a dashed black line. Hammer (circled) for scale

Full size image
  1. 1.

    [Top]: 45 cm (minimum) very weakly calcareous medium-brown clay-rich soil (humus). A very strong granular structure was developed and many modern plant rootlets were present throughout. Angular limestone clasts (on all scales) were dispersed throughout, and gastropod shells and bat bones were also present.

  2. 2.

    [Middle]: 62–70 cm light beige-yellow firm calcareous clay, with a noticeable carbonate sand and granule component. Scattered limestone and (angular) chert clasts (generally 2–7 cm) were present; however, they were less abundant in comparison to unit 1 above. A moderate to strong granular texture was developed.

  3. 3.

    [Base]: 30 cm (minimum – base not seen) medium reddish-brown calcareous silty clay. It was reasonably well sorted with a moderate to fine granular texture.

Currently, five sedimentary units (labeled A–E) have been identified in the inner chamber (see Table 2.4 and also Figs. 2.16 and 2.18), comprising a predominantly fine-grained succession that is punctuated by two horizons containing elevated amounts of coarser clasts (Units B and D, see Table 2.4, Fig. 2.18). The larger clasts in these layers comprise limestone and chert debris, some of which appears to have simply dropped from the roof above. The significance of these two horizons with elevated amounts of coarse angular cave-wall debris may be that they indicate a shift towards wetter or cooler conditions, but (as discussed above for both Azokh 1 and 2) this line of reasoning is somewhat speculative. The stratigraphic horizons which directly follow Units B and D (Units A and C respectively) both drape and infill their irregular top surfaces.

Table 2.4 Stratigraphy of Azokh 5 inner chamber
Full size table
Fig. 2.18
figure 18

Photomontage of the sediment section present inside Azokh 5 inner chamber. The five units (A–E) are marked in the photo and the tape measure scale is extended to 318 cm

Full size image

The contact between Units A and B (at the top of the inner chamber sequence in Azokh 5) is both conspicuous and significant. The top surface of Unit B is quite irregular and rough and is progressively infilled by the laminated fine clays and silts of Unit A. Thin seams of fine to medium sand-grade material also occasionally occur in the latter. Unit A represents a switch to calcareous sediment deposition within Azokh 5 (Table 2.4).

The sedimentary laminations of Unit A are inclined and appear to have banked up in this corner of the chamber, effectively onlapping the top of Unit B. This may have introduced a time gap of unknown duration into the sequence. Fernández-Jalvo et al. (2010) reported modern human remains (teeth and cranial fragments) with associated charcoal from Unit A. The latter provided a radiocarbon age of c. 2.3 14C kBP. These authors suggest a possible relationship between this material and finds from near the base of Unit I in Azokh 1. It is interesting therefore that both Unit A (Azokh 5) and Unit I (Azokh 1) have discordant relationships with the units they succeed and both infill irregular topographic surfaces of their respective underlying strata. However, despite these general similarities, and in the absence of criteria allowing direct and precise correlation, any broad correlation in terms of the infilling history of the two passageways remains entirely hypothetical.

Unit B thins to 34 cm in the northwestern corner of the chamber and is covered and cemented on the upper surface by flowstone. A mineralised cylindrical structure, possibly a calcified mammal burrow (Fig. 2.19), occurs just below this level. Flowstone has also been observed in association with Units A and D, suggesting that this portion of the cave has remained wet for some time.

Fig. 2.19
figure 19

Flattened cylindrical structure, possibly a calcified mammal burrow, from the inner chamber of Azokh 5

Full size image

The sediments of Unit C are clay-rich and display weak internal stratification. Extrapolation across the cave-chamber, southwest towards the exit, indicates that when this horizon was being deposited the passage connecting to the outside may have been sealed. Unit C most likely was produced by very quiet conditions (still-water deposition), and, given its thickness of 26–34 cm, it may have taken an appreciable amount of time to form. Therefore, although Unit A may be (perhaps late) Holocene in age, with the discordant relationship with underlying Unit B and the potential amount of time required to deposit Unit C, it is possible that Units D and E below may be significantly older.

Discussion on the Stratigraphy of Azokh 5

At present, it is unclear precisely how the two sedimentary sections (external and internal) in Azokh 5 physically and temporally relate to one another. A simple topographic assessment suggests that Units A–E in the inner chamber may overlie Units 2–3 located closer to the cave entrance (see Fig. 2.16); however, this requires more excavation work to unequivocally confirm or disprove this relationship. A further complication is that natural fracturing and displacement of Units A–E has been observed in the inner chamber. It appears that these units have begun to slip progressively downwards in the direction of the cave exit (Fig. 2.16). In addition, Domínguez-Alonso et al. (2016) note that geophysical investigations have indicated over 10 m (stratigraphic thickness) of sediments infilling Azokh 5. Thus, further investigation of the passage is likely required to improve our understanding of the sediments within Azokh 5.

Conclusions

  1. 1.

    Of the three passages connecting to the interior of the Azokh Cave system, the stratigraphy of Azokh 1, previously documented in detail by Murray et al. (2010), is the most completely known. This particular passage has been excavated since the 1960s and contains an 11–12 m thick sedimentary record dating from at least the Middle Pleistocene (and possibly even older; see Table 2.1) to the present. The Pleistocene-Holocene transition is not seen due to a marked erosional disconformity in the sequence towards the very top.

  2. 2.

    A lower-lying sub-chamber in the Azokh 1 Passage (see Fig. 2.4), close to the entrance, accommodates Sediment Sequence 1. This 4.5 m thick section includes Units IX to VI (in ascending stratigraphic order), and, with the exception of the very top, it is apparently largely unfossiliferous. For this reason, the precise age of these sediments remains unknown. Previous paleomagnetic work suggested that the base of the succession, in this part of the cave, might in fact be Early Pleistocene in age.

  3. 3.

    Sediment Sequence 2 is located further in from the cave entrance in Azokh 1 and is interpreted to have largely overlain Sediment Sequence 1, although this is not possible to verify as they are no longer physically connected due to past excavation work. This sequence is about 8.5 m thick and is divisible into five units (V–I). The lowermost Unit V accounts for almost half this thickness estimate. Units V–II have produced a rich and diverse Middle to Upper Pleistocene fauna. Associated and isolated cave bear skeletal and dental elements are particularly conspicuous throughout this part of the succession. Evidence of human activity (in the form of stone tools and cut marks on bones) has also been found in these levels. In the past, a Middle Pleistocene human mandible fragment was recovered from about the level of Unit V, although the precise datum of this find is unclear.

  4. 4.

    The sedimentary infill of the Azokh 1 passage is generally fine-grained, suggesting either very low energy water-flow, perhaps due to ponding as a result of flooding further inside the cave system, or due to possible wind-blown deposition, although this is unlikely for sediments located further inside from the cave entrance. Horizons containing concentrations of coarser limestone debris also occur. Their significance is unclear; however, they may indicate a change in paleoenvironmental conditions, such as an increase in water percolation through the cave or a marked climatic cooling. Geomorphological and tectonic factors, such as an increase in earthquake activity, cannot be discounted either.

  5. 5.

    Azokh 2 is a smaller cave located 42 m NNW from the entrance to Azokh 1. At present two stratigraphic units are clearly identifiable infilling the passage. The uppermost of these (Unit 1) appears to be Holocene in age and below this, Unit 2 is conspicuously lighter in color and more calcareous. Its total thickness is unproven and, as it did not produce any fossils or artifacts, its age is unknown. A significant boulder collapse in the rear of Azokh 2 continues to pose serious logistical problems for further excavation of the cave passage.

  6. 6.

    Azokh 5 is located 100 m NNW from Azokh 1. It is a small phreatic tube that leads to an inner chamber containing at least 4.5 m of infill, although that value is likely to be a gross underestimate as the base and top of the sequence were not seen and geophysical results reported by Domínguez-Alonso et al. (2016) suggests a total sediment infill of at least 10 m. At present five units (labeled A–E in descending stratigraphic order) have been identified. The sediments are generally fine-grained, although Units D and B both contain elevated amounts of coarse limestone and chert debris, much of which has been locally sourced. A disconformable relationship has been identified between Units B and A at the very top of the succession. Unit A has produced charcoal dating to c. 2,300 years BP (Fernández-Jalvo et al. 2010); however, the age of the units beneath remains unknown at present.

Spectroscopy Methodology

Raman Spectroscopy of Azokh 1 Sediment Samples: Analytical Methodology

Raman spectroscopy of sediment samples was conducted at the School of Natural Sciences, NUI Galway, using a Horiba LabRam HR laser Raman spectrometer. The instrument is equipped with a 600 groove.mm−1 diffraction grating, confocal optics and a Peltier-cooled CCD detector (255 × 1024 pixel array at −67 °C) coupled to an Olympus BX51 microscope. Dry, friable samples were placed on a glass slide and analysed in 180° backscatter mode using either 532 nm or 784 nm laser excitation channeled through a 50× microscope objective. Individual analyses were performed for between 60–90 s over the spectral range 100–1800 cm−1 (Fig. 2.14). The number of spectral accumulations per analysis typically ranged between 50 and 100 in order to maximize the signal-to-noise efficiency of the spectrometer. Calibration of the instrument was routinely performed between analyses using the Raman peak of a crystalline silicon wafer (520.2 ± 0.5 cm−1; Parker et al. 1967). Spectral uncertainty associated with the generation of Raman peak positions is estimated to be ±1.5 cm−1 (2σ) under 532 nm laser excitation and ±1.0 cm−1 (2σ) using the 783 nm laser.