4. Stratigraphic Distribution of Archaeological Evidence from Pin Hole Cave

4.1 Objective

The main objective of this section is to describe the archaeological remains from Pin Hole Cave and to reconstruct their stratigraphic distribution within the cave by use of the recording methods (identified in 1984). Archaeological assemblages recognised on the basis of their stratigraphic distribution are then comprehensively described.

Armstrong (1929a) noted extensive evidence of human occupation from the cave. The earlier excavations by Mello recorded an unknown but low quantity of artefacts (Mello 1875, 63). From Armstrong's excavations, there are also concentrations of charcoal, preserved in collections today, which he considered to be hearths and bone associated with lithic assemblages. Armstrong has noted that in his view, there are three discrete Middle Palaeolithic and two Upper Palaeolithic assemblages (Armstrong 1936; also Figure 6). Each of these is stratigraphically separate. Jacobi et al. (1998) questioned the studies of Campbell (1977, 48), who recognised two separate Upper Palaeolithic assemblages within the upper stratigraphic levels that include one of Armstrong's identified Middle Palaeolithic or Mousterian assemblages. In his study, Campbell noted a number of typological overlaps. It appears that both Jacobi and Campbell have compared the assemblages according to horizontal stratigraphic levels and have paid little attention to the effect of a gently sloping datum and sediment infill, the existence of which is corroborated by the nature of the surviving sediment infill. This led Armstrong to postulate the existence of three Mousterian assemblages. This is a remarkable observation, which was later to attract much criticism, particularly by Jacobi et al. (1998) who discounted Armstrong's observations and failed to understand the cave stratigraphy.

Studies by the author have demonstrated that Armstrong's excavation and views are credible based upon diagrams produced by him which clearly show the relevant layers. Copies of these are housed in Oxford, Sheffield and an archive housed in the Creswell Crags Centre itself. The plan has been incorporated within Figure 6. Additionally, excavation by the author commencing in 1981 cleared the large disturbed areas of the cave rear and located a constructed wall of limestone slabs and clast derived from Armstrong's work which also contained faual remains from his excavation. Many authors have also ignored that Dorothy Garrod kept notes concerning the slab layers and reported her views to the British Association of Science regulaly during the 1930s (Unpublished correspondence, Sheffield Museum and Creswell Crags Centre). Reconstruction of archaeological and palaeontological remains presented in Jenkinson (1984) and here clearly shows areas sterile of archaeological and biological remains which separate archaeological distributions. The occurrence coincides with the slab layers suggested by Amstrong.

All artefactual evidence within this study is itemised and summarised by stratigraphic layer in the digital archive. The stratigraphic distribution of artefacts within the cave is of a different nature when replotted according to their distance from the cave entrance and their depth below the calrete, capping, and floor. The results are shown in Figure 7.

4.2 Reconstruction, temporal and spatial distribution

The area of Armstrong's excavation has been divided into a series of 30×30cm units of varying depth along the inclined gradient of cave sediments. The location of individual artefacts follows the annotations marked upon each object by Armstrong. In some instances, the excavation record has no lateral detail and may be simply marked with a depth and to a general area. For instance, some objects are marked '12.P' indicating a depth of 12 feet (3.66m) within the passage area. Such objects are marked by horizontal lines indicating their maximum distribution. The angularity reflected in the final distribution is an artefact of re-plotting the location of remains within the reconstructed one foot cubes (30×30cm) which were the basic units of Armstrong's excavations. The overall effect is exaggerated by the sloping deposit. Despite the crude accuracy of the re-plotting, sterile areas between occupations in the locations where Armstrong suggested slab layers or concentrations of limestone clast are clear. Later studies (Jacobi et al. 1998) have utilised the same mapped calcrete capping deposit but have concentrated on lithics that have associated three-dimensional information and which, despite the exclusion of some lithics, results in a distribution of a similar nature.

The reconstructed distribution indicates the stratigraphic occurrence of lithic assemblages within the cave and indicate that there are four stratigraphically separated areas of distribution inclined along the contemporary and sloping cave floor. Large amounts of material occur within the area, referred to as the passage by Armstrong, and become more numerous and extensive in the uppermost part of the sequence. The importance of this finding is that later studies, often based upon lithic typology (Campbell 1977; Currant and Jacobi 2001), have been severely critical of the validity of Armstrong's original observation. These studies also suggest that Armstrong's comments concerning concentrations of limestone clast, referred to by him as slab layers within the cave passage and rear area, are accurate observations. The temporal disruption of lithic distribution occurs precisely at these depths. It is also clear that the contemporary cave floor has been undulating in form and that the lithic distribution aligns with this, and the major restrictions to access (caused by the downward projection of the cave roof) appears to be a significant factor controlling use of the cave rear. Evidence of limestone clast and slab layers within the cave rear, and which were reported by Armstrong, indicate extensive erosion of the cave roof within this area, which in places reaches the modern limestone surface and must have been inaccessible at times or at least inhibited use by humans. The nature of the reconstruction and the record of both two and three dimensionally recorded lithics inhibits specific comment upon the association of some objects. This is the case in the main areas of lithic distribution, which also contained limestone clast and blocks, which were of similar or greater size than individual excavated units, and is of particular significance during the Middle Palaeolithic when Neanderthal groups would have been confronted by a rock-strewn cave floor within the narrowest part of the cave and not by a relatively smooth surface grading upwards into the cave rear. This area would also have probably been one of the driest and well-drained parts of the cave. The prominent calcrete formation within the entrance area was also encountered by both Mello (1875) and Armstrong (1925b) and is also evident within the uppermost lithic distribution, which appears to be literally wrapped around it.

Excavation by Magens Mello within the cave entrance noted that much of the vertebrate bone was cemented within calcrete and appeared to be orientated by water flow (Mello 1875). The area of excavation extended from the talus into a prominent calcrete feature known as the 'Pin Hole', which is immediately below a significant rupture within the cave roof and which extends to the limestone surface. There is little doubt that water flow from this source has had a significant but localised effect upon the infilling sediments. The prominent calcrete feature, which is evident today, also seems to have acted as a trap to block the outflow of both sediment and water from the cave interior.

Leslie Armstrong published a number of comments relevant to his excavations within the Pin Hole Cave but many are of a very specific nature. In 1927, Dorothy Garrod, who was working with Armstrong at that time, wrote a series of notes intended to be presented to the Derbyshire Caves Committee of the British Association for the Advancement of Science. A very faded original of this survives within the archive of Sheffield City Museum (Garrod 1927). It was transcribed in 1979 by the author and, given its importance, is reproduced in the digital archive. The document is intended to report work at the Pin Hole Cave and several other Derbyshire caves. In the context of the following discussion, the four main points described within the report are of considerable significance and are therefore summarised as follows:

• The area from 60 foot north (18m) was excavated to a depth of 15'6" (4.8m) where the cave floor was a narrowing rock fissure. The fissure contained a deposit of red sand which was thought to be waterlain. This is the only reference (by Garrod and Armstrong) that this deposit existed. The deposit was sterile of archaeological and palaeontological remains.
  This or a similar deposit has been encountered in four other adjacent caves and in one instance within the western chamber of Robin Hood's Cave, and is below a stalagmite level dated to 164k. On this basis, it is assumed that this deposit pre-dates the lithic assemblages and is of Last Interglacial age. The molar of a Hippopotamus identified within the cave is derived from this general area at a depth of 10 foot (3m), (see 11.1). This species is diagnostic of the last interglacial and confirms that sediments of this age occur or occurred in the base of the passage.
• Garrod clearly states that the upper cave earth was sealed by a layer of stalagmite across the cave floor. Adjacent to the eastern wall there were disturbances within the floor (within the passage area). This disturbance is above the area of alcoves into the eastern cave wall, which occur at greater depth.
• Garrod notes that there is multiple evidence of hearths from areas of the cave often associated with burnt and split bones and particularly of Woolly Rhinoceros. Garrod specifically notes an association between Proto-Solutrian and Aurignacian type lithics and the hearths, Woolly Rhinoceros and burnt split bone fragments. This must surely relate directly to the invasively retouched pieces known from the cave. These are discussed within a context of the Middle Palaeolithic by Dorothy Garrod (Armstrong's Mousterian 3?).
• The report contains a very clear statement that the lithic assemblages were clearly separated by two layers of Magnesian Limestone blocks across the cave floor. The lower blocks were cemented with breccia and stalagmite in places. Vertebrate bone fragments and 'Irish Elk' bones were found associated in one layer. This is a direct reference to the 'slab layers' report by Armstrong.
  It is clear that Garrod and Armstrong thought that three middle assemblages existed within the cave and that these occurred with the passage area and were separated by layers of large limestone blocks. Later researchers have discounted the ideas noted by these two very accomplished historians. In the discussion that follows, evidence is described that supports the comments made by Armstrong and Garrod.

4.3 Reconstruction of the temporal and spatial distribution of artefacts and clast within the Pin Hole sediments

Since Armstrong's excavation, there has been constant comment upon the distribution of archaeological material from this cave (Campbell 1977; Jenkinson 1984; Jacobi et al. 1998; White and Pettitt 2011). Armstrong recorded sufficient dimensional information on the archaeological deposits to allow reconstruction of a temporal and spatial distribution within the cave sediments. This is illustrated in the digital archive (Reconstruction Lithic and Clasts) / (Reconstruction with Hearth). The reconstruction shows the distribution of:

• Charcoal samples recovered by Armstrong - During the 1977 survey, the author was able to locate six samples stored in tobacco tins together with micro-vertebrates. Five samples were found and three had been labelled, indicating their origin within the cave. Archives also described three further samples, now destroyed. The charcoal was examined by Woodhead in 1926 (Jenkinson 1984) with the following result - Level 10 foot (3m) charcoal which included Oak (Quercus and Pine (Pinus sp), Level 5-6 (1.5-1.8m) Oak (Quercus sp), and level 1-2 foot (0.3-0.60m) Oak (Quercus), Pine (Pinus sp) and Birch (Betula sp).
• Stalagmite fragments - Most of the stalagmite fragments recovered by Armstrong were from levels underneath the roof rupture of the 'inner chamber'. A total of 34 fragments were located and 21 of these have sufficient dimensional details to be incorporated in the reconstruction. Further specimens may survive within unsorted material.
• Limestone clast - Small Magnesian Limestone clasts were recorded by Armstrong. These were selected from the many that were incorporated within the deposits. Armstrong did not provide a specific reason for retaining these but various comments suggest he may have thought some pieces were used by human groups. A total of 32 pieces survive, 22% of them are incorporated into the reconstruction.
• Natural quartzite pebbles and flakes - Natural quartzite pebbles or flakes were recovered from the length and breadth of the cave. Out of 108, 23 have sufficient dimensional details to be incorporated into the reconstruction. They have been grouped as 'natural' on the basis of either their small size or in the case of flakes, a lack of evidence of fracture to suggest human fabrication.
• Chert artefacts - Out of 11 excavated chert artefacts, 3 have sufficient dimensional information to be incorporated in the reconstruction.
• Quartzite artefacts - Quartzite artefacts have been grouped according to clear criteria indicating fabrication by human groups. A total of 23 were selected, with 10 labelled sufficiently to be incorporated within the reconstruction.
• Flint artefacts - A total of 187 flint lithics were located with 146 having sufficient dimensional details for incorporation in the reconstruction exist: 37 from the 'Top 2' (2 foot), 1 from Level 5, 2 from level 12 of the passage and 1 from 'the base of the hearth'.
• Bone tools and engravings - There were 23 bone artefacts, including worked bone and engraved pieces, with six of these with sufficient dimensional detail to incorporate into the reconstruction.
• Ivory - Several pieces of ivory are known from the cave and three of these seem to have sign of human workmanship. All three have sufficient dimensional details to be incorporated within the reconstruction.
• Non-local fossils - Fossils which are not local to the Magnesian Limestone include an echinoid and marine shell fragments. Two pieces are known and both are incorporated into the reconstruction.

In each case, the number of items utilised in the reconstruction is shown along with the total number of those known from the cave. The importance of this reconstruction is that it illustrates the context of each category of object without preselection of remains and as a result localised contextual variation is easier to assess.

The initial reconstruction of archaeological material (Jenkinson 1984) summarised groups of objects by archaeological 'spit' to show the maximum distribution of several assemblages that included both Middle and Upper Palaeolithic and with a collection of unsorted or categorised artefacts in the uppermost levels (Figure 7). The more comprehensive reconstruction produced for this report offers considerable comment upon the total context and nature of the artefacts assemblages, and particularly that of quartzite objects. The inclusion of further contextual material also renders several structural features more obvious than the 1984 reconstruction. Important features include:

1. The distribution of the two Middle Palaeolithic assemblages replicates that published in 1984 including their distribution, which appears to follow the south-facing downslope of sediment. They are both described in terms of distribution with 'sterile' or no artefact areas around them.
2. The Upper Palaeolithic assemblages and overlying artefacts of younger age exhibit a more horizontal distribution within the cave with the exception of the extreme rear.
3. There is an interface or hiatus, seen in stratigraphic level 5, between the assemblage of Middle Palaeolithic type and the Upper Palaeolithic assemblages. This is indicated by an unusual distribution of lithic and clast which is concentrated and extensive within less than the dimensions of a 1 foot (0.3m) spit. This level also contains several flint artefacts that have rolled surfaces. The hiatus may in part be the result of low-energy water passage downslope and to the south. The feature has been significantly highlighted by disturbance.
4. Leslie Armstrong records a hearth at 2-3ft (0.6-0.9m) in the area 35ft (10m) north of the entrance. The construction of the charcoal, lithic, clast and vertebrate bone within the area suggests a substantial hearth area that may have existed for some time. The area within 33-42ft (9.9m-12.6m) and between depth 1-5ft (0.3-1.5m) is distinct and outlined by the sporadic occurrence of charcoal. The area may be of greater size and extend a further 10m to the north. Lithic and clast items are concentrated in depth level 5 within this area and the horizontal occurrence includes 4 quartzite artefacts and 17 flint artefacts over the 9m length. This contrasts with all surrounding layers.
5. There is also a concentration of vertebrate bone within this area which includes 28 Reindeer body parts consisting of 5 limb, 2 axial and 21 cranial fragments. Additionally there are 26 antlers, 10 of which are cast antler and 5 derived from massacred individuals. This is in complete contrast to the surrounding few metres of sediment. Finally, there are Wild Horse bone remains which include 2 limb, 1 axial and 3 cranial parts and European Bison parts, including 8 limb, 5 axial and no cranial parts. The combined evidence of structural form, clast and lithic concentration and the remains of three species (in a form that hints at sophisticated carcass processing) strongly suggest that this is a concentration centred on a fire or hearth. The presence of Reindeer antler from both massacred animals and cast indicates an activity more complex than cooking. There is at least one bone engraving associated with lithics in the northern slope of this feature.
   In an attempt to date the stratigraphic sequence within the cave, Rowe et al. (1989) obtained age estimates on stalagmite fragments from levels that can now be located to inside this feature. These produced age estimations of 32.5-38 ka. The application of radiometric dating on bone of Mammoth, Reindeer and Hyaena from within this area by Jacobi et al. (1998) produced age estimations of between 34-38Ka. Irrespective of the aim to date these particular levels at the top of the stratigraphic sequence, the dates clearly show they contained older vertebrate bone and stalagmite fragment. The most likely explanation for their presence is that the depression revealed by artefact and bone clustering has been excavated into older sediments. The presence of charcoal suggests that this is a 'prepared' area for a fire-place. This is illustrated in the digital archive. The age of this feature is more problematic. The style of predation revealed by the remains, particularly of Bison and Horse, certainly equates with late Middle Palaeolithic but the feature is more likely to be of the earliest Upper Palaeolithic. The disturbance extends to and exposes quartzite lithics of undoubted Middle Palaeolithic age at the base.
6. A significant concentration is also apparent in the area within 63-70ft (18.9-21m) north from the entrance and within stratigraphic levels 5-9 (1.5-2.7m). Charcoal was noted within this area by Armstrong but has since been lost. The area contains 9 flint lithics, 3 quartzite and 9 natural quartzite pebbles and flakes. The assemblage is of Middle Palaeolithic type. Associated with these are 21 Wild Horse bones (2 axial, 7 limb and 12 cranial) and 49 European Bison bones (19 limb, 13 axial and 17 crania). Reindeer antler includes a total of 24 pieces (11 cast, 4 massacred and 9 fragments associated with 3 axial body parts). This is clearly not a carnivore accumulation but seems to be a small collection of lithics associated with 94 vertebrate bones, which include axial body parts, cast and massacred antlers.

4.4 Early Middle Palaeolithic

The lowest occurring concentration of artefacts within the cave is between 10-12ft (3-3.6m) below the stalagmite datum and 24m north of the entrance. A total of 52 artefacts can be replotted according to Armstrong's handwritten annotations and these are of Middle Palaeolithic type. A further 64 artefacts of similar type lack sufficient stratigraphic information to be sourced to the excavation. The main concentration occurs between 15 to 24m within the cave and on an apparently sloping surface. It is also separated from other artefact assemblages by an archaeologically sterile layer, which occurs at 2.16m below the calcrete layer. Within the rear of the cave (between 19.8 to 24.7m) this sterile layer equates with the highest 'slab layer' reported by Armstrong (1928a).

The stratigraphic occurrence is further separated stratigraphically by another archaeologically sterile layer at a depth of 2.4 to 2.7m below the calcrete and this layer may equate with the highest 'slab layer' reported by Armstrong at a depth of 3.7m below the calcrete. The discrepancy in the depth of its reported occurrence may be due to the use of a reconstructed horizontal datum point. Further information and illustrations of artefacts are available in the digital archive.

4.5 Later Middle Palaeolithic

A distinct assemblage is stratigraphically distributed within levels 11 and 12 and spatially within the cave passage and slope to the 'inner chamber'.

A total of 7 lithics of 'Bunter' pebble quartzite and 15 flint waste flakes were discovered within levels 11 and 12. Their precise spatial location is unknown but it is clear from Armstrong's records that they were discovered within the cave passage. The artefacts of Middle Palaeolithic type were incompletely recorded by Armstrong and the stratigraphic distribution shown in Figure 7 indicates their maximum area of distribution. The artefacts include a biface made from an iron quartzite pebble, a unifacial side scraper or 'chopper' and simply made choppers fabricated by six flake removals from a larger quartzite pebble (Figure 9). Both raw materials are available locally where remnant pockets of 'Bunter' sandstone formations remain. Fifteen pieces of fabrication waste of both flint and quartzite were also discovered within these stratigraphic levels. Armstrong also reports a bone artefact from stratigraphic level 12, which is housed in the Manchester Museum. Examination of the object suggests that the hole is more likely to be the result of carnivore tooth penetration and it is therefore excluded from this account.

Vertebrate bone bearing cut marks were discovered within both stratigraphic levels 11 and 12 and these include the tibias of Mammoth (Mammuthus primigenius) and Woolly Rhinoceros (Coleondonta antiquitatus), which both have localised cut marks on their medial shaft (Jenkinson 1984; see Digital archive V153, V154). Interestingly, the known Middle Palaeolithic artefact assemblage appears not to contain the fine cutting edges that would have been necessary to execute such fine and closely spaced incisions. Both stratigraphic levels also contained two mandible fragments of Woolly Rhinoceros, with indications of human use and an absence of carnivore activity. Despite the very small sample, their presence within levels 11 and 12 indicate Middle Palaeolithic presence, possible fabrication of both quartzite and flint and utilisation of lithics on site in the butchering or preparation of vertebrate carcasses. The specific presence of cut marks in similar locations of selected skeletal parts for two separate species suggest a particular task, which might have involved the preparation of animal pelts (Jenkinson 1984). The relatively crude nature of the excavated record prohibits clear stratigraphic separation of human and Hyaena activity but the absence of gnawing and the mode of bone attrition on three of the pieces suggest that the two activities are separated in time.

4.6 Later Middle Palaeolithic 1

Artefacts of Middle Palaeolithic type were excavated from stratigraphic levels 6 to 9, between 2.4 to 2.7m below the flowstone datum. These are clearly separated by sediments that contained no archaeological objects. In Jenkinson (1984) the Middle Palaeolithic was described as Mousterian I and II, now referred to as an Early and Later Middle Palaeolithic. The suggestion by Armstrong of a third assemblage is likely to be the result of a misunderstanding of an isolated distribution of artefacts. Previous studies (Campbell 1977, 147) have suggested that parts of the assemblage are 'typologically' mixed within the later assemblage. Armstrong's stratigraphic recording, marked upon the relevant pieces, is clear and indicates that these pieces were found at a depth and well within the replotted distribution, the 'mixing' argument is the result of misunderstandings in the chronology of variation in artefact typology. There is also no other evidence of disturbance in deposits in this area of the cave.

The later assemblage is both more numerous and diverse and includes use of both flint and quartzite as a raw material. Quartzite raw material, in pebble form, is prolific locally but flint is less common and this may indicate collection of suitable raw material from the Lincolnshire Wolds outcrops, although it remains possible that flint erratics may have been more frequent locally during the Quaternary.

The lithic assemblage includes two finely retouched scrapers in flint, both of which are similar in style to the more well known 'Quina' scrapers known from European Middle Palaeolithic locations (Figure 10).

Other artefacts which can be replotted include 17 pieces of fabrication waste of which 6 are incomplete flakes, 5 blade fragments and 6 flake fragments. A further incomplete flake of red chert is recorded from stratigraphic level 7. An invasively retouched elongate flake of flint (Figure 10c) was discovered within stratigraphic level 7, well within the Later Middle Palaeolithic assemblage. It has been suggested by several authors (e.g. Campbell 1977) that this is of Early Upper Palaeolithic type and age, a suggestion that is not supported by its very clear stratigraphic position.

Quartzite artefacts include two bifaces from stratigraphic levels 7 and 8, both of which have evidence of edge damage. Scrapers are relatively common and comprise nearly half of the total quartzite assemblage. These are bifacial; one was discovered in the same location to the invasively retouched blade illustrated in Figure 10 and the second was found near to a surface-damaged quartzite pebble, possibly a hammerstone.

There are 64 artefacts of quartzite and flint of Middle Palaeolithic type without a clear provenance. These lack Armstrong's full excavation coordinates or have a simple depth record, below the flowstone datum so cannot be accurately replotted. All of these have been included within the identified Middle Palaeolithic assemblage simply because there is no other indication of quartzite artefacts in the overlying levels of sediment. Fortunately 11 unretouched or incomplete flint artefacts have depth coordinates. 42 quartzite artefacts have insufficient coordinate details to be sure of their location; these include 6 bifaces, 2 bifacial scrapers, 2 unifacial side scrapers, 29 unretouched pieces and 2 cores. There are a further 6 flakes that may have originally been associated with the assemblage.

The total Middle Palaeolithic assemblage is clustered in two separate assemblages both of which occur in the cave passage where they are distributed along a gradient that slopes towards the cave entrance. It has already been noted that this is a very narrow and presumable dark area of the cave but the sediment surfaces were also at their highest in this region.

Fabrication of the assemblage is distinct in that rather simple flaked bifaces are associated with more complex flint forms and where the variety of materials found indicate that they have made within or near the cave. Although the assemblage is 'rich' in Creswell terms, lithic numbers are small in comparison to other European sites. It has been suggested (Mellars 1974) that the typological style is of the well-known Mousterian or Acheulean character and implicit in this suggestion is a western European origin for the technology. The argument has no real foundation in that Levalloisian technology and style is known from large contemporary areas of Europe. The form of this assemblage, which includes sophisticated flint work associated with simply flaked quartzite artefacts, is more likely indicative of attempts to use the more commonly available raw materials, and in any case is more indicative of functionality than cultural contact. It is difficult to imagine Neanderthal groups fabricating and carrying a tool kit weighing several kilograms of readily available material when the evidence suggests the pragmatic and spontaneous use of local materials, some of which show no signs of use but have nevertheless been discarded or lost within the cave. In pragmatic terms, the assemblage is composed of low numbers of scrapers of different types and larger cutting edges or scrapers of quartzite in roughly equal proportion. They do not include examples of projectiles, formal cutting edges normally associated with hunting practice, or awls, denticulates or burins, which may normally indicate secondary processing activity. Also absent is any indication of the utilisation of wood as a raw material, despite excellent conditions of preservation. Wood is likely to have been present in the more sheltered condition of the limestone gorges and which is suggested by the presence of specific woodland and aquatic vertebrate species known from these levels. The available evidence may therefore simply represent the heavier more durable or utilised element of tool kits. This seems to indicate that tool production is specific and not related to local or regional procurement activity but is more indicative of specific on-site processing of resources. This study describes contemporary vertebrate bone from these stratigraphic levels (see Mammoth and Woolly Rhinoceros descriptions in section 6) and in the absence of other evidence for functional activity, suggests Middle Palaeolithic groups may have been preparing and processing vertebrate skins or pelts (an activity that would go some way to explain the nature of the assemblage and why it differs from nearby sites and is particular to this cave). The location is also in stark contrast to other nearby sites, which are larger and appear to be more suitable for habitation.

4.7 Later Middle Palaeolithic 2

This artefact assemblage overlies that of the Later Middle Palaeolithic 1 and is separated from it by between 0.3-0.9m of sediment containing no artefacts. The assemblage was originally described as of Middle Palaeolithic type (Mousterian 3) by Armstrong (1926). This would equate with the depth of occurrence, which is 5ft (1.5m) below the datum and the reconstructed stratigraphic level 5 (Jenkinson 1984). Campbell refers to the assemblage as Early Upper Palaeolithic, primarily due to the typological form of some of the lithics and specifically the delicate invasive retouch which was then thought to be similar to the Solutrian assemblages known from France. Both lithic types have characteristic invasive surface retouch but are quite different in many other ways, which suggests that they are not part of the same facies. A well known 'Leaf Scraper' is illustrated in Figure 11.

The assemblage is horizontally distributed in the passage area and between 10.2-24.7m from the cave entrance. Its vertical distribution varies between 0.77-1.85m. Figure 7 shows the stratigraphic occurrence and extent. Stratigraphic separation of this assemblage from the overlying one is clear cut for an extensive area of the cave, where an average of 30cm divides the two. The situation is less clear locally at several points. A Bronze Age burial, 11.4m from the cave entrance has penetrated both this and Middle Palaeolithic levels. Disturbance is also present at 18.5-20.4m from the entrance, where human bones have been found and which may be part of a post-glacial burial. Artefacts with sufficient stratigraphic information (c.72) and their distribution is shown in Figure 7.

The context of this small group of lithics is of considerable interest and there has been much comment on their relevance. Campbell (1969) has suggested the existence of an Early Upper Palaeolithic within the UK where invasively retouch lithic are characteristic. Several invasively surface-retouched lithics are known from the cave, particularly from its rear, in the region of 66ft (20m) from the entrance and at various depths. The largest number occur within the 'hearth' feature. Given their importance the context is discussed in detail. They were discovered during 1927 and it seems that both Leslie Armstrong and Dorothy Garrod were familiar with the discovery at that time. Other lithics that derive from the same area 30-35ft north of the entrance are within the 'hearth' area. The groups consisted of:

• Large stemmed point (Font Robert point) - found at 35ft (10.5m) north at 5ft 6in.' (1.68m)
• Quartzite hand axe - found at 35ft (10.5m) at 5ft 6in. (1.68m)
• Quartzite, large retouched flake found at 31ft north (9.3m) at 5ft depth (1.5m)
• Dorothy Garrod found an end scraper on blade at 49ft (14.9m) north at 5ft depth (1.68m)

All of the lithics were on or above the 5ft depth (1.68) upper slab layer commented on by both Armstrong and Dorothy Garrod. Garrod (1927) comments that this layer was continuous across the cave floor, with the exception of some areas adjacent to the east wall. The purpose of this detail is to suggest that the apparent mixing cited in many reports is very difficult to explain under these circumstances. Even allowing for movement, the lithic was resting on a solid rock floor that coincidentally happened to be the base of the later disturbance associated with the hearth.

A total of 72 artefacts have sufficient stratigraphic information to replot their distribution, which is shown in Figure 7. Descriptive details of the total assemblage are shown in the archive, which shows that the major concentration is within stratigraphic level 5. Artefacts recorded from the assemblage include a complete invasively retouched 'point' (Figure 9) from stratigraphic level 3 and a broken example from level 4. Stratigaphic level 4 also produced an exceptionally large and finely worked stemmed point (Figure 12a), which was found toward the cave entrance, and an obliquely truncated blade with retouch around the truncation (Figure 12b). Stratigraphic level 5 has produced two points with extensive retouch.

A total of four burins are known, three of which have no retouch but the fourth has been backed along one edge. End scrapers on blades and flakes (Figure 13) are recorded from stratigraphic levels 3-5. The total does not include the 'laurel leaf points' already mentioned and which may in fact be side scrapers. Other artefacts include a large backed blade, a retouched blade and 6 retouched flakes.

The stratigraphic distribution of this assemblage is of considerable interest. In his study of the UK Upper Palaeolithic, John Campbell (1977) grouped 72 of the artefacts into an Early Upper Palaeolithic assemblage. The inclusion of some bone artefacts and flint within this assemblage is the result of a misunderstanding of Armstrong's records and these can be shown to be of later age. Jacobi (1980) has suggested that the Early Upper Palaeolithic can be divided into at least two chronologically different parts. He suggested that an earlier phase is characterised by invasively retouched pieces (leaf points) and a second later phase, which includes Upper Perigordian tool types (the stemmed point in Figure 12). In his study, he correlated these phases with those known from French Palaeolithic sites (Jacobi 1980, 83) and suggested interstadial periods. The argument rests heavily upon consideration of variation in artefact typology and is not supported by the stratigraphic evidence for distribution from this cave. Armstrong's recording methods are not sufficiently refined to demonstrate the possibility of different phases of the Middle Palaeolithic assemblage (or Early Upper Palaeolithic) assemblages, a situation that does not discount the suggestion but the existence of a hiatus between the rear cave and front cave distributions cannot be proved.

In addition to the lithic remains, Armstrong records a concentration of quartzite and Mammoth bone. In the area 29ft (9.7m) from the entrance at a depth of 5ft (1.5m), a green quartzite biface and one worked quartzite pebble were discovered with a Mammoth (Mammuthus primigenius) tusk section, worked at both ends, and a worked but fractured piece of tusk. All of the pieces were found under the eastern wall overhang and thought by Armstrong to be associated. This is the only evidence from the cave of Middle Palaeolithic bone working of Mammoth bone (see Digital archive, items 9, 10 and 11). This is the same stratigraphic level 5 that has produced invasively retouched flint laurel leaf points. Unfortunately the worked Mammoth tusk has not been located during this study. The discovery is unusual and rare evidence of Middle Palaeolithic use of ivory (see below).

Within this general area (32ft from the cave entrance) and within the area of the Middle Palaeolithic assemblage 5ft depth (1.5m), a worked quartzite was discovered associated with a juvenile Mammoth molar and a small fragment of Hyaena mandible but neither bone has any indication of surface cutting or carnivore damage. Later studies of this area within 33-48ft (9.9m-14.4m) from the cave entrance suggest that a prepared 'hearth' existed here, which may have been excavated within the latest Middle Palaeolithic and particularly as the fill within the depression contains vertebrates typical of the sort produced by Neanderthal predation, which is known from other areas of the cave at this depth.

The purpose of this study is to give an account of Armstrong's excavation and to reconstruct, as far as possible, the original location of this assemblage within the cave. It is clear that the excavation indicates:

• The assemblage present in stratigraphic levels 5 and 6 (1.5m-1.8m) contains lithics fabricated in both quartzite and flint. There are several instances where the typological style has been used as evidence that they are of Upper Palaeolithic age.
• There are associations between quartzite tools and vertebrate bone and ivory that occur in circumstances that cannot suggest mixing of remains (rock alcoves within the eastern cave wall).
• Suggestions of disturbance have been based upon the occurrence of variations in the type of artefacts present within the cave. In most cases it cannot be demonstrated that these are stratigraphically separated within stratigraphic levels 5-6 even allowing for the presence of a hearth disturbance in one area. Arguments that they are in reality separate within the cave relies upon suggestions of 'mixing' for which there is no sedimentary evidence, the discounting of apparent inappropriate associations (other lithic and bone working) and radiometric dating where clear association and context cannot be demonstrated other than a two-dimensional location specific to a 0.3m depth and 1.8m distance range. The difficulty is aggravated by the fact that little account is taken of the effects of gradient within the sediment build up, which suggests that stratigraphic level 5 could in fact be stratigraphic level 6 lower down the sediment gradient.
• There is at least one instance of a 'hearth' area where there has been penetration of Middle Palaeolithic levels at a later, but unknown, age. This is confirmed by radiometric dating. It is also suggested by a dramatic occurrence of vertebrate remains which are distinct from other populations known from the cave. This population includes Wild Horse, European Bison and Reindeer, which are preserved in a form that suggest hunting by human groups and not carnivore activity. There is no indication of disruption or discontinuity within the vertebrate assemblage.
• The stratigraphic situation is aggravated by the reported existence of a 'slab layer' reported by Armstrong within stratigraphic level 5. There is enough evidence to know that this existed in some form, and extensive clasts probably derived from this have been recovered from the cave during the 1980s. Armstrong also notes that some lithics were discovered 'under stone'.

The evidence from the cave unfortunately is controversial in the sense that it does not independently support the concept of two separate assemblages (a later Middle Palaeolithic assemblage overlain by an early Upper Palaeolithic assemblage) interleaved between a clear occurrence of Middle Palaeolithic occupation (levels 5?-6) and Upper Palaeolithic occupation (level 4 and above).

4.8 Later Upper Palaeolithic

A total of 152 artefacts are recorded from the Later Upper Palaeolithic of the cave. Armstrong recorded a two-dimensional location detail (distance into the cave from the entrance and depth) on the surface of some of these but large numbers of lithics have no stratigraphic details. The stratigraphic distribution of 137 artefacts has been used to replot distribution in Figure 7. In addition, Armstrong recorded 10 bone artefacts, 3 marine shells, amber, and 2 Jurassic fossils. An additional 30 flint artefacts have no stratigraphic details but are included in this account on the basis of their typology. A significant difficulty in any consideration of the Later Upper Palaeolithic 'assemblage' is the presence of microlithics of a typological form that would normally be considered to be of later Mesolithic age, and many of these have no accompanying stratigraphic details that would allow comment on their actual age and their relationship with the Palaeolithic evidence. This is a shame in a site that is otherwise so data rich, and particularly as archaeological evidence from Mother Grundy's Parlour (Armstrong 1924a) and Thorpe Common Rock Shelter (Jenkinson 1979), both local to the cave, suggest that there is evidence of an archaeological boundary between the two 'types' of assemblages. This is of considerable interest in that it is associated with major environmental change, and the change in character may be of a functional nature rather than a chronological development (see Figure 14). The collection is summarised in the digital archive and the distribution is shown in Figure 7.

Consideration of the assemblage is undertaken in reference to the contents of individual stratigraphic levels and lithics that can be located specifically from Armstrong's coordinates. The approach has been applied to 5 levels (levels 4-0) described below.

Stratigraphic level 4 (i.e. 4ft/1.2m below the stalagmite datum)

Stratigraphic level 4 contains 4 flint waste flakes and a possible bone artefact in the form of a sub-triangular bone flake where both sides have a highly polished surface.

Stratigraphic level 3 (i.e. 3ft/0.9m below the stalagmite datum)

Stratigraphic level 3 contains the largest and most diverse evidence for the assemblage within this cave. This level contains 16 lithics with sufficient information to be replotted. These include a plano-convex point, which is retouched along the oblique truncation and the opposing distal edge, a retouched scraper, a backed blade fragment and a truncated and retouched blade (see Figure 15).

Bone/Ivory artefacts and engravings

Three artefacts are known from this level which include:

• An ivory pendant. The piece is 35mm in length and tapers at one end (Figure 16). The other end has been perforated by a small hole 2mm in diameter and the surface on the shaft outer surface has been polished. The other end is broken across the width of the piece. The break is of considerable interest and may indicate that the owner deliberately discarded the piece after it was damaged.

  

• A further bone artefact is a damaged rod of bone, 41mm in length. The piece is a half cylinder-shaped rod, which has been polished on one side (Figure 17). The lateral surface has two separate intertwined undulating lines incised along the length of the piece. It is not clear if the engraving is complete or truncated by the fracture. The fracture on the left side is vertical and is likely to the result of cutting the bone, while the right side oblique fracture suggests later damage. Both deliberate and accidental end damage is very similar to at least one other piece from these levels (Figure 19) that suggests these artefacts have a similar function and are prone to be damaged in the same way. The purpose of the artefact is unclear; it is skilfully fabricated, and the incision is very finely executed which indicates that both the piece and function were likely to be of some significance. Despite the lack of direct evidence of function, it is possible that the piece may have been part of a projectile and that the incisions along its surface would have aided removal from an animal carcass.

  

• Engraving of a human figure. This substantial artefact is illustrated in Figure 18. It has been formed on a large ungulate rib, and the object retains the original and natural curvature of the rib.

  

  
  At the time of discovery Armstrong (1929b) named the image as that of a 'Masked Human figure'. This has remained controversial and much of the comment has centred upon the reliability of Armstrong's observation rather than the piece itself. The engraving has been executed on the distal part of the rib and the distal section of the image has been partially obscured by later surface polishing, which is the result of use and was not part of the fabrication process. The engraving itself remains difficult to interpret, particularly as the entire surface on both sides of the bone show numerous marks and incisions. In recent years, this has been the subject of a study and restoration by Cook (Cook pers. comm.), who has confirmed that the outline is of ancient manufacture and probably on a piece of fossil bone. Armstrong published the original image, which had been outlined by him (1929b). The image was photographed again in 1978 by the author using oblique light sources and recorded on black-and-white film. The resulting image clearly shows a prolific number of 'marks' concentrated around the outlined image, which suggests that the well-known image may in fact be part of a more complex one and which offers a rather different possibility of interpretation. The issue is not as yet resolved and discussion has concentrated upon the contemporary nature and existence in reality of associated lines (Cook pers. comm.). The suggestion that some lines are the result of root marks caused by vegetation, while a possibility, is problematic as there is no evidence that live vegetation has ever existed within the passage cave environment.
  Putting to one side the discussion and interpretation differences, it is clear that the incised lines are an abstract image of a figure with a form that is more animalistic than humanoid and as such is very unusual piece of evidence for Upper Palaeolithic abstract art.
  The location of discovery is clear from Armstrong's records, which suggest that it was discovered in an area between 18.5-21m from the entrance and within the upper 1.2m of sediment. This area is well away from the main focus of Later Upper Palaeolithic remains. There are no lithics in the immediate vicinity. It was also discovered near the base of the Later Upper Palaeolithic lithic distribution.

Stratigraphic level 2 (i.e. 2ft/0.6m below the stalagmite datum)

Artefact evidence from stratigraphic level 2 is extensive and 44 flints have sufficient stratigraphic detail to be replotted. Both retouched and unretouched blades are an important element of the assemblage and represent 44% of the total. Other tools include an extensively retouched point, three scrapers, two burins and a borer or piercing tool.

Bone artefacts

• Ivory rods
  Artefacts from this level include an ivory rod 37mm in length, a second example 31mm in length and a third which is rectangular in cross-section and 31mm in length. All three rods are fully shaped with some surface polish but have no trace of incisions on the surface.
• Engraved bone rod
  This artefact is illustrated in Figure 19. The rod has an approximate length of 40mm and is probably a rib fragment derived from a small vertebrate species, possibly a carnivore.

  

  
  The bone fragment has been partially prepared by edge and surface polishing and has been engraved elaborately on one surface. The surface engraving consists of nine separate elaborately engraved groups of oblique designs that are diametrically opposed to form a 'cross hatching' effect along the length of the rod. Two groups, one at either end of the rod, are damaged by fracture. The total engraving has been extremely finely executed with considerable precision, which suggests that it was of significance. The form is very similar to the previously described rod (Figure 17) and has been damaged through use in the same general fashion, indicating a common function and use. Both pieces are enigmatic in the sense that their function is not clear. They are obviously intended to be of an artistic nature and display two different sets of abstract concept and are therefore probably decorative.
• Marine shells
  Armstrong discovered two unusual shell fragments. The first is a fragment of mother of pearl, found 13.6m from the cave entrance, where it might have been associated with human skeletal material, and which was immediately in front of a concentration of charcoal that Armstrong interpreted as a hearth. The charcoal fragments survive today with the Manchester Museum collection.
  The second shell is a complete Cypraea Moneta, which was discovered 22.3m from the cave entrance at a point almost at the rear of his excavations. It is difficult to be sure if the shell fragments are 'modern' in the sense of being of Quaternary age or if they are fossilised examples. Their presence indicates that Upper Palaeolithic groups were familiar with the outcrops within the Lincolnshire Wolds and may have similarly been familiar with the then coast.
• Fossils
  At a point 6.1m from the entrance a small Belemnite fragment was discovered in association with Upper Palaeolithic lithics. As in the case of the shell fragments, its presence indicates familiarity with the outcrops of the Lincolnshire Wolds.

Stratigraphic level 1 (30cm below the stalagmite datum)

Artefacts are at their most numerous within this level and a total of 56 have sufficient stratigraphic information to be plotted within the distribution illustrated in Figure 7. The lithics are dominated by retouches and unretouched blades, which comprise 39% of those known from the level. Other elements of the assemblage include backed blades and burins and a truncated 'Cheddar' point from the front of the cave at 8-10.2m. Also from this area there is a bifacial flint knife.

Bone artefacts

• This solidly formed bone fragment has been extensively worked and is vertically truncated on the left side where an elongate section has been removed from one surface to produce a bevel (Figure 20). The right-hand side has been beveled in both surfaces to produce a pointed end that has also been obliquely incised by a series of at least 9 lines on both surfaces. One surface has been finely incised with undulating lines along its surface. The whole piece has been surface polished as part of the manufacturing process. The artefact may be part of a larger piece and the engraved bone found in stratigraphic level 1 might be part of the same object. The two pieces do not actually fit together but the form and decoration are very similar. Campbell has suggest that both pieces are the fractured parts of the same object (Campbell 1977, fig.102).
  It has been suggested by both Armstrong and later by Campbell (1977) that this is a projectile point. The suggestion seems to be very reasonable, particularly as the nature of the left-hand bevel is suitable for hafting and the manufactured right side forms a sharp wedge where the length has been incised in a fashion clearly intended to reduce suction in circumstances where the point has penetrated flesh. The engraved end (right side in Figure 20) has edge damage that may be the result of an impact. The object itself is quite small and is more likely to have been the more durable end of an arrowhead than that of a javelin or spear point and intended for use in the hunting of smaller vertebrate species.
  

  

  
  The possibility that this ivory fragment fits the similar piece from level 3 was suggested by John Campbell (1977) and points to post-mortem movement. A greater difficulty is that they are made of ivory, presumably from Mammoth, a species currently thought to be locally extinct during the Late Glacial. The fact that ivory has been used suggests that the source was an older fragment of ivory that was collected and used as a raw material during the later Palaeolithic.

Stratigraphic level 0 (containing stalagmite datum and the immediately underlying deposits)

The uppermost stratigraphic level contains relatively few lithics (19) and over half of these are retouched and unretouched blades. Two fragments of retouched blades were also discovered.

• Fossils
  A single, well-preserved fossil of the Jurassic echinoid from this level was identified by Eager as Pseudecliadena depressa (Eager pers. Comm).
• Typologically similar artefacts
  The Armstrong collection contains 30 lithics of Upper Palaeolithic type, which do not have any individual stratigraphic coordinates but are stored as a group of lithics from the upper levels of the cave. These are included in the digital archive. The material is dominated by retouched and unretouched blades and includes a penknife point, burins, scrapers and broken retouched blades.
  A second collection contains 18 flints for which there is no stratigraphic information. The collection includes microlithic forms, a point, 3 unretouched blades and 8 broken flint flakes. It is unfortunate that these cannot be located in the lithic distribution and many are likely to be Later Upper Palaeolithic in age. The presence of unstratified microliths is of particular interest as there are only two other Creswell sites (Mother Grundy's Parlour and Thorpe Common Rock Shelter) where evidence for the Mesolithic occurs stratigraphically overlying the Upper Palaeolithic.
  One aspect, which has not been systematically examined within this study, is the reuse of bone and ivory as raw material. There are a number of artefacts made of ivory at a time when Mammoth is generally thought to have been regionally extinct and there is at least one example of an engraved form on an ungulate rib which was fossilised prior to its use as a raw material. Corroborative evidence of interest in such material is also suggested by the deliberate collection of Jurassic fossils and amber. The question is of some interest in that it suggests that human groups were perhaps more environmentally aware than previously supposed.
• Human occupation
  The cave offers extensive evidence of human use at least twice in its early history. The first occupation is known from a few quartzite tools and cut bones during the Middle Palaeolithic, and following a period of sedimentation with no archaeological evidence a second occupation takes place but now represented by significantly more artefacts of quartzite and sophisticated flint tools. These assemblages are separated from and overlain by an enigmatic assemblages of distinctive invasively retouched tools of flint ('laurel leaves') dominated by cutting edges. The uppermost sediment sees the appearance of numerous backed blade tools of much more diverse type and accompanied by bone artefacts and curios.
  Occupation or use of the cave has, surprisingly, been in the same area during the Middle and Early Upper Palaeolithic where remains are concentrated on the sloping bank of deposit within the passage of the cave, its narrowest point, and presumably benefiting from this location's relative dryness and away from the more unstable cave rear. Considerable expansion of the occupied area, toward the cave entrance, occurs during the Later Palaeolithic. Much of the occupation terminates a small distance from the entrance, inside the cave at a point with evidence of increased erosion associated with solution cavities in the cave roof that reach the surface and where there is consistent evidence of water inflow.
  The cave interior, throughout most of prehistory has offered a difficult and restricted area for occupation and the selection of a specific area inside the narrow passage seems to indicate that shelter, while important, may not have been the main motivation for the cave's use. Tools normally associated with the processing of resources rather than hunting kits dominate the relatively restricted range of artefacts. It is possible that one of the main motives in selection of the cave is that, despite its rather difficult nature, the location is the only one within the gorge with an open northern location, that offers a relatively drier and lighter environment and commands a visual access along the limestone gorge and large areas to the west, none of which are available from adjacent sites within the gorge and issues that must have been of some importance in the Quaternary environment.

4.9 Postmortem movement- a dimension of human and vertebrate curiosity

Many comments have been made concerning movement and the possibility of 'mixing' within archaeological assemblages. Similar comment has been made concerning vertebrate bone. This study has presented a case that minimises the effect of mixing and disturbance outside of clearly known and understood examples. It is also clear that the deposition of sediment and the incorporation of artefacts and vertebrate bone is not in circumstances that create horizontal 'slices' of evidence through time. There is multiple and clear evidence of activity by an enormous number of animals and human groups over a long time period and within a small space. Explanation of the taphonomic circumstance is often difficult or impossible in many cases. The purpose of this account is to describe some of these examples, which can occasionally be explained. In other examples, there is a possibility that movement is undertaken deliberately by human or animal agency and motivated by curiosity. Relevant examples are:

• The area 54-56ft (16.2-16.8m) north of the cave entrance contained a Hyaena mandible in two parts, a complete left and right mandible. They are derived from the same individual. They were discovered in separate locations. One mandible was at 54ft (16.2m) north and 4ft (1.2m) below the datum. The second mandible was discovered at 56ft (16.8m) north of the datum and 5ft (1.5m) below the datum. The fitting parts are separated by what may have been as much as 2m. They are separated vertically by 30cm. The most significant difficulty is that they are in sediment that is thought to post-date Hyaena occurrence within the cave and the sediment to the north behind them is of younger age.
• At the location 66ft (19.8m) north of the entrance and within the 2ft layer (0.6m), a complete mandible of Hyaena was unearthed. The mandible is derived from an old individual and both sides are completely fused together at the ephysis. The specimen is also in pristine condition with no indications of damage or rolling. There is nowhere to the north of this area where this object can be derived from, particularly as the horizontal sediment behind it is younger in age. In summary, a mid-Quaternary Hyaena mandible, well-preserved, occurs within areas of an Upper Palaeolithic lithic assemblage of late age. The most obvious suggestion is that this was placed or lost here after collection by a human occupant of the cave.
• The depth and stratigraphic levels referred to contain lithic fabricates in flint and chert derived from regional sources. They also contain Jurassic Belemnite and an Echinoid that must have been from a Lincolnshire rock outcrop. In addition, the layers contain a piece of amber, an ornamental marine shell, and a section of mother of pearl of marine origin. There are also ivory pieces, derived from Mammoth which had disappeared from the region long before collection. These items are quite a comprehensive and direct indication that curiosity by human groups may offer one explanation why unexpected associations between objects have been encountered and which do not indicate movement or mixing by geomorphic event.
• Lastly, there are direct records of associations of objects that are a product of human activity. These are recorded within notes kept by Leslie Armstrong.

Use the navigation bar (top) to go to other sections of this publication