7. Middle Devensian Vertebrates (The MAZ assemblage)

7.1 Objective

Both the concept MAZ (Mammal Assemblage Zone) and the evidence is reviewed to assess the nature of the 'Mammoth Steppe' fauna at this locality.

7.2 Background

Studies by Currant and Jacobi (2001) have developed a useful concept in recognising a series of distinctive mammal assemblages known from British cave localities and which include formal recognition of a Mammoth Steppe type faunal assemblage. Their studies of the Pin Hole Cave vertebrates from levels between 6ft (1.8m) to the cave base at 20ft (6m) have suggested that there is a distinct group dominated by Wild Horse, Woolly Rhinoceros and Mammoth, which have become formally referred to as a Mammal Assemblage Zone (MAZ). The zone is referred to OIS 3 and thought to date between 50-38 ka. The vertebrate population appears to equate well with the suggestions of a 'Mammoth Steppe' biome (Guthrie 1982) and the fluctuating, extreme holoarctic climatic conditions suggested by Coope (1977).

There is no doubt that this is a useful concept but there are a number of difficulties in its application to the Pin Hole vertebrates known from the lower levels ('yellow cave earth'). The material reported in Jenkinson (1984) records 14,978 vertebrate bone fragments, which were identified to 77 species and 7 generic groups. An additional issue is that concerning the identification of Ground Squirrel, which is a rare species within a last glacial context in the UK. In 1977, the author identified three specimens housed in Manchester Museum (see Digital archive). There are difficulties with including these specimens in MAZ and these are discussed in the species description within this report. The circumstances are made more difficult by the very wide stratigraphic separation of each specimen (one is near the surface, possibly in Holocene sediments). Stratigraphic levels 5 and 6 contained 45 Rodent cranial parts identified to five other species and a total of 4840 Rodent post-cranial bones identified to genera, and some of these are without doubt of Holocene age.

Studies by Currant and Jacobi (2001) appear to have been confined to an occurrence of 3093 vertebrate bones derived from large species and which were identified to 14 species. They have therefore based their definition of a mammalian zone within this cave on less than 10% of the known evidence from these levels. There are several problems with this approach; firstly a formal definition of a mammalian zone based on such a low percentage of the evidence may not be sound. Secondly, if the total information had been included, the zone's components would differ from other localities with apparently similar evidence to the zone. Thirdly, the components of the mammalian zone within Pin Hole Cave do not suggest the regional environmental conditions suggested by Guthrie's 'Mammoth Steppe' (1982) or the extreme conditions suggested by Coope (1977).

Even if consideration is given to the possibilities of bioturbation or geomorphic disturbance of the sedimentary infill, it is an unlikely explanation for the possible downward movement of such a large number of vertebrate bones and any such process does not seem to have affected the lithic evidence. It does seem that Armstrong's observations describe a real occurrence and one that is significantly more diverse than the assemblage suggested. The species present also suggest a biome that includes aquatic environments, mudflats, shrub vegetation and seasonal fresh-water sources. The vertebrate faunal evidence suggests:

1. A large population of seasonally migrating ungulates who are presumably grazing on a diverse herbaceous and grass flora. Based upon bone frequency these are dominated by the remains of Reindeer with Woolly Rhinoceros and Wild Horse. A smaller frequency of Mammoth and Bison are associated with this main group.
2. The ungulate population is sufficiently large to support a numerous and diverse carnivore population, dominated by Hyaena and the less frequent occurrence of Red Fox and Brown Bear.
3. Eight species of Rodent occur in some numbers. These include arctic species such as Northern Vole and Tundra Vole but these are associated with more temperate preferring species that are usually associated with the presence of ground cover and the ability to burrow.
4. Bird species are commonly of single individuals and a large component of the population are Geese, Ducks and wading birds that indicate the regional occurrence of open areas of slow-flowing water. Geese egg fragments (153) also indicate that these species are regionally resident in their spring or breeding area.
5. Bird species include 15 species of perching birds that require scrub or open woodland. These are associated with Short Eared Owl, a vole predator, known today from open area habitats.
6. Large quantities of Common Frog and four species of fish, all of which prefer slow-flowing freshwater, occur in virtually all levels.

The vertebrate population has no comparison but is dominated by species usually found in Holarctic environments where they occur in a range of biomes and seasonally migrate. It is likely that the ungulates indicate open and extensive grassland or steppe environments but the occurrence of other species clearly indicate lacustrine and riverine habitats and areas of open woodland. It is very likely that such sheltered areas existed within the limestone gorges and valleys within the Creswell area. In contrast to the impression given by descriptions of the components and biome preferences of the MAZ assemblage, the evidence suggests a more temperate regime that was used by a greater diversity of species during the spring and summer.

The vertebrate fauna discovered within the upper levels ('red cave earth') of Pin Hole has few parallels in diversity and frequency. The fauna from the upper levels is distinct and indicates an episode of great change. The numerous vertebrate bones have been identified to 130 species. A number of species appear for the first time and include 9 bat species, 4 insectivores, and a large range of rodents and birds.

Hyaena may have occasionally been present but effectively Wolves and Brown Bears replace them. Smaller carnivores include Stoat, Weasel, Badger, Wolverine and Arctic Fox. Many of these are stratigraphically associated with a significant increase in faunal diversity. Direct evidence of predation is available for Reindeer (by humans and wolves). Evidence of a punctured goose egg (Figure 95) and punctured goose bones highlight what must have been extensive small carnivore predation. Cast antlers of Reindeer occur having been collected and transported into the cave where a number have been gnawed by Wolves.

The species Mammoth, Woolly Rhinoceros, Wild Horse and European Bison and Giant Deer are present in the earlier stratigraphic levels but slowly decline and in some cases disappear within the most recent levels. Reindeer is present in relative and declining numbers. Extensive studies by Řičánková et al. (2015) have demonstrated that Palaeo-arctic mammal faunas have retreated eastwards during the mid- and late Quaternary. This suggests radical shifts in species ranges and not extinction.

A further significant development in the vertebrate fauna is the enormous increase in the bird and fish population. The population is diverse and indicates a considerable degree of change in regional habitat and particularly the development of regional woodland and ground cover.

The evidence available from the excavation by Armstrong is seemingly more diverse than any other UK Quaternary site. I contend that it is not reasonable to suggest that the sediments are mixed and/or that there has been a downward movement of bone material. It is reasonable to consider the entire faunal population from these levels which far exceeds similar populations from other sites. There is a dominance of arctic species (Woolly Rhinoceros, Mammoth, Reindeer and Bison) but significantly these are associated with a much more diverse fauna. The frequency and long-term presence of individual species, such as European Wolf, does not point to sudden change related to climatic variation. The frequency of minimum individuals for species shown in Figure 108 has no evidence of species reduction or change in stratigraphic levels 6-10. Sedentary species are shown in Figure 109, which shows a slight reduction in sedentary species but an increase in those that regionally migrate.

I suggest that the concept of a diagnostic Mammoth Steppe fauna requires further research at this locality. The very small sample of the vertebrate population used to define the existence of this distinct fauna within this locality is not reliable.

7.3 Dating of the Quaternary deposits

The application of modern dating techniques has been frustrated by the controversy surrounding the excavations of Armstrong and the uncertainties related to disturbance of both sediment and sediment and palaeontological samples. Much of this report has been concerned to reconstruct the original temporal and spatial distribution of archaeological and palaeontological remains known from the cave. Difficulties both in the accuracy of available records and those presented by the lack of any record for some finds are discussed. All of the reconstructions within this report include a clear statement illustrating the total known collection and the sample size used for individual reconstruction. It is important, despite some repetition, to clearly restate the methods of reconstruction of finds and subsequent selection for dating on the basis of specific assemblages or sediments. It is also crucial to draw attention to geomorphic activity, which has obviously had an effect on some samples used in age estimations and which have not been adequately taken into account.

7.3.1 Geomorphology stratigraphy - biostratigraphy

In considering the stratigraphy of this cave, my main concern has been to examine the work of Leslie Armstrong. In 1924, Armstrong was working under the auspices of Sir William Boyd Dawkins and was well informed about the large excavation projects of the 19th century. These excavations had historically established a stratigraphic sequence for Quaternary remains and this was particularly the case for the Creswell Caves. It is to Armstrong's credit that he examined the remaining sediments within the Pin Hole Cave and realised that much of the cave was not explored. It is also to his credit that, by accident or design, he choose to record his excavated finds in two dimensions a distance north from the cave entrance and a depth measured from the robust stalagmite floor deposits that seem to cap much of the infilling deposits for the entire length of the cave. It is unclear if he appreciated the full significance of this act. He discovered that the sediment infill was sloped toward the entrance and he had chosen a capping deposit that replicated the entire slope as his datum. Despite the various difficulties in interpreting the excavation today, if he had used a horizontal datum the results of the excavation would have cut horizontally through sloping deposits and made interpretation unintelligible. The significance of this method was not finally realised until noticed in 1977 in some offhand comment within a private letter from Armstrong to Michael Eager at the Manchester Museum.

Despite minor difficulties with Armstrong's two-dimensional records, he extensively published excavation accounts not based upon this empirical record. The accounts grouped excavated finds within subjective frameworks dependent upon archaeological typology, vertebrate biostratigraphy and palaeoclimatic sequences. The results have kept academics in employment since and has generated to this day discussions concerning the same conflicts, particularly where the empirical data conflicts with subjective sequences. Anomalies are soon explained away as the result of bad excavation techniques or of mixing.

It is in this context that Armstrong's two-dimensional recording method excels if not used in isolated contexts. The main criticism of Armstrong's endeavours is that he was either not aware of or didn't believe that the detailed taphonomic circumstances associated with the excavated finds were essential for a full understanding of both archaeological and paleontological remains. Having stated this, Armstrong did record a series of specific associations between artefacts and vertebrate bone that are of considerable interest today.

This report has dealt exhaustively with the issues relating to Leslie Armstrong, his excavation, his methods and results. As an aid for comment, an enormous archive of finds and contexts is made available in the digital archive. The site is exceptional in a UK context and is fairly unusual in a global context. The locality has not been fully appreciated for the information it provides, mainly due to the difficulties in correlating individual studies to more widely based concepts. The difficulties of recording techniques in the excavation have led to many studies that have commented upon apparent 'mixing' and post-mortem movement that undermines attempts to correlate it with similar sites. This has been particularly acute in archaeologically based Palaeolithic studies. Many comments point out that correlation with either the Mendip group of caves in the late Palaeolithic and the many fluvial and open-air sites in southern Britain is difficult. The question has to be asked why this site should correlate. This locality has one of the deepest sedimentary sequences containing elaborate archaeological sequences and includes over 58,000 vertebrate bones derived from over one hundred species, several of which are unique to the locality. The archaeological record is diverse and includes lithics, bone work, bone fabrication, engravings, amber, mineral fossil shell and mollusca, many of which are not to be found in other sites.

These rather defensive points underline the sheer importance of clarifying the context within this cave in order that the very significant body of data helps to clarify the interpretation of middle and later Palaeolithic presence in the UK.

7.3.2 Stratigraphy sequence - contextual relationships association and the misleading concepts of 'mixing'

Jenkinson (1984) shows significant areas of ruptured cave roof through which fine sediment has fallen or washed into the void below. This has formed a distinct deep fill within the cave rear which slopes to the south and the cave entrance. The main area of slope occupies what has become known as the 'passage'. This is precisely the area that contained much of the archaeological and palaeontological deposits. The preserved stalagmite formation that covered much of the unexcavated deposit is preserved and clearly seen on the cave wall today. Both deposits have been mapped in detail during 1977-1980. The relative height varies from the lowest point at the entrance to the highest known point in the cave rear. The western stalagmite is approximately 0.6m higher than the stalagmite preserved upon the eastern wall. The western wall stalagmite was used by Leslie Armstrong as the datum from which the dimension north into the cave and the depth of deposits were recorded. The eastern wall was only referred to in notes where relevant to specific finds.

Following mapping of the western stalagmite, the original position of all finds were replotted into a reconstructed series of 0.3m square excavation spits or boxes. The results were also summarised in the context of 0.3m deep stratigraphic levels along the northern length of the excavation. It is important to note that Armstrong's handwritten records refer to both horizontal sections and those that were divided into boxes, an approach which is clear from the dimension recorded on finds. For this reason the distribution of archaeological remains was within the context of 0.3m boxes that are illustrated in Figure 6 and Figure 7. Figure 7 illustrates the distribution of artefacts within a 0.3m box and where they are associated with similar contents in other boxes. This is significant in that it shows the sloping occurrence of the Middle Palaeolithic assemblages which were originally demonstrated in 1984 (Jenkinson 1984). A similar phenomenon has been illustrated for vertebrates.

The combination of geomorphic observation, current studies, and the results of vertebrate studies and the distribution of archaeological assemblages allow a multi-facetted reconstruction of events in the formation of the cave infill and occupation. This can be considered as:

1. The out-washing of some pre-Devensian glacial sediments by fluvial action passing through three roof ruptures in the cave rear and two within the entrance area. Uranium Thorium series estimations have provided a timescale for this event and the presence of a Hippopotamus molar in stratigraphic level 11 may indicate a remnant older interglacial age sediment surviving further north in the rear. This is further confirmed by the presence of a red sand at the basal levels of the cave where it occupied a narrow fissure within the cave floor (Garrod 1927).
2. In-washing of sediment from the roof ruptures in the form of fine particles derived from the glacial deposits surviving on the limestone scarp surface. These contained Quartzite 'Bunter' type pebbles. Evidence of this action is available in the form of very many small 'Bunter' pebbles and flakes that are in collections today and which were originally excavated underneath the rear cave roof area and below the ruptures above the Pin Hole within the entrance. Interestingly some of these Quartzites from the entrance have been considered as artefacts.
3. Groups of Hyaena took up residence in the cave rear in an area currently about 80ft (24m) from the cave entrance. Evidence included a large number of adult bones of this species associated with fragmentary foetal and juvenile remains. These include an unusual near-complete foetal skeleton that may have been buried in a collapse of sediment. These remains display an inclined distribution, where some elements from adjacent stratigraphic levels belong to the same individual Hyaena. This evidence confirms that the occupation area was on the inclined surface of the sediment mound underneath the roof fissure. Excavation records using an inclined datum for depth often show the same animal spread between at least one sloping depth and excavation spit. These are associated with other species such as Woolly Rhinoceros, Mammoth and Horse. Many of these vertebrates have chewing, gnawing, biting and surface licking marks on their bones, which confirms that they were of interest to Hyaena and particularly to young and juvenile individuals. The intense activity of this species in living, breeding and eating in this area of the cave undoubtedly had a significant effect on the distribution of remains that became buried within the deposits.
4. The first evidence of Neanderthal use of the cave is associated with the appearance of a limited range of heavy cutting tools made from local quartzite pebbles. The occupation is concentrated within the passage, in an area within 55-60ft (16.5-18m) from the cave entrance; the assemblage distribution extends onto the inclined slope occupying the rear of the cave. There is no evidence of the assemblage in the adjoining small eastern chamber. The area of lithic assemblage is associated with Woolly Rhinoceros, Mammoth and Horse remains. There is direct evidence of Neanderthal interest in Woolly Rhinoceros and Mammoth, which is known from two surface cut-marked tibias and a range of bones undamaged by carnivores. The evidence suggest that occupation by Hyaena and Neanderthal groups are separate in time. The presence of juvenile animals demonstrated that Neanderthal occupation was during spring and summer.
5. Use of the cave at this time appears to be in temperate conditions. There is clear evidence that fine sediment is washing into the cave through the roof rupture and forming the rear mound and slope into which the Neanderthal lithics are deposited. Additionally, Hyaena is a species associated with temperate environmental conditions and seems to be thriving in this contemporary habitat. Grayling and Pike are relatively frequent in these levels and their presence directly indicate clean cool water nearby and mild winters. The so-called cold-loving species (Mammoth and Woolly Rhinoceros) are shown by the presence of juveniles to be summer visitors. This does not compare very favourably with the concept of a cold 'Mammoth Steppe' ecosystem.
6. Following a distinct hiatus between lithic distributions, there is evidence of a further phase of Neanderthal use of the cave, which is more extensive in terms of distribution and now includes a greater area of the passage. There is very distinct evidence of great change in the lithic assemblage. Artefacts now include significant numbers of flint and fewer robust quartzite tools. Lithics continue to include cutting edges but are now accompanied by more frequent scrapers. There is very clear evidence that the middle assemblage is associated with the remains of Wild Horse, European Bison and Reindeer. The presence of a total range of body parts suggests that these animals were hunted and much of the carcass was transported into the cave. All of the bone material continues to contain very large numbers of flakes and gnawed bones, which suggest that there are episodes of Hyaena and European Wolf residence. The presence of Reindeer is of great interest. Antler remains are prolific and while some show evidence of massacre many are cast, from both male and female individuals. This not only suggests late summer and autumn presence but also deliberate collection of cast antler. The identity of the collector is unknown. Some are gnawed, indicating that carnivores may have been attracted to the fresh antler. The majority are undamaged by carnivores and are concentrated in levels associated with Middle Palaeolithic type lithics. An obvious conclusion is that these were collected as a resource.

This original system, utilising the mapped stalagmite datum, has been used by Jacobi et al. (1998) to establish a reconstruction using specific individual dimensions recorded upon individual finds. The result is of considerable interest but in reality replicates Jenkinson's (1984) study using dimensional spot measurements instead of excavated boxes. Extant samples of charcoal and biological remains are not included.

Both relative radiometric and uranium series dating relies upon confidence in the method of reconstruction for remains that was developed in 1984. Subsequent studies also illustrate several issues of importance for considering samples for dating. These are discussed prior to discussion of individual dating results as it is clear that they are relevant in assessing stratigraphic context. These are:

1. The horizontal variation is the height of the capping stalagmite, seemingly used in reference to the use of Armstrong samples selected for dating. This approach requires caution. There is already one example of a Mammoth tibia from the base of the cave deposits that has a coating of the stalagmite derived from three metres above its excavated location where it is dated to the Late Glacial.
2. There are a range of items, particularly in the 5-8ft (1.5-2.4m) stratigraphic levels, which as isolated pieces have been treated controversially and are thought to be mixed. Some of these items were originally located by vertebrate bones that have been subsequently dated.
3. There is a major geomorphic feature and surface evident within the 'inner chamber' and passage boundary that is rarely featured when the context of remains is considered. Virtually all of the stratigraphic distributions (illustrated in Figure 7 and the vertebrates in the digital archive) clearly show a cluster of objects between 60 to 70ft (13-21m) north and between depth 6-12ft (1.8-3.6m). This clearly illustrates objects accumulated and accumulating down the slope from the rear of the cave. There is an abrupt decrease in the frequency of objects at the base of the slope. This is of relevance in the discussion of radiometric dating below. Above the sloped deposits, the archaeological and palaeontological remains appear to be much more horizontal in aspect. This is particularly clear to the south. Near the cave entrance the sediment movement is blocked. This is dramatically illustrated by a huge stalagmite block (the Pin Hole).
4. Post-mortem movement of remains is a taphonomic process known from both the sedimentary, typological and paleontological evidence. This report has demonstrated that this is a complex process that is of local temporal and spatial significance. This reinforces the need to fully understand the context and post-movement history of specific objects to be used for dating. This is not a uniform process that can be used to explain discrepancies discovered at a later date and which do not 'fit' perceived ideas. Attention has already been drawn to the fact that this cave has a concentration of 22,516 vertebrate bones.There has been vertebrate bone visible all over the cave floor throughout its entire history.
5. Finally it is necessary to comment upon the soundness of the scientific theoretical concept that can accept a stalagmite fragment which is demonstrated from its date to exhibit 'post-mortem' movement in time and location based upon a uranium series date. While this is a valid observation that is largely accepted as an idea, it does not follow that a vertebrate bone inscribed with two-dimensional information seventy years ago should be taken at face value in terms of context and post-mortem history within what was obviously an extremely busy cave.

7.3.3 Uranium series

The application of uranium series dating was undertaken by Peter Rowe during 1986-7. The research was part of a PhD study at the University of Norwich and operated under the auspices of the Creswell Project in general terms and at the Pin Hole Cave in particular. The results were published for several of the Creswell Caves in 1989 (Rowe and Atkinson 1985; Rowe et al. 1989). There are three groups of uranium dates from Pin Hole which include:

• During the cave survey and mapping in 1978, it became clear that the roof in the rear of the cave was composed of several large blocks that had fallen into the lower fissure and had become lodged within the fissure. This forms the cave roof that can be seen today. Two blocks are held in place as they are jammed against the limestone fissure wall by 'Bunter' quartzite pebbles, which have become lodged between fallen blocks and the wall. It is perhaps a coincidence that this area is 4m directly above the Middle Palaeolithic quartzite tool assemblages. The upper surface on one of these blocks contains a small stalagmite formation which was dated to 18.6 ka (Rowe et al. 1989, UEA100). The cave wall in the vicinity of these blocks reveals a small body of sediment adhering to the cave wall. The sediment contained a small stalagmite formation which has been dated by Rowe to 110 ka. (Rowe et al. 1989, UEA156). This is clear evidence of the one-time presence of last interglacial sediments filling the cave to this point, 3m above the current pre-excavation level.
• During excavation by Jenkinson, a small sediment block was removed from the cave roof within the rear chamber. This was examined using micro-excavation techniques. A small stalagmite formation within the block has been dated by Rowe to 14.7 ka (Rowe et al. 1989, UEA101). This region is very close to the depth of the 'lower slab' layer noted by Armstrong and by Garrod (Garrod 1927), which contained brecciated limestone block with thin stalagmite covering in places. The depth level is above the occurrence of the Mousterian assemblage. This is yet another enigma that questions the age of the earliest Middle Palaeolithic within the cave.
• A collection of stalagmitic sediment fragments were excavated by Armstrong and survive within the Manchester collections. Some of these were recorded in 1977 and are detailed in the digital archive. Several more samples are reported by Jacobi from the same collection. Some of these were dated by Rowe (Jacobi et al. 1998). Samples selected for dating were selected from the lower levels excavated by Armstrong. The standard two-dimensional level recorded by Armstrong was used. As a result the dates relate directly to the description given by stratigraphic levels within this report. The difficulties resulting from either contamination or leaching of uranium are full discussed by Rowe et al. (1989) and Jacobi et al. (1998). They suggest that despite considerable variation, all estimates suggest that the stratigraphic levels containing the Middle Palaeolithic assemblage date between 40-50 ka. This significant finding is in agreement with previous suggestions that these sediments date to what has been previously referred to as the Upton Warren Interstadial.
• A significant criticism of the dating estimates is that the formal association of the sedimentary infilling deposits and the fragments of variable age stalagmite fragments have been assumed to be coeval. It remains possible that the dated fragments are younger in age than the sediment ('yellow cave earth').

7.3.4 Radiometric dating

There have been isolated attempts to apply radiocarbon dating to the Pin Hole sedimentary sequence but the more comprehensive application has been undertaken by Jacobi (Jacobi et al. 1998). The account described within this report is based upon the original studies by Peter Rowe, Tim Atkinson and the author (Rowe et al. 1989) and the later study by Roger Jacobi (Jacobi et al. 1998), which aimed to date the Middle Palaeolithic sequence from the cave. These report a range of dates from two north locations and from within levels below 6ft (30cm). The age estimations are comprehensively published and all fall within an estimated range of 30-50,000 years. The results offer an extremely valuable insight into the age of vertebrate bones discovered in the lower part of the cave. The significance of the age estimations and particularly their relevance to the age of the Middle Palaeolithic assemblage is difficult to assess. There are a number of complicating factors:

• The age estimations are all near the limit of application of these radiometric techniques. Despite this uncertainty, it seems reasonable to assume that the age is greater than 40,000 years.
• The suggested ages of the differing deposit depths do not completely synchronise with the progression of depth of location from the capping stalagmite recorded by Armstrong.
• The lack of synchronisation and in some cases the inversion of estimated ages suggests that either there is disturbance of deposits or the association of the dated object and its context are not fully understood.

This is an important issue which potentially gives some insight to our understanding of the early prehistory of this cave. The issues relating to context are discussed in detail as a result and can be considered as follows:

Samples for radiometric dating were selected by Jacobi et al. (1998) from the excavated vertebrate bone that had been excavated by Leslie Armstrong. These were housed within the collection in the Manchester University Museum. The dimensional records recorded on individual finds by Armstrong were used to pinpoint the original location. The stalagmitic floor deposit was used as a datum to plot the location of specimens. Confusingly, this process was not correlated with the method published by the author in 1984, which in reality is exactly the same procedure and the results can therefore be directly compared with the stratigraphic level reconstruction of that date. Dimensional records written by Armstrong refer to arbitrary spits or boxes of excavated deposit, which were extended and contracted according to need but usually to not less than 6in. (15cm) for the north from the entrance measurement. Depth of spits seems universally to have been 1ft (30cm). The radiometric study has treated these dimensions as accurate actual spot measurements. The difference is minor but there are several examples of apparent 'mixing' that are the result of utilising Armstrong's measurements in this fashion.

A second difficulty is that the nature of association has been confused in two cases. Radiometric dating of a sample from 10ft (3m) depth at 64ft (19.2m) north from the entrance has produced a date of much younger age than anticipated and younger that the four radiometric dated samples above it. Jacobi cites the nearby presence of a European Badger bone as a potential reason for missing vertebrate remains. This area equates with stratigraphic level 10 described in this report and which uses the stalagmite datum on the western cave wall. The area from which this sample (and others) is derived at 64ft (19.2m) north of the entrance has an eastern wall underhang that is coated with a stalagmite deposit. This is coeval with the western wall remnant and survives as a deposit today, where it is formed across the cave floor sediments and runs down the eastern cave wall into the 'inner chamber' (referred to by Jacobi as the Trefoil chamber). The sediments adjoin the eastern wall and the contents of the chamber are capped by the stalagmite, which has been dated by uranium series estimates to be one and the same deposit and of Late Glacial age. In concise terms, the radiometrically dated bone is not associated with the basal deposit of layer 10. A second example is a nearby Mammoth bone of undoubted Level 10 age, which has a stalagmite coating dated to the Late Glacial.

The reported age estimations are derived from vertebrate bone samples that have been selected with a view to sampling the sediments on the basis of recorded depth (from the capping stalagmite). At least two separate areas of the cave sediment have been sampled. The area in the region of 30-40ft (9-12m) north from the entrance is the narrow beginning of the passage at the point where the cave reduces in width. The second area, normally referred to by Armstrong as the passage, occurs between 55-58ft (16.7-17.6m) extending to 60-65ft (15.2-19.8m) north from the entrance and, lastly, the rear of the cave, the structure of the cave profile and the depositional conditions for sedimentary remains have already been discussed and are different in these three locations.

• The area at the rear of the cave has sediments that are near horizontally deposited between limestone clasts of variable size. This is clearly illustrated in the series of 1cm excavated spit records (see digital archive).
• The area within 58/65ft (17.6/19.8m) north of the entrance is the location of the mound of deposit directly below the roof ruptures. In this the area to the north is near horizontal but the area to the south has a distinct gentle slope to the south. This phenomenon is clear from the profile of the surviving stalagmite deposit, and it is very clearly illustrated in the occurrence of the artefact assemblages (Jenkinson 1984), which replicate the south-sloping profile. There are multiple illustrations within this report where virtually all distribution reconstructions for vertebrate remains display this feature.
• The area of cave from the entrance to 40/45ft (12/13.7m) north appears to have a relatively horizontal deposition in comparison to the mound but is also very likely to receive sediment, archaeological and palaeontological items moving downslope where they may have accumulated against a smaller mound of infill. This is now indicated by the prominent stalagmite remnant, narrowing of the cave walls and evidence from distributions.

The purpose of this description is to show that the context of the dated vertebrate bone includes samples derived from the base of an inclined and sloping deposit which are recorded in such a way that is sympathetic to the slope (i.e. the sloping stalagmite datum) but there is no clue to the third spatial dimension. The use of 1ft (30cm) 'spit' records as a specific and accurate 'spot' measurement adds further confusion and this environment alone has two sloping profiles, firstly toward the cave entrance 55ft (16.7m) to the south and secondly 3ft (0.9m) to the east into the 'inner chamber' and its known Late Glacial deposits. The age estimations therefore offer an insight in the sense that the levels are shown to be older than 40,000 years but they also confuse our understanding of the specific age of the differing Middle Palaeolithic assemblages and associated fauna, as they appear to apply to items where there may well be an association but also include dated items that have undoubtably slipped down the south-facing slope and in at least one sense slipped eastward into the Late Glacial.

Lastly the report mentions that the upper levels of the cave, particularly toward the rear, contain lithics that are mixed and include Neolithic types (Jacobi et al. 1998), indicating that the upper levels are disturbed. Unfortunately it is not clear if there is stratigraphic evidence of this or if this is a suggestion to explain the radiometric dating. The issue is of some importance. Within the cave today the remnant of the flowstone floor on both sides of the cave, but particularly upon the western wall, shows no evidence of disturbance since deposition during the Late Glacial. There are three small disturbed areas mentioned by Armstrong - a known burial of Bronze Age near the cave entrance and two further small disturbances in the cave rear, not in the vicinity of lithic assemblages. Prior to excavation, the area of cave behind the flowstone blockage near the entrance presented a very narrow fissure with a flowstone floor that had several large blockages within the void. One of these is very evident from the traces of flowstone visible mid-cave (see Figure 5). This shows that just a little further north from the entrance and the Bronze Age disturbance, the cave was of extremely limited access following the Late Glacial flowstone formation. Interestingly, historical artefacts, ceramic and later debris have been discovered within the entrance but all discoveries further north than the flowstone blockage are of Quaternary age.

These comments are minor in real terms but underscores the trust and reliance upon two dimensions recorded 90 years ago as the only guarantee to support the Principle of Association, which is used in isolation from other data relevant to context. These examples highlight the need to clearly understand context and therefore association.

7.3.5 Conclusion - age of the sedimentary infill archaeological and paleontological assemblages and populations

The application of relative and absolute dating techniques has provided an extremely useful insight concerning the age of the sedimentary infill, archaeological assemblages and paleontological remains. It clear that none of the techniques have provided a clear and comprehensive chronology. The age estimations are of considerable use for understanding the chronology for the cave infill but the also reinforce the need for extreme care in the process of sample selection and the need to be aware of context within the sedimentary infill.

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