ramblingon

[A similar version of this report was published in Cave & Karst Science, Volume 52, No. 2 (2025) p. 67-70. Transactions of the British Cave Research Association.]

Abstract

Following the recovery of human remains from a site at Kewstoke Road, near Weston-super-Mare, several samples were sent for radiocarbon dating. A total of four radiocarbon dates have so far been obtained on a representative selection of samples from the Kewstoke Road rock shelter burial site, the obtained dates ranging between 3784-3651 BC and 3501-3351 BC. These radiocarbon dates place the burials at Kewstoke Road rock shelter within the Early Neolithic period. The findings from Kewstoke Road are compared with other discoveries of Early Neolithic remains in the caves of the Mendip Hills in Somerset, UK.

Figure 1. Locations of Kewstoke Road rock shelter excavation and other Mendip caves containing Early Neolithic human remains mentioned in the text. OpenStreetMap contributors (2015) Planet dump [accessed 15/12/24]

Introduction

The excavation site at Kewstoke Road, Weston-super-Mare, North Somerset is located on the north facing flank of Worle Hill and lies at approximately 60 – 65 metres above Ordnance Datum (aOD). The site of archaeological investigation is found at the base of a ‘rock shelter’ in a prominent rock outcrop with soil and gravel covered platform and adjacent to a scree slope (Plate 1). It was during works by the landowner to construct a pathway cutting across the scree-slope, that human remains were uncovered. A visit from Police Forensic officers confirmed the human remains were of considerable age and, therefore, no further investigation by the police was required.

At the request of the landowner(s) an archaeological investigation was conducted, and a Ministry of Justice Licence was applied for and granted (No. 24-0123 dated 30^th May 2024) . The archaeological excavation was conducted in line with current guidelines and standards with regard to an excavation that included the recovery of human remains (Mitchell & Brickley, 2017). The site is on private land and there is no public access to it.

Geology

The underlying bedrock comprises Clifton Down Limestone Formation, a sedimentary bedrock formed between 344.5 and 337 million years ago during the Carboniferous period. It is described as splintery dark grey calcite and dolomite mudstones, pale grey oolitic, dark grey bioclastic and oncolitic limestones and some mudstones, with scattered cherts and silicified fossils. It was deposited in a barrier/back barrier/shelf lagoon setting. To the south lies a boundary with sedimentary bedrock consisting of Mercia Mudstone formed during the Triassic period, 252.2 and 201.3 million years ago (BGS, 2024). The Carboniferous limestone crops out as a discontinuous elongated west/east trending ridge, the Weston-Worle inlier, surrounded by Mesozoic, or younger rocks and deposits (Whittaker & Green, 1983). Although there are no superficial deposits recorded in the immediate area around the site location there are extensive tidal flat deposits – clay, silt and sand formed during the Quaternary period, 11.8 thousand years ago and the present, approximately 500m further to the south below 10m aOD (BGS, 2024). In the wider field, human activity is known from Palaeolithic, Mesolithic, and Neolithic remains from the limestone caves of the Mendip Hills (Whittaker & Green, 1983).

Figure 2. Continuous section showing general distribution of human remains (highlighted in red) prior to commencing the excavation and recovery of human remains (see also Plate 2). Scale bar = 1m

Radiocarbon dates on human remains from Kewstoke Road rock shelter and other Mendip caves

Lab. No.	14C	±	Date BC		d13C	d15N
SUERC-126136	–	–	3784	3651	-20.8	10.2
SUERC-128083	4967	21	3791	3653	-20.9	10.3
SUERC-131382	4757	27	3635	3383	-21.0	10.9
SUERC-131383	4616	23	3501	3351	-21.3	13.2

Table 1. Radiocarbon (¹⁴C) date ranges and stable isotope analyses for the human remains recovered from Kewstoke Road rock shelter.

An initial radiocarbon analysis, on behalf of Avon and Somerset Police, was conducted on a human partial left mandible that reported a date range with 95.4% probability covering the period 3784-3651 BC (early Neolithic). A second analysis on material recovered during the police forensic investigation (a human long bone fragment) confirmed this result providing a date range 3791-3653 BC. A further two radiocarbon analyses were conducted on material recovered during the 2024 archaeological investigation; a partial petrous (part of the skull) from a juvenile (SK 112, SUERC-131383) provided a date range with 95.4% probability covering the period 3501-3351 calBC; and an adult (SK 104, SUERC-131382) fibula (left, distal end) provided a date range 3635-3383 calBC (see Table 1).

Location	NGR	14C	±	Date BC
Picken’s Hole, Compton Bishop	ST 3965, 5502	4800	55	3695	3500
Chelm’s Combe, Cheddar	ST 4634, 5447	4680	45	3630	3365
Totty Pot, Cheddar	ST 4826, 5357	4706	35	3630	3370
Hay Wood Cave (dates range)	ST 3398, 5824	–	–	3946-3781	3622-3368

Table 2. Other Mendip caves with earlier Neolithic radiocarbon (¹⁴C) dates on human remains

Whittle et. al. (2011) estimate that the beginning of the Neolithic in south-west England occurs 3940-3735 BC. An outline summary of other Mendip caves that have produced radiocarbon dates pertaining to the Early Neolithic period is presented here (Simmonds, 2014, and Table 2). At Picken’s Hole, Compton Bishop (ST 3965, 5502), a radiocarbon date of 4800 ± 55 bp, calibrated to 3695-3500 years BC (86.8% confidence) was obtained for a human premolar indicating an early Neolithic date; however, there was nothing else found at the location, neither finds nor context, to add support to a Neolithic date (ApSimon and Mullan, 2018). In view of the limited number of surviving remains it has been suggested that Picken’s Hole might have been used during an intermediary period before the bulk of the remains were removed to another site for secondary burial (Peterson, 2019). At Chelm’s Combe in Cheddar, Somerset (ST 4634, 5447) a medium-sized rock shelter was located, unfortunately, the site has now been lost to quarrying activities. The site was unusual; in the cliff face below the main rock shelter, covered by limestone scree, the excavation team discovered a small rock-cut chamber that had been used for burial. There were large quantities of human remains recovered from the main rock shelter upper fills (Balch and Palmer, 1926). A single radiocarbon date of 4680 ± 45 bp, calibrates to between 3630-3365 years BC obtained on a long bone from either the rock-cut chamber or the main rock shelter levels (Ambers and Bowman, 2003). Peterson (2019) suggests two different stages of secondary burial occurring at Chelm’s Combe. An intermediary period where bodies were placed in the rock shelter long enough for some fragmentation to occur before selected elements were removed for secondary burial elsewhere, perhaps to the rock-cut chamber. The bone assemblage recovered from that part of the site seems to represent a secondary burial assemblage, dominated by disarticulated crania, long bones, and axial elements. Carnivore damage on two of the femurs provides further evidence that they were exposed prior to their final burial. To the southeast of Chelm’s Combe, excavations conducted at Totty Pot, Cheddar (ST 4826, 5357) produced an assemblage of human and animal bone (including wild and domesticated species) with artefacts including microliths and Beaker to Middle Bronze Age pottery sherds (Gardiner, 2016). A programme of AMS dating on six individuals with the earliest date Mesolithic, 7445-7080 cal BC, the other five individuals produced dates ranging 3630-3370 cal BC (early Neolithic) to the latest at 2830-2460 cal BC (late Neolithic), with three intervening results indistinguishable at c.3340-3000 cal BC. The prominent Neolithic funerary presence was unexpected, given the lack of any diagnostic material remains relating to this period. It is clear that at least four distinct events involving the deposition of human remains occurred at Totty Pot. The earliest during the Mesolithic and at least three temporally distinct episodes of deposition in the Neolithic (Schulting et al., 2010). At the northwestern extent of the Mendip Hills, Hay Wood Cave, near Hutton, North Somerset (ST 3398, 5824) is a limestone cave located on the north-facing flank of Bleadon Hill. The cave was excavated by members of the Axbridge Caving Group and Archaeological Society from 1957 until 1971 (Everton, 1972). The cave occupies a steep, north-facing slope, at present tree covered. A large assemblage of human remains and artefacts, including Mesolithic microliths, were recovered from the site. An AMS dating programme in 2013 by Schulting, et al. confirmed the attribution of the assemblage to the earlier Neolithic period. The use of the cave for burials was modelled as commencing in the period 3930-3715 BC and ending 3580-3350 BC (95.4% confidence), with the majority of the 10 dated individuals centred on 3600-3500 cal BC. Stable carbon and nitrogen isotope data indicating a predominately terrestrial diet, despite the proximity of the coast (Schulting et al., 2013). Hay Wood Cave is located approximately 5km south of the Kewstoke Road rock shelter excavation site.

Chelm’s Combe, and probably, Picken’s Hole are likely associated with secondary burials; whereby bodies were decomposed in one location over an unspecified period of time before elements of the body were removed and transported for deposition at an alternative location. Successive inhumations, where multiple individuals are deposited in the same place for an extended period, such as at Totty Pot took place for between 530 and 830 years, and at Hay Wood Cave for a shorter episode of between 150 and 400 years (Peterson, 2019). Sequential inhumation, or repeated use of the same place for later burials would imply social continuity at that location.

This small group of Mendip caves, chosen as suitable locations for Early Neolithic burials, in general, all have a northerly facing aspect. Chelm’s Combe, Hay Wood Cave, and a postulated alternative entrance to Totty Pot face due north, and Picken’s Hole faces north-east (Peterson, 2019).

Stable isotopes and diet

Isotopic analysis of bones and teeth can help to shed light on the diets of people in the past. Stable carbon (d¹³C) measurements reveal that humans obtaining 100% of their protein from marine sources will have the same values as recorded from fish or seal bone, circa -12 ± 1‰, while humans obtaining all of their protein from terrestrial plants and/or animals will have values, circa -21 ± 1‰. Stable nitrogen (d¹⁵N) measurements reflect trophic levels and can inform on the relative proportions of plant and animal protein in the diet (Schulting, et al. 2013). Human consumers of terrestrial plants and animals typically have d15N values in bone collagen of about 6-10‰ whereas consumers of freshwater or marine fish and/or seals may have d¹⁵N values of 15-20‰ (Tykot, 2004).

The samples analysed from Kewstoke Road rock shelter show δ¹³C (range -20.8 to -21.3‰) and, generally, δ¹⁵N (range 10.2 to 10.9‰) values of bone collagen that might be indicative of people who did not have significant marine resources in their diet. However, it is noted that SK112, a juvenile, returned an elevated level of δ¹⁵N, +13.2‰, this is still below the range expected to indicate a significant freshwater or marine fish input to the diet. This higher level of δ¹⁵N might reflect a residual nursing effect, i.e. breastfeeding or weaning (Waters-Rist, et al., 2011). A similar occurrence was noted by Schulting et al. (2013) when reporting stable isotopes and Neolithic diet for samples recovered from Hay Wood Cave.

Comments

A total of four radiocarbon dates have so far been obtained on a representative selection of samples from the Kewstoke Road rock-shelter burial site, the dates ranging between 3784-3651 and 3501-3351 BC. A comparison of these dates to those obtained from the nearby Hay Wood Cave show a remarkable similarity, where use at that cave for burials has been modelled as commencing 3930-3715 BC and ending 3580-3350 BC (Schulting et al., 2013).

The open texture of scree deposits and the highly active, mobile processes mean that it is extremely common for human bones and artefacts which have been placed on the surface to be moved, redeposited, and combined in complex ways (Peterson, 2019). This mobility of scree materials is reflected in the widespread distribution of the human (and faunal) remains and artefacts recovered at the Kewstoke Road rock shelter. There is also an indistinct boundary between some of the contexts recorded.

A full and detailed report on the findings from Kewstoke Road rock shelter is currently underway and will be published at a later date. Further archaeological excavation concentrating on the rock shelter platform is scheduled to commence in spring/summer 2025.

Acknowledgements

Thanks, are especially extended to the landowners, Cye and Zoe Hayes, who requested and assisted the archaeological investigation; to Valentina Bernardi for help with identification of human skeletal elements; to the attending members of the Mendip Caving Group (MCG) for their help with the excavation; and to Yvie Barbier for digitizing site drawings.

Map data copyrighted OpenStreetMap contributors and available from https://www.openstreetmap.org

Funding statement

The Kewstoke Road rock shelter project benefitted from a British Cave Research Association (BCRA) – Cave Science and Technology Research Fund (CSTRF) grant. This award was used to provide radiocarbon dates on the human remains, and the results obtained are included in this summary.

References

Ambers, J. and Bowman, S. 2003. Radiocarbon measurements from the British Museum: datelist XXVI. Archaeometry 45 (3), p.531-540

ApSimon, A.M. and Mullan, G.J. 2018. The human teeth from Picken’s Hole, Somerset. Proceedings of the University of Bristol Spelaeological Society, 27 (3), p. 339-342

Balch, H.E. and Palmer, Rev. Preb. 1926. Excavations at Chelm’s Combe, Cheddar. Proceedings of the Somerset Archaeological and Natural History Society, 72/2, p.93-123

British Geological Survey Geology Viewer, https://www.bgs.ac.uk/map-viewers/bgs-geology-viewer/ Accessed online July 2024

Everton, A., and R. 1972. Hay Wood Cave burials, Mendip Hills, Somerset. Proceedings of the University of Bristol Spelaeological Society, 13 (i), p. 5-29

Gardiner, P.J. 2016. Totty Pot, Cheddar, Somerset: History of the archaeological finds and excavations from 1960 to 1998. Proceedings of the University of Bristol Spelaeological Society, 27 (), p. 39-72

Mitchell, P.D. and Brickley, M. (Editors). 2017. Updated guidelines to the standards for recording human remains. Chartered Institute for Archaeologists (CIfA), Reading

Peterson, R. 2019. Neolithic cave burials: agency, structure and environment. Manchester University Press

Schulting, R.J., Gardiner, P.J., Hawkes, C.J. and Murray, E. 2010. The Mesolithic and Neolithic human bone assemblage from Totty Pot, Cheddar, Somerset. Proceedings of the University of Bristol Spelaeological Society, 25 (1), p. 75-95

Schulting, R.J., Chapman, M., and Chapman, E.J. 2013. AMS 14C dating and stable isotope (carbon, nitrogen) analysis of an earlier Neolithic human skeletal assemblage from Hay Wood Cave, Mendip, Somerset. Proceedings of the University of Bristol Spelaeological Society, 26 (1), p. 9-26

Simmonds, V.J. 2014. An overview of the archaeology of Mendip caves and karst (revised 2016). Accessible at https://www.mcra.org.uk/wiki/lib/exe/fetch.php?media=archaeology:mendip_archaeo_simmonds.pdf

Tykot, R.H. 2004. Stable isotopes and diet: You are what you eat. Proceedings of the International School of Physics “Enrico Fermi” Course CLIV, M, Martini, M. Milazzo, and M. Piacentini (Eds.) IOS Press, Amsterdam

Waters-Rist, A.L., Bazaliiski, V.I., Weber, A.W. and Katzenberg, M.A. 2011. Infant and child diet in Neolithic hunter-fisher-gathers from Cis-Baikal, Siberia: intra-long bone stable nitrogen and carbon isotope ratios. American Journal of Physical Anthropology 146: 225-241 (2011)

Whittaker, A. and Green, G.W. 1983. Geology of the country around Weston-super-Mare. Memoir Geological Survey of Great Britain, Sheet 279 with parts of 263 and 295. Institute of Geological Sciences, Natural Environment Research Council, Geological Survey of Great Britain, England and Wales. Available at: https://webapps.bgs.ac.uk/Memoirs/docs/B01667.html

Whittle, A., Healy, F. and Bayliss, A. 2011. Gathering time: dating the early Neolithic enclosures of southern Britain and Ireland. Oxbow Books, Oxford

A selection of representative photographs:

Plate 1. Overview of Kewstoke Road rock shelter and exposed human remains prior to excavation and recovery. Some initial cleaning of loose debris has been done (Scale = 0.4m).

Plate 2. North and northwest facing sections prior to excavation commencing. Exposed human remains can be seen centre right (cranium and long bone, SK104) and inset (Scale = 0.5m). Scale = 2metres(m)

Mendip Times article regarding the discovery of a woolly rhinoceros skull and other faunal remains from a small cave located in Fairy Cave Quarry, Mendip Hills, Somerset.

https://issuu.com/mediafabrica/docs/issue_7_-volume_21-_mendip_times/80

Corrections: Caption for image (lower left) should read: Woolly rhino, distal end of humerus. Photograph by Roz Simmonds, 14/02/2025. Also, presentation by Danielle Schreve was given at ‘Mendip Rocks Festival‘

17th Annual J’Rat Digging Awards held at the Hunters’ Lodge Inn, Priddy, Mendip.

Well done to Ali Moody & team for a tremendous effort and discovery in Longwood Valley Sink.

Below are the slides from the PowerPoint presentation I gave regarding some faunal and human remains found and recovered from a three Mendip sites in 2024/2025.

The common name Fly Agaric comes from the practice of breaking the cap into platefuls of milk, used since medieval times to stupefy flies. It is a strong hallucinogen and intoxicant and is used as such by the Sami of northern Scandinavia. In such cases the cap is dried and swallowed without chewing, the symptoms begin 20 minutes to 2 hours after ingestion. The central nervous system is affected, and the muscles of the intoxicated person go into spasm. This is followed by dizziness and a deathlike sleep; during this stage the mushrooms are often vomited, but the stupor and drunkenness continue. While in this state of stupor, the person experiences vivid visions, and on awakening is usually filled with elation and is physically very active. This is due to the nerves being highly stimulated, the slightest effort of will producing exaggerated physical effects; for example, the intoxicated person will make a gigantic leap to clear a small obstacle. The Sami may have picked up the habit of eating Amanita muscaris through observing the effects of the fungus on reindeer, which are similarly affected. Indeed, reindeer like it so much that all one has to do to round up a wandering herd is to scatter pieces of the mushroom on the ground. Another observation the Sami made from the reindeer was that the intoxicating compounds can be recycled for consuming the urine of an intoxicated person. The effects of this species are exceedingly unpredictable due to the fungus containing different amounts of the toxins ibotenic acid and muscimol according to season and the method of cooking and ingestion, as well as the subject’s state of mind. Ibotenic acid is mostly concentrated in the coloured skin of the cap. This very unstable compound rapidly degrades on drying to form muscimol, which is 5 to 10 times more potent. This is an extremely dangerous mushroom to ingest, even in small amounts, and should always be treated as deadly poisonous.

Reference: Roger Phillips. 2006. Mushrooms: A comprehensive guide… Macmillan

[A similar version of this report was published in Caves & Karst Science, Vol.48, No.3, p.95-96. 2021. Transactions of the British Cave Research Association]

Abstract:

During 2011, a small-scale excavation in Toil and Trouble, located in the lower series of passages in Hallowe’en Rift, recovered a single animal bone from the fine sediments. The bone was later identified as Bison priscus 1^st phalanx. In April 2019, the bone was sent for radiocarbon dating, unfortunately, this was unsuccessful due to insufficient collagen in the sample. However, there is supporting evidence to suggest it is the find is of Pleistocene date and adds to the existing list of Mendip caves that have yielded ice age mammalian faunas.

Introduction

Hallowe’en Rift (Halloween Rift), NGR ST 5353 4809, altitude 130 metres(m) above Ordnance Datum (MCRA, accessed 2021), located in wooded hillside lying to the north-east of the Wookey Hole Cave ravine (Mendip Hills, Somerset). Excavation of the cave was commenced in 1982, but by the end of the 1980s interest at the site had waned. Then, in the early 1990s activity in the cave re-commenced and it was during these excavations that the dig on the east side of the entrance was started. Interest, once again, waned as the participants moved onto pastures new. In 2009, the current phase of activity began, and this phase of work concentrated on expanding the cave to the east-side of the entrance. The exploration and investigation of Hallowe’en Rift remains an ongoing project and several potential leads are being actively pursued.

Figure 1. Survey of the eastern passages in Hallowe’en Rift, drawn by Duncan Price (2018) with additional annotation by author.

The cave consists mostly of a low bedding, partially filled with sandy silt with cobbles and boulders of dolomitic conglomerate and frequent fragmented speleothems, including stalagmites, stalactites, and flowstone. The low bedding occasionally intersects rifts, that are, generally, northwest/southeast aligned.

Faunal remains

During 2011 an excavation in the lower series of passages known as Toil and Trouble, a single animal bone was recovered from the fine sediments. The bone was later identified as Bison priscus 1^st phalanx (Danielle Schreve, personal communication). In April 2019, the bone was sent for radiocarbon dating, unfortunately, this was unsuccessful due to insufficient collagen in the sample.

Bison priscus 1^st phalanx: peripheral view (top), ventral view (bottom). Scale = 5cm.

Family: BOVIDAE; Tribe: BOVINI; Genus: BOS Bison priscus, 1st phalanx:
Length (mm)	Bp (mm)	Bd (mm)	SD (mm)	Weight (g)
69	45	36	34	39

Table 1. Measurements of recovered Bison priscus bone.

The Steppe Bison (Bison priscus) existed in the British Isles, Europe and beyond from the Middle Pleistocene before becoming extinct at the end of the last Ice Age, about 10, 000 years ago (Marsolier-Kergoat et. al., 2015)

In the  British faunal record, mammal assemblages identified in caves are a valuable resource. Bison (Bison priscus) is recorded as part of the mammal fauna assigned to the Joint Mitnor Cave mammal assemblage-zone (MAZ), Marine Isotope Stage (MIS) 5e, c.128-116 ka and to the Bacon Hole MAZ, MIS 5a. However, a faunal grouping dominated by bison and reindeer, the Banwell Bone Cave MAZ, is believed to correlate closely with the Early Devensian, MIS 4, c.71-59 ka. Mendip sites containing a Banwell Bone Cave MAZ fauna includes the nearby cave site, Hyaena Den; and in the wider  environs at Picken’s Hole, Bleadon Quarry, Brean Down and Limekiln Hill Quarry. Also, to the north of the Mendip Hills at Pen Park Quarry in Bristol. The Banwell Bone Cave MAZ is a vertebrate assemblage from a cold climate (Currant and Jacobi. 2001). The mammal fauna of Banwell Bone Cave includes Lepus timidus, mountain hare; Vulpes lagopus, arctic fox; Vulpes Vulpes, red fox; Canis lupus, wolf; Ursus arctos, brown bear; Lutra sp., an otter; Gulo gulo, wolverine or glutton; Rangifer tarandus, reindeer and Bison priscus bison. Bison has also been recorded from the Lower Cave Earth deposits at Pin Hole, Creswell Crags, Derbyshire, and listed as part of the Pin Hole MAZ, Middle Devensian, MIS 3, c.59-24 ka. The Pin Hole MAZ also includes woolly mammoth (Mammathus primigenius), woolly rhinoceros (Coelodonta antiquitatis), wolf (Lupus lupus), as well as mountain hare, red-cheeked suslik, red fox, brown bear, stoat, polecat, spotted hyaena, lion, wild horse, giant deer, and reindeer. No evidence of bison was found in deposits attributable to the Gough’s Cave MAZ, MIS 2, c. 12.9-9.9 ka (Currant and Jacobi, 2001).

Comments

Speleothems from Hallowe’en Rift have been sampled and uranium-series dates obtained from the speleothems span MIS’s 15-13, 7c, 5e and 3,  with the youngest date 51.26 +0.31 −0.32 ka (Simmonds, 2019). Higham (2006) gives dates for Bison priscus bones recovered from Banwell Bone Cave as 52,700 ±1900 14C age BP to >59,000 14C age BP and for Hunters Lodge Inn Sink as >54,800 14C age BP. These dates fall within the range 59-24 ka, Middle Devensian, MIS 3. It is likely that the bone recovered from Hallowe’en Rift is contemporaneous with these dates. This Pleistocene date adds Hallowe’en Rift to the existing list of Mendip caves that have yielded ice age mammalian faunas.

Further excavation in this passage was abandoned due to the instability of the area. No other faunal remains have been recovered from the cave, so far. We are still looking.

Acknowledgements

In 2019 an application to fund radiocarbon dating was accepted by the British Cave Research Association (BCRA) Cave Science and Technology Research Fund (CSTRF). Unfortunately, the funding was not used due to a lack of collagen in the sample making it unsuitable for radiocarbon dating.

References

A Currant and R Jacobi. 2001. A formal mammalian biostratigraphy for the Late Pleistocene of Britain. Quaternary Science Reviews 20 (2001), Elsevier Science Ltd. 1707-1716

Marsolier-Kergoat, M-C., Palacio, P., Berthonaud, V., Maksud, F., Stafford, T., Bégouën, R., and Elalouf, J-M. 2015. Hunting the extinct Steppe Bison (Bison priscus) mitochondrial genome in the Trois-Frères Palaeolithic Painted Cave. PLoS one. 2015; 10(6); e0128267

MCRA (Mendip Cave Registry & Archive) www.mcra.org.uk/registry/sitedetails.php?id=150 Accessed 2^nd November 2021

Simmonds, V J. 2019. Evidence for Pleistocene frost and ice damage of speleothems in Hallowe’en Rift, Mendip Hills, Somerset, UK. Cave and Karst Science, Vol.46, No.2, (2019) 74-78

T F G Higham, R M Jacobi, and C Bronk Ramsey. 2006. AMS Radiocarbon Dating of Ancient Bone using Ultrafiltration. Radiocarbon, Vol 48, Nr 2, 2006, p179-195

All photographs by the author unless stated otherwise.

British National Grid ST 5353 4811

[A similar version of this report has previously been published in Caves & Karst Science Vol. 52 (1), 2025, p.27-31. Transactions of the British Cave Research Association (BCRA)]

Abstract

Excavation for speleological purposes is an intrusive process and as such it is vital that sediments and other deposits contained within caves are recognised, recovered where necessary, fully recorded and reported so that information about them is not lost and can be disseminated allowing for further research. During the ongoing excavation and exploration of Hallowe’en Rift, Mendip Hills, Somerset, a faunal assemblage has been recovered comprising steppe bison Bison priscus, brown bear Ursus arctos and ?horse Equus ferus. The assemblage is consistent with those found from nearby sites, such as Hyaena Den and Rhinoceros Hole, and taken together with U-series dating of speleothems suggest a Mid-Late Devensian date, MIS 3, c.59 – 24 ka. A Pleistocene date for the faunal assemblage recovered from Hallowe’en Rift further extends the list of ice age mammalian faunas found in Mendip cave sites.

Figure 1. Survey of Hallowe’en Rift and locations of faunal remains. Survey BCRA Grade 5c by D. Price, R. Taviner & V. Simmonds (latest update August 2024)

Introduction

Hallowe’en Rift, at British National Grid coordinates ST 5353 4811 and altitude 148m above Ordnance Datum (aOD), is located in a wooded hillside lying to the north-east of the Wookey Hole Cave ravine (Mendip Hills, Somerset). Excavation of the cave was commenced in 1982, but by the end of the 1980s interest at the site had waned. Then, in the early 1990s activity in the cave re-commenced briefly until interest, once again, waned as the participants moved onto pastures new. In 2009, the current phase of activity began, this phase of work concentrated on expanding the cave to the east-side of the entrance where significant discoveries were made in 2018. The exploration and investigation of Hallowe’en Rift remains an ongoing project and several potential leads, currently in the Soft South area, are being actively pursued (2024). The cave consists mostly of low passages, either partially or completely filled with sand, silt and clay containing cobbles and boulders of dolomitic conglomerate and frequent fragmented speleothems, including stalagmites, stalactites, and flowstone. The accessible low passages generally trend northeast/southwest, occasionally intersecting rifts, aligned northwest/southeast.

During 2011 an excavation along a lower passage known as Toil and Trouble, a single animal bone was recovered from the fine sediments. The bone was later identified as Bison priscus 1^st phalanx (Professor Danielle Schreve, personal communication). The bone was sent for radiocarbon dating, unfortunately, this was unsuccessful due to insufficient collagen in the sample (Simmonds, 2021). Excavations elsewhere in the cave continued and, in 2024, efforts were concentrated on pursuing a large fossil passage trending NE/SW. At the northeast extent, the excavation intersected a previously discovered section of the cave known as ‘Trick or Treat’ and the area beyond now further extended by continuing excavation. The passage currently trends eastwards, sloping down about 25 degrees. While emptying bags of spoil following a digging session several mineralised ‘bone’ fragments were recovered (later identified as ?horse). This discovery prompted a closer inspection of the excavation area and the potential source of the ‘bones’ was located. Following on from this initial find this area was carefully excavated, and more faunal remains were recovered from a ‘naturally-disturbed’ deposit underlying a calcite flowstone ‘floor’ (now degrading). This calcite layer was, in turn,  overlain by a thick (circa 0.60m) deposit of banded/laminated fine sediments mostly sand, silt, clay. The deposits fill the cave passage in this location to the roof. Where they have been exposed the conglomerate walls and ceiling are adorned with a variety of speleothems including curtains and drapes, stalactites, helectites, straws.

Faunal remains

The majority of the faunal remains so far recovered from the excavation of Hallowe’en Rift consist of Brown bear Ursus arctos, comprising a sub-adult and an adult (Professor Danielle Schreve, personal communication). As mentioned previously, a fragment of mineralised bone recovered from the spoil heap is probably of wild horse Equus ferus, although the recovered fragment is small.

Element	Number	Element	Number
Epiphysis (unfused)	3	Vertebra	3
Phalanx	26	Astragalus	1
Metapodial	20	Calcaneus	1
Tarsus	3	Humerus	1
Carpus	1	Scaphoid	1
Tooth	6
Total number of identified elements:		66

Table 1. Number of identified elements (Ursus arctos) recovered from Hallowe’en Rift, beyond Trick or Treat up to the end of 2024 when excavation was postponed due to adverse conditions.

The brown bear remains recovered so far consist mostly of ‘foot’ bones (phalanges, podials, tarsals/carpals, etc.), but also includes some vertebra and dentition. The smaller sizes of these faunal remains suggest these have not been transported any distance and are more indicative of an in-situ assemblage and likely to be close to their life/death position, perhaps, speculatively, a mother and cub perishing in hibernation. There is no evidence of carnivore gnawing on any of the faunal remains so far recovered. The assemblage does, however, suggests that there was an opening to the surface somewhere nearby. There are a number of thoughts to be considered; the faunal remains were found within a sandy deposit with frequent fragmented flowstone (frost/ice damage?) then sealed by calcite deposition. For this process to occur the cave passage must have been more open than it is at present, to allow access for bears and perhaps other mammals as yet undiscovered, and for it to be later sealed by calcite precipitation. The current human access into the cave is through a narrow, blasted fissure opened in the early 1980s, it is thought unlikely any larger mammals entered via this route. Therefore, access must have occurred through another, as yet undetermined, route. Subsequent to this ‘open’ phase, the cave then becomes filled with finer sediments over an undetermined time period. A red brown clay layer covering the calcite floor is the primary fill, followed by a hiatus as a thin grey silt is deposited, this, in turn, is overlain by a composite deposit consisting of  laminated/banded grey silt and red clay in fining upwards sequences (known as rhythmites) laid down with obvious periodicity and/or regularity. In Hallowe’en Rift, permafrost conditions on Mendip during the Pleistocene Epoch are thought to have penetrated to significant depth, with subsequent periods of thaw reaching lesser depths resulting in a deep impervious ice plug causing meltwater to be trapped and ‘ponded’. The ponded meltwater ‘topped up’ by the ingress of surface-derived water, probably reflecting seasonal changes (Simmonds, 2024). The ‘pulsing’ movements of the water resulting in layered sediment deposits encountered.

Figure 2. Mendip (and nearby) caves with faunal deposits containing remains of Brown bear, Ursus arctos (see Table 2). [OpenStreetMap contributors (2015) Planet dump [accessed 15/12/24]

Bears in [Mendip] caves

Excavations by the Natural History Museum in the 1970s at Westbury Quarry discovered an abundance of mammal bones including the extinct Deninger’s bear Ursus deningeri in deposits laid down during the early Middle Pleistocene, c.620 ka. In Britain, U. deningeri was replaced by the Cave bear Ursus speleaus after the Anglian glaciation, c.480-423 ka. which, in turn, was replaced by the Brown bear Ursus arctos appearing during MIS 9, c.339-303 ka. Faunal remains of Brown bear are relatively common in cave assemblages throughout the British Middle and Late Pleistocene during both warm and cold stages. Its presence in Britain in association with herbivores of cold open landscapes (woolly mammoth, woolly rhinoceros, and horse), as well as with those of temperate conditions, shows it to have been adaptable to a range of environments. Brown bears have evolved a generalist omnivore strategy; foraging for plants, tubers, berries, scavenging carrion, and preying on small mammals, and weak, older ungulates, and their calves. Temperature and snow conditions are reported to be the most important factors determining the composition of brown bear diet (Scott and Buckingham, 2021). Where it is still extant today the brown bear occupies a wide variety of habitats from tundra to temperate forests.

With general regard to Mendip cave sites, Brown bear (Ursus arctos) is recorded as part of the mammal fauna assigned to the Joint Mitnor Cave mammal assemblage-zone (MAZ), Marine Isotope Stage (MIS) 5e, c.128-116 ka, a faunal assemblage consistent with this MAZ was recovered from the nearby Milton Hill Quarry. Banwell Bone Cave MAZ (initially believed to correlate closely with the Early Devensian, c.71-59 ka, and formerly assigned to MIS 4) is assigned to MIS 5a, c. 83-71 ka. Within the Banwell Bone Cave MAZ, the remains of Ursus arctos appears to represent a larger form of the species. Brown bear has also been recorded from the Lower Cave Earth deposits at Pin Hole, Creswell Crags, Derbyshire and, therefore, listed as part of the Pin Hole MAZ, Middle Devensian, MIS 3, c.59-24 ka. The Pin Hole MAZ also includes steppe bison (Bison priscus) and wild horse (Equus ferus). Mendip sites with faunal assemblages attributable to the Pin Hole MAZ include sites near to Hallowe’en Rift at Hyaena Den and Rhinoceros Hole at Wookey Hole, and further afield at Picken’s Hole near Compton Bishop, and Uphill Quarry in North Somerset (Jacobi and Currant, 2011). It is likely that the Hallowe’en Rift faunal assemblage comprising brown bear, steppe bison and wild horse can be attributed to the Pin Hole MAZ and was deposited during the later stages of MIS 3. Brown bear also occurs in deposits attributable to the Gough’s Cave MAZ, MIS 2, c. 12.9-9.9 ka (Currant and Jacobi, 2001) in Cheddar, Somerset.

Site (Mendip Hills area)	NGR (ST)	Age	Age (ka BP)
Bleadon	3418, 5674	MIS 7	245 – 186
Hutton Cave	3611, 5811	MIS 7	245 – 186
Picken’s Hole	3965, 5502	MIS’ 5a and 3	c.83 – 24
Banwell Cave	3822, 5881	MIS 5a	c.83 – 71
Goatchurch Cavern	4758, 5822	Devensian	116 – 11.55
Hyaena Den	5322, 4795	MIS 3	59 – 24
Sandford Hill	4314, 5906	MIS 3	59 – 24
Dulcote Fissure	5670, 4430	Devensian	116 – 11.55
Uphill Quarry	3153, 5841	MIS 3	59 – 24
Sun Hole	4674, 5408	12378 BP	–
Gough’s Cave	4668, 5388	c.12200 BP	–
Soldier’s Hole	4687, 5402	Late Glacial	24 – 11.55
Bridged Pot	5260, 4866	Late Glacial	24 – 11.55
Walton, Clevedon	4184, 7265	Late Glacial	24 – 11.55
Aveline’s Hole	4761, 5867	Late Glacial	24 – 11.55

Table 2. The occurrence of Brown Bear Ursus arctos in archaeological sites in the Mendip Hills area. Dating of most records is indirect (adapted from Yalden, 1999, Table 4.3, pp. 113-115) 

Dating

Speleothems from Hallowe’en Rift have been sampled and uranium-series dates obtained spanning MIS’s 15-13, 7c, 5e and 3,  with the youngest date 51.26 +0.31 −0.32 ka (Simmonds, 2019). The spread of dates suggesting that the passages in Hallowe’en Rift were open for long periods during the Early to Late Pleistocene before becoming filled with sediment through the latter part of the Pleistocene and continuing throughout the Holocene.

In January 2025, an application to fund radiocarbon dating was accepted by the British Cave Research Association (BCRA) Cave Science and Technology Research Fund (CSTRF). A sample, consisting of a partial canine tooth, was selected and sent for analysis. Unfortunately, the sample failed due to a lack of collagen in the sample making it unsuitable for radiocarbon dating. The lack of collagen might be due to several factors including the age of the sample was beyond the limits of radiocarbon dating (c.50 ka) or mineralisation of the faunal remains recovered. It is hoped that another sample might be found that is suitable and sent for radiocarbon dating in the near future.

The faunal assemblage from Hallowe’en Rift comprising steppe bison Bison priscus, brown bear Ursus arctos and ?horse Equus ferus is consistent with those assemblages recovered from nearby sites, such as Hyaena Den and Rhinoceros Hole that have been attributed to the Pin Hole MAZ, Middle Devensian, MIS 3, 59-24 ka. In the absence of absolute dating a Middle Devensian date is suggested for the Hallowe’en Rift faunal assemblage including Ursus arctos.

Comments

In Britain, the faunal assemblages recovered and identified from cave sites are a valuable resource providing useful information and adding to the understanding of past environments and how organisms adapted to fluctuations between cold glacial and warmer interglacial periods. A Pleistocene date for the faunal assemblage recovered from Hallowe’en Rift further extends the list of ice age mammalian faunas reported from Mendip cave sites.

The excavation of caves for speleological purposes is an intrusive process, as such, it is vital that sediments and other deposits contained within caves, in this case faunal remains, are recognised, recovered where necessary, fully recorded and reported so that information about the deposits is not lost and can be disseminated widely and so allowing for further research.

Acknowledgements

Without the commitment, determination, and camaraderie of a dedicated group of diggers, including (in no particular order) Graham ‘Jake’ Johnson, Paul Brock, Nick Hawkes, Robin Taviner, Jonathon Riley, Mike Moxon, and Roz Simmonds, the discoveries made in Hallowe’en Rift could not have happened.

A special thanks to Professor Danielle Schreve (University of Bristol) who is always ready to give advice and confirm identification of faunal remains whenever asked.

Thanks are extended to the anonymous reviewer(s) whose comments added to the clarity of this report.

Map data copyrighted OpenStreetMap contributors and available from https://www.openstreetmap.org

References

Currant, A. and Jacobi, R. 2001. A formal mammalian biostratigraphy for the Late Pleistocene of Britain. Quaternary Science Reviews 20 (2001), Elsevier Science Ltd. 1707-1716

Jacobi, R and Currant, A. 2011. The Late Pleistocene Mammalian Palaeontology and Archaeology of Mendip. In Lewis, J. (Ed.) The Archaeology of Mendip: 500,000 years of continuity and change. Oxbow Books, Oxford, UK

Kearey, P. 2001. The New Penguin Dictionary of Geology, 2^nd Edition. Penguin Books

Scott and Buckingham. 2021. Stanton Mammoths and Neanderthals in the Thames Valley: Excavations at Stanton Harcourt, Oxfordshire. Archaeopress Archaeology

Simmonds, V. 2019. Evidence for Pleistocene frost and ice damage of speleothems in Hallowe’en Rift, Mendip Hills, Somerset, UK. Cave and Karst Science, Vol.46, No.2, (2019) 74-78. Transactions of the British Cave Research Association

Simmonds, V. 2021. A brief note on faunal remains from Hallowe’en Rift, Mendip Hills, Somerset, UK. Cave and Karst Science, Vol.48, No.3, (2021) 95-96. Transactions of the British Cave Research Association

Simmonds, V. 2024. Notes and observations relating to sediments containing ferro-manganese spherules found in Hallowe’en Rift, Mendip Hills, Somerset, UK. Cave and Karst Science, Vol.51, No.3, (2024) 129-133 Transactions of the British Cave Research Association

Yalden, D. 1999. The History of British Mammals. T & A.D. Poyser Ltd. London

All photographs by the author unless stated otherwise.

A representative selection of photographs:

[A similar version of this report was published in Cave & Karst Science, Vol. 52 (1), 2025, p.32-34. Transactions of the British Cave Research Association.]

Quarter Way Up Hole (QWUH) is a small cave in Fairy Cave Quarry, Mendip Hills, Somerset  (NGR ST 65529 47500). It is being dug actively by cavers in an effort to extend its current open length. It was during this process, early in 2025, that they uncovered some interesting and significant faunal remains, the digging team immediately sought advice and assistance. On 9^th February 2025, the excavation leading to the recovery of a woolly rhinoceros (Coelodonta antiquitatis) skull took place at the cave. The skull had been located within an unstable boulder ‘choke’ and, because it was lying in an especially vulnerable position, it was clearly evident that immediate recovery was the only safe option. Several other partial skeletal elements, including femora and humeri, had previously been recovered, and these together with the skull form a substantial assemblage.

Subsequently, the skull and other remains were cleaned carefully, to remove fine-grained sediments and granular material, after which they were allowed to dry slowly under controlled conditions. The skull and associated assemblage of faunal remains (including other species) have been deposited in the care of Professor Danielle Schreve at the University of Bristol. After an initial examination, Danielle has provided the following observations:

“The skull is absolutely magnificent, and I can confirm it is that of a large, male woolly rhinoceros. It’s also an old animal. Not only are its teeth very worn but it has a completely ossified nasal septum (the bony plate separating the two nostrils), which is a feature that gradually develops as the animal gets older, in order to strengthen the skull and support the very large nasal horn. Other features of the skull and shape of the nose also support this being an old individual and based upon comparison with the longevity of the extant white rhino, which has a similar morphology and grazing diet to the woolly rhinoceros, my provisional estimate is that the QWUH specimen was around 40 years old at time of death…

I haven’t looked in detail at the bones of the skeleton yet and there is work to be done refitting the fragments but there are parts of the major limb bones (humerus and femur) present. I am sure this must have been a whole animal that fell into the cave. It is very rare to find a complete individual in the fossil record and I only know of two other complete or partially complete woolly rhino skeletons in Britain, both of which were also accidental deaths…

The skull is beautifully preserved and covered with thin flowstone encrustation. It might be possible to get a Uranium series date on the stal, but the best initial option would be a radiocarbon date on the bone. For the time being, in terms of the geological age of the specimen, it is probably 40-50 000 years old and dates to the middle part of the last Ice Age…They are extinct in Britain by about 35 000 years ago.”

Currently, the recovered faunal assemblage from QWUH is undergoing more detailed study and conservation work, as mentioned above several of the fragmented bones refit.

Other sites on the Mendip Hills where woolly rhinoceros have been found include Gully Cave (Ebbor), Hyaena Den (Wookey Hole), Picken’s Hole (Compton Bishop), and Sandford Hill. Woolly rhinoceros, Coelodonta antiquitatis, has been recorded from the Lower Cave Earth deposits at Pin Hole, Creswell Crags, Derbyshire and has therefore been listed as part of the Pin Hole Mammal Assemblage Zone (MAZ), attributed to the Middle Devensian, MIS 3, c.59-24 ka. The Pin Hole MAZ also includes other large herbivores such as woolly mammoth, Mammathus primigenius,  steppe bison, Bison priscus, wild horse, Equus ferus, and reindeer, Rangifer tarandus (Currant & Jacobi, 2001). During the Middle Devensian conditions were generally cold and dry, although the period is characterized by sharply oscillating climates ranging between milder periods and short cooling episodes, in which dry grassland ‘mammoth-steppe’ environments were dominant (Webster, 2008).

Sources consulted

Andrew Currant & Roger Jacobi. 2001. A formal mammalian biostratigraphy for the Late Pleistocene of Britain. Quaternary Science Reviews 20 (2001) p. 1707-1716

Chris Webster. 2008. The Archaeology of South West England, South West Archaeological Research Framework (SWARF). Somerset County Council

Professor Danielle Schreve personal communications

https://www.mcra.org.uk/registry  Mendip Cave Registry & Archive Online search, accessed 18^th March 2025

A similar version has previously been published in Caves & Karst Science, Vol.51, No.3 (2024) pp.129-133. Transactions of the British Cave Research Association.

Abstract

During 2024 excavations in the area known as ‘Can of Worms’ in Hallowe’en Rift, an interesting sediment deposit was exposed and later sampled. A sub-sample was washed to remove the silt and clay fractions and revealed an abundance of ferro-manganese spherules. On closer examination, the spherules exhibited a concentric layered internal structure with several having a central, angular mineral grain around which the layers were formed, analogous to the growth of pearls in oysters or the formation of ooids, which are a type of carbonate or iron-coated grain, commonly with a cortex of fine laminae lacking biogenic features, and with a nucleus, such as a shell fragment or sand grain. Permafrost conditions on Mendip during the Pleistocene Epoch are thought to have penetrated to significant depth, with subsequent periods of thaw reaching lesser depths resulting in a deep impervious ice plug causing meltwater to be trapped and ‘ponded’. The ponded meltwater was ‘topped up’ by the ingress of surface-derived water, probably reflecting seasonal changes. The rise and fall of water created currents within the flooded cave, leading to agitation of particles and precipitation of minerals that coated particulates by a process analogous to that which creates ooids. The ‘pulsing’ of water also led to the deposition of rhythmites, which are observed in sediments throughout the cave.

Introduction

Hallowe’en Rift, at British National Grid coordinates ST 5353 4811 and altitude 148m above Ordnance Datum (aOD), has a current surveyed length exceeding 300m, with a vertical range of about 25m. The cave is located in a wooded hillside northeast of Wookey Hole Cave (Mendip Hills, Somerset). Initial excavation of the cave began in 1982, when the current access route was blasted open, but by the end of the 1980s interest at the site had waned. In the early-1990s, activity at the cave re-commenced briefly before other projects and interests took precedence. The current phase of activity began in 2009, and the exploration and investigation of the increasingly complex cave is ongoing (2024); several potential leads are actively being pursued.

Figure 1. Hallowe’en Rift survey. BCRA 5c by D. Price, R. Taviner and V. Simmonds

The ‘upper’ cave passages consist mostly of low passages leading to the more spacious area of An Unexpected Development (Fig.1). The ‘lower’ sections of passage, known as the Soft South, are more generally walking- to stooping-sized with several good exposures of sediment deposits. The cave passages are all formed within strata deposited during the Triassic Period, 251.9 to 201.3 million years ago. Most of the now accessible passages were completely or partially filled with deposits of sandy silt/clay, containing many scattered cobbles and boulders of dolomitic conglomerate and common clasts of fragmented speleothems, including broken stalagmite, stalactite, and flowstone; in places there is also evidence for bioturbation, in the form of root disturbance. U-series dates of speleothems range from c. 500 thousand years (ka) to 50 ka, indicating that the cave passages were open to some extent during that time. Sediment deposition within the cave passages occurred mostly late in the Pleistocene Epoch and in the Holocene Epoch. Most of the known passages have been made accessible by excavation. Recently discovered sections of open cave passage were found to contain abundant shattered speleothems and a variety of interesting geomorphological features relating to frost/ice damage within the cave (Simmonds, 2019). Currently, work continues to follow a large sediment-filled fossil passage trending towards northeast-to-east, leading from the base of a c. 3m deep pot, where a connection to Trick or Treat has been achieved and access to passages beyond is currently being pursued.

Sediments

During 2024, a change of circumstances prompted a review of the digging strategy and the development of an ‘alternative’ dig (now known as ‘Can of Worms’) leading approximately south-westwards from the ledge at the top of the pot in the Soft South. During digging activities, an interesting-looking sediment deposit was exposed (Fig.1). A bulk sample (about 1kg) was taken on 03/02/2024, from which a sub-sample weighing c.200g was later washed over a set of nested test sieves (2mm, 0.6mm, 0.063 mm), removing the silt/clay fraction and retaining the sand- and gravel-size particles (Fig.2 and Plate 1). Visibly evident in the residue was an abundance of iron/manganese (Fe/Mn) spherules. Most of the spherules were coarse-sand size (passing 2mm mesh, retained 0.6mm mesh) with a smaller quantity classified as granule-size (passing 6.3mm mesh, retained 2mm mesh).

Figure 2. Particle-size analysis, sample 006/COW (test sieve sizes: 2mm, 600-microns, 63-microns)

The presence of these Fe/Mn spherules had previously been noted in other sediment samples taken in Hallowe’en Rift although not in such high concentrations.

To investigate their mode of formation, a representative selection of spherules was set in a small block of epoxy resin and polished down to determine their internal structure. The resulting polished epoxy block revealed that most of the spherules examined had a concentrically layered internal structure. Several of the spherules also contained a central, angular, mineral grain around which the layers had formed, in a manner analogous to the growth of ‘pearls in oysters’ or the formation of ooids. Broadly in the current context, ooids are a type of carbonate or iron-coated grain with a cortex of fine laminae, lacking biogenic features, with a nucleus, most commonly comprising a shell fragment or a sand grain (Kearey, 2001). The term ‘pisoid’ (or ‘pisolith’) might also be used, because structurally they are similar to ooids, but pisoid is generally, applied to larger sizes (≥ 2mm in diameter), whereas ooids are normally considered as ≤ 2mm (Kalinina et al., 2024).

Plate 1. Sample: 006/COW, washed over 63-micron test sieve (ungraded). Abundant iron/manganese spherules are clearly evident ranging from sand- to fine gravel-size

Initially, it was thought that these spherules had been accumulated through turbulent eddies during multiple flood events occurring throughout the Pleistocene, the denser sand- and gravel-sized Fe/Mn particles being deposited as pockets or lenses. Subsequently, as energy and turbulence dissipated, the interstices between the coarse-grained particles becoming filled by silt and clay transported by water movement through the sediments over an undetermined period. However, it is now thought more likely that the laminated coarse- to fine-sediments (fining upwards) deposited in Hallowe’en Rift are a result of freeze/thaw ponding causing water to ‘pulse’ through the cave in a cyclical process (Bull, 1980).

Rhythmites  are finely laminated sediments in which two or three different lithologies are regularly repeated, they are common features of glacial lakes (Kearey, 2001). Permafrost conditions on Mendip during glacial periods throughout the Pleistocene period might have reached depths of ~80–100m, resulting in the blocking of Hallowe’en Rift (and other caves) by ice plugs. During warmer interglacials and interstadials, thawing might occur to a lesser depth, perhaps ~50m. Effectively the cave would still be ‘plugged’ by deeper ice, causing meltwater outflow and ‘ponding’ (Simmonds, 2019). The ponded meltwater ‘topped up’ with the ingress of surface-derived water, probably reflecting seasonal changes. The rise and fall of water creating currents within the flooded cave allowing for agitation of particles and the precipitation of minerals coating particulates in the same process that creates ooids. Precipitation is dependent upon the concentration of ions in solution and factors such as temperature and pressure. Water with high oxygen saturation is favourable for the formation and precipitation of ooids (Kalinina et al., 2024). Precipitation can occur when the temperature of the solution falls, when solute evaporates, or with changing chemical conditions in the solution. Saturation is the point at which a solvent can hold no more solute. Observations in Hallowe’en Rift confirm that ongoing ingress of water through surface soils into the cave readily precipitates carbonate, and that only short time-scales are required for this to occur. Water presently encountered in the cave originates via percolation, there are no sinking streams.

Cave sediments can be divided into two broad categories: clastic sediments and chemical sediments. Clastic sediments are moved mechanically whereas chemical sediments are formed in place, precipitated from solution in seeping, dripping, or flowing water. Clastic sediments can be further subdivided into materials that are derived locally within the cave (autochthonous) and those that are transported into the cave from the surface (allochthonous). Chemical sediments are subdivided into categories based upon their composition (White, 2007), including travertines, evaporites, phosphates, resistates, and ice. In the case of the Hallowe’en Rift spherules, clastic sediments provide the nucleus around which chemical sediments precipitate. Iron (Fe) and manganese (Mn) are commonly found in cave environments, especially as oxides and hydroxides. Their presence is generally indicated by characteristic colours, including reddish-brown, orange, and yellow in the case of iron, and dark brown to black in the case of manganese (Kotula et al., 2019). These compounds are essentially insoluble in neutral pH water and so are known as resistates (White, 2007), i.e. they are minerals that are resistant to chemical weathering. Black coatings that commonly occur to cover stream sediments are usually described as manganese oxide. These oxides appear as thin (<1 mm) coatings although thicker and more massive deposits are known to occur (White, 2007). In general, the clastic sediments found within Hallowe’en Rift comprise variably red to red-brown silt/clay abundantly layered with variably grey to grey-black silty sand containing many subangular or subrounded cobbles and boulders of the locally occurring (Triassic) Dolomitic Conglomerate and of calcite flowstone and fractured stalagmites and stalactites. There are also less extensive deposits of yellow and greenish-yellow, slightly silty, sand, though this layer generally lies towards the base of sediment sequences in the deeper sections of the cave, in some cases underlying a degrading calcite layer.

Chemical Analyses

The local host rock mostly comprises part of the Triassic Dolomitic Conglomerate. Most of the clasts within the conglomerate bedrock are of Carboniferous limestones and less commonly Devonian ‘Old Red Sandstone’; the rock clasts are also interspersed with various minerals. Weathering of the local bedrock releases iron, manganese, and other trace elements to be transported throughout the cave by water. Generally, any ‘metals’ released into caves from bedrock or allochthonous sediments appear as oxides and/or hydroxides. The mineralogy of Fe and Mn compounds in caves is also affected by the microclimatic conditions of the cave environment (especially humidity), which govern the hydration or dehydration of compounds. A wide variety of Fe and Mn compounds can be observed in caves developed in carbonate rocks, because these rocks and their weathering products are commonly enriched in minerals that contain Fe and Mn, including pyrite, hematite, rhodochrosite, marcasite, limonite, and goethite. Many fissures and caverns in karst limestones are filled or partially filled with red masses of iron oxides. For this reason, soils formed on limestone under some climatic conditions exhibit a characteristic red colour (Kotula et al. 2019). Fe and Mn compounds  are sensitive to changes in pH and Eh (oxidation/reduction) of water environments and oxidize when geochemical conditions change from anaerobic to aerobic, resulting in precipitation of minerals (Kotula et al, 2019).

Sample	As	Ba	Fe	Mg	Mn	Zn	pH	Carb.	SOM
(Samples = mg/kg, except pH = units; carbonate (Carb.) and soil organic matter (SOM) = %)
Surface	42	179	36500	3925	3147	487	6.4	3.8	0.9
Red/brown	47	160	43400	4468	1988	810	7.3	14.3	0.8
grey	54	361	52130	4724	7944	384	7.3	1.8	1.2

Note: for the purposes of comparison between Tables 1 and 2, 1 ppm = 1 mg/kg

The samples presented in Table 1 were taken at the surface to provide background data, and underground from successive layers about 5m east of the base of the cave entrance, in the south bank of the approach passage. It is noted that the analysis was limited to three samples.

Sample	As	Ba	Fe	Mg	Mn	Zn	Ca	Pb	Si
(all samples = PPM – parts per million)
1	<LOD	823	57004	10474	24710	387	3151	6352	37255
2	<LOD	324	13578	31275	10623	306	257	2342	30686
3	<LOD	1134	43118	7928	48279	543	4680	9132	31427
4	<LOD	1193	73518	8324	34602	514	3854	9251	29849

The results for portable hand-held X-ray fluorescence (XRF) analysis of the ferro-manganese spherules sampled are presented in Table 2, showing a representative selection of elements allowing for comparison with the elements analysed and the results presented in Table 1.  Underground, a mean soil pH of 5.1 was recorded for the fine-grained sediments in-situ at the sampling location (006/CoW) in Can of Worms.

X-ray diffraction analysis

X-ray diffraction (XRD) analysis was conducted at the School of Earth and Environment at the University of Leeds using the following methodology. The XRD scan was conducted on fine powder derived from a single spherule using a Bruker D8 with the scan range of 2-86deg two theta, step size of 0.02deg and speed of 1s/step. Bruker Eva Search/Match and the ICDD PDF database were used to match the observed peaks and thus identify the phases present.

Figure 3: Details of  the XRD scan and the ‘fits’ that the automated software generated (quartz, muscovite, and calcite as well as hematite). The fact that the quartz and muscovite peaks are larger than the hematite ones does not mean they are more abundant, just that they are more crystalline and so give stronger peaks (Professor Simon Bottrell, personal communication).

The XRD analysis indicates that the principal Fe mineral is hematite. Quartz is probably detrital, possibly providing the nucleus grain for the spherule, whereas the calcite, which is consistent with the carbonate content revealed by the above analysis (Table 1), could either be detrital or an authigenic precipitate. The muscovite is probably detrital mica that was incorporated as the spherule grew. Assignment of the mica as muscovite is a generalization and the material present could lie within the range of compositional range illite-muscovite (Professor Simon Bottrell, personal communication).

From the analytical results presented above, it is clear that the spherules exhibit elevated mineral/trace-element levels, especially Fe and Mn, as might be expected when compared to the background sediment samples, and due to the enriched mineral precipitation that occurs around a particulate core.

Comments

• Much of the cave passage found in Hallowe’en Rift has been formed along mineral veins and throughout the cave there is evidence of hydrothermal activity. Scallops in the deeper sections of cave are indicative of upwelling fluids.
• There are thick deposits of calcite throughout the cave, and the flowstones and other speleothems probably derive from supersaturated fluids. Supersaturation might be a consequence of the of cave passages at shallow depth being subject to more ‘aggressive’ surface waters that leached carbonate minerals from the host rock.
• Further samples taken from a section of cave beyond Trick or Treat also contained abundant ferro-manganese spherules confirming the wider distribution of spherules throughout the cave. They are especially common in ‘grey’ bands within rhythmite sediments. Scope remains to study sediments in the wider catchment area, to investigate whether the distribution of Fe/Mn spherules is more widespread.
• In future it is hoped that suitable speleothem sample(s) might provide U-series dating results and so contribute to establishment of a time-frame for sediment deposition.

It is acknowledged that the research to date is limited with regard to the number and scope of analyses conducted; therefore, a detailed interpretation is not presented at this stage. Hopes are that this can be addressed in the future by more sampling and analyses, leading to an increase in the quantity and value of the information available.

Acknowledgements

Special thanks to Professor Simon Bottrell and Lesley Neve, at the School of Earth and Environment, University of Leeds for providing the XRD analysis and in doing so adding valuable information about the composition of the Fe/Mn spherules.

My thanks also the BCRA referees for their constructive and helpful comments, which helped to improve the clarity of this Report.

Without the commitment, determination, and camaraderie of a dedicated group of diggers, including (in alphabetical order) Paul Brock, Nick Hawkes, Graham Johnson, Mike Moxon, Jonathon Riley, and Robin Taviner, the discoveries made in Hallowe’en Rift would not have happened.

Reference sources consulted

Bull, P.A. 1980. Towards a reconstruction of timescales and palaeoenvironments from cave sediment studies. In Cullingford, R.A., Davidson, D.A. and Lewin, J. (1980) Timescales in Geomorphology. John Wiley & Sons Ltd.

Huang, L. 2022. Pedogenic ferromanganese nodules and their impacts on nutrient and heavy metal sequestration. Elsevier, Earth-Science Reviews 232 (2022) 104147 https://doi.org/10.1016/j.earscirev.2022.104147

Kalinina, N. A., Rudmin, M. A., Sherstyukov, M., Maximov, P., and Kerimov, A-G. 2024. Origin of iron-rich minerals ooids and pisoids in the Jurassic ooidal ironstones of the Labino-Malkin region (Caucasus). Journal of Palaeogeography, 2024, 13(3) : 475-494

Kearey, P. 2001. The New Penguin Dictionary of Geology, 2^nd Edition. Penguin Books

Kotula Piotr, Andreychouk Viacheslav, Pawlyta Jacek, Marynowski Leszek, Jendrzejewska Izabela. (2019). Genesis of iron and manganese sediments in Zoloushka Cave (Ukraine/Moldova) as revealed by δ13C organic carbon. “International Journal of Speleology” (2019, iss. 3, s. 221-235). https://digitalcommons.usf.edu/ijs/vol48/iss3/1/

Simmonds, V. Evidence for Pleistocene frost and ice damage of speleothems in Hallowe’en Rift, Mendip Hills, Somerset, UK. Cave and Karst Science, Vol.46, No.2, (2019) 74-78. Transactions of the British Cave Research Association

White, W.B., “Cave Sediments and Palaeoclimate” (2007). KIP Articles. 784. https://digitalcommons.usf.edu/kip_articles/784