Pickhill Mound, North Yorkshire

Pickhill, North Yorkshire.


The hill is called Picts Hill. The village – Pickhill takes its name from this hill. Pickhill is in the Domesday book.

On the 1st edition OS, there is a long mound and a short mound marked in the field next door, as well as some “mound foundations” and an Earthwork. None of these are easily spotted today.

“A large artificial mound here, apparently raised for defensive purposes, bears the name of Picts’ Hill, and an improbable belief prevails that the Picts defeated the Romans in battle at a spot, not far off, called Roman Castle. This mound is also known as Money Hill, but, though partially cut away for the construction of the railway, the traditional hidden treasure was not found.”

Bulmer’s History and Directory of North Yorkshire (1890)

“There is a large artificial mound at Pickhill called Pict’s Hill (or Money Hill), which now forms part of the railway embankment. Mr Longstaffe gives the following particulars, for which, he says, he is indebted to his friend M. M. Milburm, Esq., land agent, Thirsk.

Mother Shipton is said to have prophesied that Pickhill would never thrive till a certain family became extinct, and Picks, or Money Hill was cut open. Some years ago an old man dreamed that there was an archway in the hill, beneath which was a black chest, with three locks, containing the money which gives the name to the mound.

Well, the family did become extinct in 1850, and the Leeds and Thirsk railway Company cut the hill open in 1851 – though it naturally formed part of its embankment, and their line passes over it. Still the directors ordered it to be excavated – the old man, the dreamer, was still alive, and pointed out the spot where the archway lay.

The hill was cut through in all directions, but nothing was found, save in the foss, where portions of tile and a small brick, both vitrified on one side, and fragments of urns, and a piece of thin iron, like the crest of a helmet, were discovered. The mound was squarish, 113 by 80 feet, and surrounded by a moat.” Wellans History and Topography of York and North Yorkshire. 1859.

Scorton Cursus, North Yorkshire

Scorton Cursus, North Yorkshire

Scorton Cursus marked in blue on the 1850’2 OS 1st Edition map.

The cursus was originally about 2.1km long and aligned SE-NW. Clustered round the monument were a number of ring ditches, one that was excavated had a single burial with a beaker. This would date the site as being in use from around 3,500BC until at least the Bronze Age c.2,000 BC.

The cursus was cut almost dead straight over it’s entire length and was 37.2m wide. The cursus was constructed by digging two parallel ditches 4.10m wide on the east side and 3.90m wide on the west.

Cursus monuments are thought of as “ceremonial ways” – avenues, carved though the landscape that created a sacred or ritual space. Often these are associated with water, and Scorton is no exception – it sits close by the River Swale and also sits in a remnant alluvial flood plane. At the time when Scorton Cursus was built, it is likely that it sat amongst a large number of lakes. These monuments are often given their “ritual” function due to the fact that burial mounds tend to be placed in close proximity to them.

Almost the entire site was destroyed by quarrying, very little record of it exists yet it is one of the first and largest “communal ritual structures” in the north of england.

Research Notes

“The excavation of a section accross the complete width of the cursus, now reduced to a cropmark, produced evidence from recut ditches of at least two major phases of use. The first was indicated by narrow almost V-profiled ditches, which, having silted, were recut to a wide and shallow form. A low mound of indeterminate shape, possibly Upcast from the ditches, was situated in the centre of the cursus, while a pit complete with a ramp was discovered as an external feature of the eastern ditch. A sampling strategy recovered a flaked stone assemblage of flint and chert, which included debris from knapping activities in addition to actual tools. In the area excavated no evidence was recovered for any recognisable form of occupation of any period.” Peter Topping 1978.

“A section cut accross the cursus by P. Topping of Newcastle University showed that both ditches had been re-cut from a norrower almost V-shape to a wide shallow form. A truncated post hole was found dug into the silting of the eastern ditch. A large low mound apparently lay within the ditches at the north west end of the cursus. There was also evidence here os external pits possibly holding upright timbers” YAJ 51, 1978

“NZ240005 Cropmarks revealed in aerial photographs taken by Prof. J. K. F. St. Joseph are threatened by gravel workings. The Richmondshire Excavation Group, directed by Mrs F. Thubron for the DOE sectioned the cursus mark to expose two double ditches 32 m apart, 1 m deep, and from 2-3m wide. Two of these showed evidence of re-cutting. No dating evidence was found. Several sherds of Roman pottery was picked up from plough soil near the circular cropmark.” from Yorkshire Arch. Journal, Vol 48, 1976, P.2.

“The site was discovered by Prof J.K. St Joseph in 1949, and from his series of aerial photographs it was possible to trace the course of the cursus for approximately 2.1 kilometers. Evidence from soil stripping and exacavations has shown that the monument extends even further to the north-west, the north west terminal not as yet having been discovered. The south west terminal, which shows clearly on the aerial photographs, consists of a straight transverse ditch which joins the two main ditches at right angles. Clustering around it was a series of ring-ditch crop marks. The aerial photographs also show a series of bleach marks between the ditches at the southern end of the cursus, which may represent a series of contiguous mounds. This area of the cursus also features what appears to be smaller outer ditches, although they may be restricted to the southern end of the cursus as they were absent in the excavated area.
Another noteworthy feature brought out by aerial photography is the accuracy with which the ditches have been laid out, so they are remarkably straight considering the distance over which they extend”.

“from the evidence available at Rudston, it would appear that cursus monuments in Yorkshire developed during the late Neolithic, and flourished, as can be seen at Rudston in its magnificent complex of monuments, into the Early Bronze Age. It is within this local chronological framework that the cursus at Scorton must have developed, although not to as great an extent as the important centres of Rudston or Thornborough”. Excavation at the Cursus at Scoton North Yorkshire 1978 Peter Topping, Yorkshire Archaeological Journal, Volume 54, 1982

Kirklington Tumulus, North Yorkshire

Kirklington Tumulus

“Prehistoric vessels dug out of the mound at Stapely Hill, Kirklington, in 1903. Fragments of several pottery urns of the Bronze Age, C. 1,000 B.C., one containing cremated human bones” Description and photo’s from Kirklington Church.

“SE 326828 S. White reports that a polished flint axe was found by Mrs J. Fothergill in 1976. The axe is in perfect condition with a cream patination, 14 cm long, 4.4 cm thick and 6 cm wide at the blade. It is of red coastal flint. An examination of the field in November 1977 produced no further finds. In the possession of the finder.” YAS Vol 50, 1978, P.8.

Devil’s Arrows, North Yorkshire

The Devil’s Arrows – BoroughBridge

Introduction

This Bronze Age site comprises of three large standing stones, it is thought originally there were as many as five stones in this alignment. Being Bronze Age little is known about the origin of the Devil’s Arrows, the name reflecting a more recent myth. The monument is strongly linked with an alignment withseveral others covering a line of over 50 miles and heading north south through North Yorkshire.

The site itself, being just off the A1 has been an important communications route for several thousand years and it is likely that in the Iron age the Brigantes adopted this as a centre for religious or tribal gatherings. It’s importance is further indicated by the creation of the tribal capital of Isurium Brigantium in c. 120AD less than three miles away at Aldborough.

Research so far has been limited to a site visit and scans for document records.

In the Brigantium context it is likely that this site, together with the triple Henge alignment and cursus of Thornborough some 8-10 miles to the north of the Arrows formed one of the most significant religious and tribla gathering point in Brigantia.

Location

Grid Ref: SE391665, Latitude: 54.092736, Longitude: -1.402114.

Three naturally shaped stones in an alignment thought to have originally included up to five stones. These are almost in a straight line, pointing north to south. The outer stones are 200 and 370ft away from the central stone.

Access to these is relatively simple, they are just off the side of an elevated section of the A1(M) trunk road at Jct 48 to Boroughbridge.

The stones have heights of 18ft (northern stone), and 22ft. They are of millstone Grit which was quarried at Knaresborough, 6.5 miles to the SW. The grooving is the result of weathering.

The Devils Arrows are the centre of the most important alignment of standing stones, henges and other remains in Yorkshire, all running N/S alongside the ancient line of the A1, Stretching from Hutton Moor to Thornborough and beyond. Current estimates indicate they were erected during the Bronze Age c. 2,200 – 1,400 BC.

Views of the northern and central stone

Left, central stone and northern stone, right.

southern-most stone

Other research

The name stems from a legend traced back to 1721, where the Devil was supposed to have thrown the stones, aiming at the next town of Aldborough. He stood on Howe Hill and shouted,

“Borobrigg keep out o’ way,

For Aldborough town

I will ding down!”

But obviously his aim was not that good, and so they landed short of their mark.

At their maximum of 22 feet in height, The Devil’s Arrows are the tallest standing stones in the United Kingdom apart from the Rudston Monolith.


Since Antiquarian William Stukeley’s time it has been believed that the arrows are in a straight line, running North to South. The fact is that they’re visibly not. When they were dragged across from miles of countryside from Knaresborough (seven miles away), the stones where constructed to not be in line, but to be slightly “westward.”

In the 1970’s Paul Devereux wrote in The Ley Hunter’s Companion that “the functions of the monoliths was to act as a multi-directional sighting or reference instrument.” Devereux also quotes G Bernard Wood on “the Devil’s arrows stand in line “with an ancient ford across the River Ure.”

The Devils Arrow alignment. Ref: Lines on the Landscape, Devereux and Pennick

This place is remarkable for those monuments called the Devils Arrows, but whether Roman or British, is uncertain. “Here was, in the British times,” says Dr. Stukeley, “the great Panegyre of the Druids, the Midsummer meeting of all the country round, to celebrate the great quarterly sacrifice; accompanied with sports, games, races, and all kinds of exercises, with universal festivity. This was like the Panathenian, the Olympian, Nemean meetings, and games among the Grecians. These obelisks were as the Metae of the Races; the remembrance hereof is transmitted in the present great Fair held here, on St. Barnabas Day.”

In Leland’s time there were four, but in the seventeenth century, one of them was pulled down; the remaining ones are placed at unequal distances from each other. The tallest one is 30 feet 6 inches from the bottom, about 6 feet of which are buried in the ground; its greatest circumference 16 feet.

Richard Frank, a singular traveller, and famous peripatetic angler, in his tour to the northern parts of Scotland, to enjoy his favourite amusement, which he published in 1694, says that he saw near Boroughbridge, seven of these stones, in which he must have been mistaken, as it is not likely that they have increased since the days of Leland. Evident marks of the chisel appear below the surface of the earth. It is of the common coarse rag stone or mill grit; a large rock of this stone from which, probably these obelisks were taken, is at Plumpton, near Knaresborough. Doctor Stillingfleet considers them as British Deities: Leland, Camden, and Drake, suppose them to have been the work of the Romans, and erected by that people as trophies, to commemorate some important victory.

Near this place, in 1322, that unfortunate Prince, Thomas Earl of Lancaster, with some of the nobility, disgusted with the royal favourites, the Spencers, made stand against the forces of his nephew, Edward II. but was taken by Sir Andrew de Harcla, who, being insensible to entreaties and solicitations, and after suffering every possible indignity that cruelty could suggest, was mounted on a sorry horse, and brought before the King, who ordered, without any form of trial, his head to be struck off, on an eminence near Pontefract. One of his partisans, the powerful John de Bohun, Earl of Hereford, in passing over the bridge, then made of wood, was run through with a spear, by a soldier, cowardly placed beneath for that execrable purpose. It sends two Members to Parliament, a privilege it derived from Queen Mary in 1553.

[From Langdale’s Topographical Dictionary of Yorkshire. (1822)]

Mote of Mark, Dumfries

 

[caption id="" align="aligncenter" width="1024"]Mote of Mark from sea shore pathMote of Mark from sea shore path – geograph.org.uk – 6273954” by Andrew Curtis is licensed under CC BY-SA 2.0

Mote of Mark – a Dark-Age citadel above Rough Firth

Setting & basic layout

Located on a granite knoll (45 m OD) on the east shore of Rough Firth between Rockcliffe and Kippford, Dumfries-and-Galloway (NGR NX 845 540). The west and south faces drop almost sheer to the estuary; access is by a narrow neck on the north-east. (hillforts.arch.ox.ac.uk, Britain Express)

Defences – A single timber-laced stone rampart, c. 4 m thick, once ringed the 0.14 ha summit; most blocks were tumbled downslope after a fierce burning that fused parts of the core into green-black glass. A slighter outer bank and ditch skirt the easier north-east approach. No definite entrance has been located. (Canmore)

Chronicle of investigation

Date Investigators & method What they added
1755–1893 Roy’s Military Map, R. Riddell (1790) & J. Coles (1893) sketch-survey First published notice of a “vitrified fort”; rough plan and section. (hillforts.arch.ox.ac.uk)
1913 Alexander O. Curle cut 13 trenches across rampart and interior Proved timber-lacing + vitrification; recovered continental glass, E-ware pottery, 400+ clay mould fragments, crucibles and high-status metalwork, revealing industrial activity. (journals.socantscot.org)
1973 & 1979 Lloyd Laing & David Longley reopened Curle’s area and trenched the N & S walls Produced a full stratigraphic sequence, mapped rampart faces, identified five structural/occupational phases and sampled vitrified slag. (books.casematepublishing.com, hillforts.arch.ox.ac.uk)
2006-13 Watching briefs, bracken-die-back surveys & UAV imagery (HES) Monitored erosion, located terrace platforms below the summit. (Canmore)
2022 Publication of The Mote of Mark monograph Synthesised all finds, provided new scientific dating and specialist analyses. (books.casematepublishing.com)

Key finds & specialist results

Imported table-wares – 55 sherds of Gaulish E-ware and two Late-Roman (LR 2) amphora fragments place peak occupation in the mid-6th century AD. (books.casematepublishing.com, Canmore)

High-status craft debris – 482 clay mould fragments (Penannular brooches, enamel studs), crucible slag and bronze/iron off-cuts indicate on-site non-ferrous metal-working aimed at élite goods. (books.casematepublishing.com)

Glass & gaming pieces – Vessel shards from Frankish glass beakers and a bossed glass gaming counter underscore long-distance connections. (books.casematepublishing.com)

Animal bone & food waste – Dominance of cattle and high proportions of red-deer venison fit a short-lived, high-status residence rather than a farming hamlet. (books.casematepublishing.com)

Dating & historical horizon

Radiocarbon assays on rampart charcoal and occupation layers converge on c. AD 550–700; artefact typology agrees, framing the fort within the post-Roman kingdom of Rheged and the wider Irish-Sea trading zone. (hillforts.arch.ox.ac.uk, books.casematepublishing.com)

Why the Mote matters

Classic vitrified wall south of the Clyde/Forth line – a laboratory for studying firing techniques beyond the better-known Highland forts. (Canmore)

Industrial powerhouse – unparalleled quantity of moulds and crucibles shows that prestigious metal-working was embedded inside a royal seat, not farmed out to satellite workshops. (journals.socantscot.org, books.casematepublishing.com)

Trade cross-roads – Imported wine amphorae, fine pottery and glass prove direct contact with Atlantic Gaul and the Mediterranean during Britain’s so-called “Dark Ages”. (Canmore, books.casematepublishing.com)

Tightly dated destruction – Coherent 6th-century radiocarbon suite plus vitrification raise the prospect that the fort was deliberately torched during early Northumbrian expansion. (hillforts.arch.ox.ac.uk)

Outstanding questions & research potential

Issue Why it matters Next step
Who burned the rampart? Could link the fire to named conflicts in Historia Brittonum. Pair archaeomagnetic & micro-CT slag studies to refine burn episode.
Extent of craft zoning Interior still partly unexcavated. Targeted geophysics and micro-excavations in central hollow.
Outer terrace platforms Possible worker huts or later reuse? Coring & OSL dating of terrace fills.
Landscape integration How did the fort control estuary traffic? Viewshed & catchment modelling tied to LiDAR.

Vitrified Fort

The Mote of Mark is a defended hilltop overlooking the Urr estuary. It was the court or citadel of a powerful Dark Age chieftain, possibly one of the princes of Rheged. The site was occupied during the 6th century and appears to have been destroyed by fire in the 7th century.

The top of the hill was enclosed by a massive stone and timber rampart. Inside was a timber hall surrounded by a huddle of workshops and stables. This was a wealthy site with trading contacts across Europe. Finds from the excavations include glass beads and wine jars from central France and glassware from Germany. Local craftsmen produced elegant bronze jewellery in a distinctive Celtic style.

The tumbled remains of the ramparts can still be seen, and an on-site interpretation panel has an atmospheric reconstruction of the fort.

size: 8 ha (20a)

Legendary and Literary Background – https://panther.bsc.edu/~arthur/others.html
This fort was occupied from the 5th to 7th centuries, smack dab in the Arthurian time frame. At the pinnacle of its prominence, it was a well-fortified trading and manufacturing center. Excavations in 1913 and 1973 unearthed a large, circular timber hut and evidence of metalworking. These people seemed to have imported raw materials–iron from the Lake District and jet from York–to produce interlaced jewelry, brooches, and sundry metalwork. They imported luxuries as well–pottery from Bordeaux and glass from the Rhineland were found. Such prosperity suggests that this fort may have been the stronghold of a smaller British subkingdom.

The primary defences consisted of stone and timber walls, and there was a timber gate for the main entrance on the southern slopes. In the 7th century, though, these defences failed. The outer wall shows evidence of vitrification, a condition when extreme heat causes stones to fuse together. Many believe that this was the result of an attack by the Angles–Anglian runic inscriptions were found at the site–though some say that the walls were purposely vitrified to strengthen them.

The only thing truly connecting this fort with the Arthurian legend is the name, its period of occupation, and its proximity to Trusty’s Hill.

Vitrified Forts Distribution

Vitrified Forts Geographic Distribution

One of the great mysteries of classical archaeology is the spartan worldwide distribution of vitrified forts, except for Scotland and France. However, there are a number in the wider world, especially Europe. Here is the current breakdown:

Global (really pan-European) picture of vitrified hill-forts

To date, all firmly identified vitrified forts lie in western and Northern Europe. More than 200 sites are accepted; c. 70 of them are in Scotland—the densest cluster anywhere. No authenticated examples are recorded in the Near East, Africa, the Americas or Australasia.

Region / country Approx. number & key examples Notes on distribution & research status
Scotland ≈ 70 sites from Galloway to Caithness. Flagships: Tap o’ Noth, Craig Phàdraig, Dunnideer, Dun Deardail, Finavon, Burghead. Scottish forts were the first recognised (19th c.); systematic surveys began in the 1930s and continue. (Amusing Planet)
England 3 certainly vitrified: Wincobank (South Yorks), Almondbury/Castle Hill (W. Yorks), Castercliff (Lancs). Localised fusion in a few Cumbrian ring-works is still debated. (Nature)
Wales & Isle of Man Isolated, limited vitrification reported at Pen Dinas (Ceredigion) and South Barrule (Man), but not as extensive as the Scottish walls. Requires modern material’s analysis; both often cited but thinly published.
Ireland Fewer than ten candidate sites; best-known are Dun Aengus (Aran Is.), Knocknashee (Sligo) and possibly Creggankeel. Vitrification tends to be patchy. Field confirmation lagging; dating work only recently begun.
France c. 20 forts, concentrated in Brittany (e.g., Le Yaudet), Normandy and Burgundy (e.g., Château de la Roche). Roland Comte, however, has suggested that there may be many more. French excavations since the 1980s apply high-temperature petrology; some walls show repeat vitrification events.
Iberian Peninsula Portugal: Citânia de Sanfins, Castelo Velho de Freixo; Spain: some Oppida in Galicia and Asturias show fused masonry. Iberian examples, often linked to the castro culture; fewer laboratory studies than in France/Scotland.
Germany & Switzerland Germany: Schloßberg (Schramberg), Kienberg (Black Forest). Switzerland: Mont Vully rim. Generally single-rampart hill-top enclosures; vitrification limited to beam-slots. (Nature)
Scandinavia Sweden: Broborg (Uppland) and Götavirke sector walls; Norway: uncertain cases in Østfold. Swedish investigations demonstrate temperatures ≥ 1 050 °C identical to Scottish samples.
Central Europe (Bohemia, Slovakia, Hungary) Small number (< 10) of hill-forts with slag-like fused zones, e.g., Švédské šance (CZ), Liptovská Mara (SK), Somló (HU). Continental research is scattered; many identifications rest on visual inspection rather than thin-section work. (ResearchGate)

Key distribution patterns

  • Atlantic/North-Sea façade cluster: Scotland, Brittany and southern Sweden share the densest concentrations, suggesting cultural transmission along maritime routes during the mid-first-millennium BC.
  • Sparse but widespread inland occurrences: Single forts in Germany, Bohemia and the upper Danube imply that vitrification was known but never commonplace on the mainland.
  • Absence outside Europe: Claims of vitrified “desert forts” or Asian walls have not survived petrological scrutiny; temperatures reached in those structures derive from natural lightning or later lime-kiln intrusions, not deliberate timber-lacing fires.

Research gaps

  • Analytical parity: Scotland, France and Sweden have applied SEM, XRD and Lead-isotope analysis; many Iberian, Irish and Central-European sites still rely on macroscopic description.
  • Chronological resolution: Fewer than 25 forts have direct radiocarbon or OSL dates for the firing episode; comparative regional chronologies therefore remain coarse.
  • Functional explanation: Whether vitrification was accidental, tactical, or ritual continues to be debated across the entire distribution area; consistent excavation strategies are needed if we are to compare motives between the Scottish core and outlier regions.

In sum, vitrified forts form a European phenomenon centred on Scotland but echoed in discontinuous belts from Iberia to Scandinavia, with isolated outliers deep in Central Europe. Their patchy yet recognisable distribution hints at shared construction methods—and perhaps shared symbolic practices—diffusing across Iron-Age Europe, never extending beyond it. (Nature, ResearchGate, Amusing Planet)

Commentary

Scotland

There are at least 70 such forts throughout Scotland. Among the most well-known are Dunnideer, Craig Phadraig (near Inverness), Abernathy (near Perth), Dun Lagaidh (in Ross), Cromarty, Arka-Unskel, Eilean na Goar, and Bute-Dunagoil on the Sound of Bute off Arran Island. Another well-known vitrified fort is the Cauadale hill-fort in Argyll, West Scotland.

Others include, Dun Mac Uisneachain (Dun Macsnoichan), the ancient Beregoiium, about 9 m. N.N.E. of Oban; Tap o’ Noth, in Aberdeenshire; Craig Phadraic, or Phadrick, near Inverness; Dun Dhardhail (Dunjardil) in Glen Nevis; Knockfarrail, near Strathpeffer; Dun Creich, in Sutherland; Finhaven, near Aberlemno; Barryhill, in Perthshire; Laws, near Dundee; Dun Gall and Burnt Island, in Buteshire; Anwoth, in Kirkcudbright; and Cowdenknowes, in Berwickshire. Dun Mac Tjisneachain is the largest in area, being 250 yds. long by 50 yds. broad. In Barryhill and Laws, the remains of small rectangular dwellings have been found.

The evidence from elsewhere shows very few vitrified forts elsewhere, indeed the total number of vitrified forts worldwide is thought to be less than 100. Some examples are as follows:

France

Vitrified forts in France are discussed in the American Journal of Science (vol. 3, no. 22, 1881, pp. 150-151) in an article entitled “On the Substances Obtained from Some ‘Forts Vitrifiés’ in France”, by M. Daubrée. The author mentions several forts in Brittany and northern France whose granite blocks have been vitrified. He cites the “partially fused granitic rocks from the forts of Château-vieux and of Puy de Gaudy (Creuse), also from the neighbourhood of Saint Brieuc (Côtes-du-Nord)”. Daubrée, understandably, could not readily find an explanation for the vitrification.

Vitrified stones, hillfort in Highland, Scotland, UK, Rubh' Ard Ghaunsgail

Vitrified stones, hillfort in Highland, Scotland, UK, Rubh’ Ard Ghaunsgail

Very vitrified. Possibly part of the pictish fort that was here.

Very vitrified. Possibly part of the pictish fort that was here.

Turkey

Similarly, the ruins of Hattusas in central Turkey, an ancient Hittite city, are partially vitrified. The Hittites are said to be the inventors of the chariot, and horses were of great importance to them. It is on the ancient Hittite stelae that we first see a depiction of the chariot in use. However, it seems unlikely that horsemanship and wheeled chariots were invented by the Hittites; it is highly likely that chariots were in use in ancient China at the same time.

Iran

Some of the ancient ziggurats of Iran and Iraq also contain vitrified material, sometimes thought by archaeologists to be caused by the Greek fire. For instance, the vitrified remains of the ziggurat at Birs Nimrod (Borsippa), south of Hillah, were once confused with the Tower of Babel. The ruins are crowned by a mass of vitrified brickwork–actual clay bricks fused together by intense heat. This may be due to the horrific ancient wars described in the Ramayana and Mahabharata, although early archaeologists attributed the effect to lightning.

Other locations

Vitrified forts have also been found in Yorkshire and Lancashire, in England; Londonderry and Cavan, in Ireland; in Upper Lusatia, Bohemia, Silesia, Saxony and Thuringia; in the provinces on the Rhine, especially in the neighbourhood of the Nahe; in the Ucker Lake, in Brandenburg, where the walls are formed of burnt and smelted bricks; in Hungary.

Why Vitrify a Fort?

Why Vitrify a Fort?

Why was it done and by whom?

Practical reasons?

Originally, it was thought that the forts had become vitrified due to an enemy attack. A theory proposed by Childe in the 1930 thought it was that it was invaders, not the builders, who were assaulting the forts and then setting fire to the walls with piles of brush and wood; however, it is hard to understand why people would have repeatedly built defences that invaders could destroy with fire, when great ramparts of solid stone would have survived unscathed. Also, this theory does not stand up to the geographic distribution of hill forts versus the known warring area where hill forts were in use. For example the south of England suffered wave after wave of hostile invasion from other Gaulish tribes, yet no vitrification has been noted – surely if it was a natural effect of a battle then these forts would be more likely to occur in the south of England (given the large concentration of timber laced ramparts and the frequency of fighting in the area).

This idea was amended with the theory that the builders of the walls had designed the forts in such a way that the vitrification was purposeful in order to strengthen the walls. This theory postulated that fires had been lit, and flammable material added to produce walls strong enough to resist the invading armies of the enemy. It is an interesting theory, but one that presents several problems. The main problem with this theory is there is no indication that such vitrification actually strengthens the walls of the fortress; rather, it seems to weaken them. In some cases, the walls of the forts seem to have collapsed because of the fires, however this may show an error in the calculations of the builders.

To further illustrate this point, Julius Caesar described a type of wood and stone fortress, known as a murus gallicus, in his account of the Gallic Wars. This was interesting to those seeking solutions to the vitrified fort mystery because these forts were made of a stone wall filled with rubble, with wooden logs inside for stability. Caesar notes how the flexibility of the wood adds to the strength of the fort in case of battering ram attack.

Some researchers are sure that the builders of the forts caused the vitrification. Arthur C. Clarke quotes one team of chemists from the Natural History Museum in London who were studying the many forts:

“Considering the high temperatures which have to be produced, and the fact that possibly sixty or so vitrified forts are to be seen in a limited geographical area of Scotland, we do not believe that this type of structure is the result of accidental fires. Careful planning and construction were needed.”

Our own research into how forts were vitrified does indeed suggest a deliberate and planned action on behalf of the builders. Looking at the hill forts in evidence today, it appears that although many seem to be specifically built in strategic locations, some do not take full advantage of the natural defences available. Another common feature is that many vitrified forts have two rings of ramparts, but only the inner is vitrified. This suggests that the vitrified rampart was for the benefit of the fort’s users (not visible outside the fort) and that although linked with battle, the vitrification served a purpose other than strengthening the fort.

Beyond this reasoning, any further comment is pure speculation, but our research does indicate that the vitrification process could have been part of a lengthy ‘ceremony’ and will have been directed by the most powerful members of the community. Although it is strongly felt that vitrification of forts represents a cultural or religious element, further comment is reserved until further investigative work can be performed.

Experimental attempts to replicate the vitrification seen at Tap o’ Noth

The rampart of Tap o’ Noth hill-fort is one of the best-preserved examples of a stone wall whose core has melted and fused into green-black glass. Because the temperatures required (> 1 000 °C) far exceed what an uncontrolled brush-fire can achieve, archaeologists have tried—twice and at full scale—to discover how such heat might have been generated and whether the melting was accidental, structural or deliberate.

Date & lead investigator Where the replica wall was built Construction details Firing regime & peak temperature Result Main lessons
1937 – V. Gordon Childe & W. Thorneycroft Plean Colliery, Stirlingshire 3.7 m long, 1.8 m wide, 1.8 m high “murus gallicus”: fire-clay bricks as faces, 30 cm timber lacing, basaltic rubble core. (the urban prehistorian) 4 t of kindling & logs piled against both faces; fuel replenished for 20 h in sleet. Wall collapsed inward after 5 h; core stones reached c. 900 °C; 3 kg of bubbly glass recovered. Timber-laced stone can indeed vitrify, but only where flame and oxygen reach the rubble; a single firing would not strengthen a whole hill-fort but could melt patches.
1980 – Ian Ralston (Yorkshire TV / Aberdeen Univ.) East Tullos landfill, Aberdeen 8 m long rampart based on Tap o’ Noth section: outer skin of granite blocks; gabbro rubble core; horizontal oak beams (≈ 30 % timber by volume). (the urban prehistorian) Continuous pyres plus paraffin & animal fat; wind-shielding tarps; fire stoked for 28 h. Core probe after 15 h read 1 050 °C. Partial vitrification in beam sockets and mid-core; 3 kg of fused granite-gabbro glass. Wall bulldozed for safety after 28 h while still hot. Confirmed Childe’s results; showed that > 1 000 °C can be reached in a timber-rich rampart without industrial bellows; but vitrification remains patchy and weakens the structure.

Key technical findings

Fuel-to-stone ratio – Both experiments needed c. 1 part dry timber to 2–3 parts stone by volume. That equates to many thousands of mature trees for a full hill-fort wall: mass felling and organised labour are implied.

Airflow control – Heat concentrated where through-drafts fed the core (beam sockets, gaps between facing stones). Stagnant pockets never melted.

Glass chemistry – Melted granite and gabbro at 1 050 °C produced the same green-black glass and vesicles seen in Tap o’ Noth samples, vindicating the experimental design.

Behavioural implications for Tap o’ Noth

Deliberate conflagration is feasible: a well-planned firing could vitrify selected stretches in one episode, whether as an act of destruction or as a dramatic closure rite.

Structural strengthening unlikely: both replica walls became unstable once timber burned out; vitrification weakens rather than “welds” a rampart.

Labour & resource cost: harvesting and hauling the timber needed for a 100 m-long, 3 m-thick rampart would require a supra-household workforce—fitting a scenario of elite display or punitive destruction, not accidental fire.

Unresolved questions

Who lit the fire? —No siege debris or arrowheads were found in Tap o’ Noth’s vitrified tumble; ritualised self-burning remains as plausible as enemy attack.

Why re-occupy a melted wall seven centuries later? —Pictish re-occupation (7th c. AD) suggests the ruined, glass-fused rampart still carried prestige or ancestral authority despite its weakened fabric.

Regional pattern: comparable vitrification and radiocarbon brackets at Craig Phadrig, Dunnideer and Dun Deardail hint at a shared cultural practice of “fire-finishing” forts in north-east and Highland Scotland.

Take-away: the Tap o’ Noth experiments demonstrated that Iron-Age builders could intentionally melt a timber-laced wall using only locally available fuel and simple draught control. They did not prove whether the goal was tactical, structural or symbolic—but they push the balance of probability toward a planned, labour-intensive fire that turned a granite rampart into a smoking, glassy monument of power.

Cover image for the YouTube Glass Castles video
Unwanted cookie prevention
Clicking on this image will load the YouTube video, and you will get exactly the same cookies as if you were on the YouTube website directly

Vitrified Fort References

Cook, M., Watson, F., and Cook., G. (2016). Burning Questions: New Insights into Vitrified Forts. In Erskine, G., P. Jacobsson, P. Miller, and S. Stetkiewicz (EDS.).

Proceedings of the 17th Iron Age Research Student Symposium, Edinburgh. Oxford: Archaeopress Publishing.

Friend, C.R., Kirby, J.E., Charnley, N.R. and Dye, J., 2016. New field, analytical data and melting temperature determinations from three vitrified forts in Lochaber, Western Highlands, Scotland. Journal of Archaeological Science: Reports, 10, pp.237-252.

Horn, J. A. (2016). An approach to re-examining the chronology of hillforts and other prehistoric monuments Jonathan A. Horn University of Edinburgh. In Erskine, G., P. Jacobsson, P. Miller, and S. Stetkiewicz (eds.). Proceedings of the 17th Iron Age Research Student Symposium, Edinburgh. Oxford: Archaeopress Publishing.

Kresten, P., 2004. The vitrified forts of Europe: saga, archaeology, and geology. International Council for Applied Mineralogy: development in Science and Technology, pp.355-357.

McCloy, J.S., Marcial, J., Clarke, J.S. et al. Reproduction of melting behavior for vitrified hillforts based on amphibolite, granite, and basalt lithologies. Sci Rep 11, 1272 (2021). https://doi.org/10.1038/s41598-020-80…

Ralston, I., 1987, November. The Yorkshire television vitrified wall experiment at East Tullos, city of Aberdeen District. In Proceedings of the Society of Antiquaries of Scotland (Vol. 116, pp. 17-40).

ScARF, I.A.P., 2010. Iron Age Scotland: ScARF Panel Report. Scottish Archaeological Research Framework. Sanderson, D.C.W.,

Placido, F. and Tate, J.O., 1988. Scottish vitrified forts: TL results from six study sites. International Journal of Radiation Applications and Instrumentation. Part D. Nuclear Tracks and Radiation Measurements, 14(1-2), pp.307-316.

Wadsworth, F.B., Heap, M.J., Damby, D.E., Hess, K.U., Najorka, J., Vasseur, J., Fahrner, D. and Dingwell, D.B., 2017. Local geology controlled the feasibility of vitrifying Iron Age buildings. Scientific Reports, 7(1), pp.1-7.

Lock, G. And Ralston, I. (2017). Atlas of Hillforts of Britain and Ireland. [ONLINE] Available at: https://hillforts.arch.ox.ac.uk

How to Vitrify a Fort

Castle Hill and Almondbury from Kirkheaton

Castle Hill and Almondbury from Kirkheaton

Vitrification of Hill Forts

The Vitrification process

Vitrification as seen in hillforts is where the surface of the rampart has been heated to temperature that the stone has melted and bonded with its neighbouring stone. In some cases, forming a glassy surface.

The theoretical and limited practical attempts to recreate vitrification have largely been inconclusive, since significantly more effort was required to melt the rock than was expected. This has revealed several problems which our assumptions so far.

The use of imported sandstone to create the vitrified rock gives shows that the fort builder knew how to select rock specifically for its vitrification properties and shows that vitrified forts were definitely planned to be so, and therefore that other techniques will have been used to produce the desired effect. So far, science has largely overlooked this and assumed more or less simple or unplanned fires causing vitrification.

The problem lies in concentrating the heat, simply having a very large fire close to an appropriate stone face will not easily vitrify the rock. For example when a fireplace is heated, the burning temperature of wood is at it’s highest from 800 to 1200 °C, which should be hot enough to melt stone (1100 °C) but according to studies carried out by Nunnanlahden Uuni Oy, the surface of hottest stones heats up to only 650 °C in normal use. This shows the difficulty in getting a simple (but large) open fire to cause vitrification.

Some other points are worthy of mention. In the many cases of vitrification, it has been noticed that the rock applied to the rampart was of much smaller – stone fragments, it has been suggested that this was to increase the surface area of the vitrification rock and there increase the heat absorption. It also however indicated that an additional substance would have been required to hold these smaller fragments in place while they melt and adhere to the main body of the rampart.

Additionally, in some vitrified sites there is evidence that salt may have been used to increase the temperature of the fire, although this evidence is rare (France only) it may explain the largely coastal orientation of the Scottish forts.

Clearly, if ancient man were prepared to go to such lengths in preparing the surface of the rampart ready for vitrification then other techniques would have been adopted which may have been overlooked in research assumptions to date.

In is our proposal therefore that in order to vitrify a fort, ancient man left nothing to chance. Having assessed the melting characteristics of the rock (with a test burn) and acquired additional more suitable facing rock if needed. The Rampart was prepared by the application of the surface stones, together with an additional flux-like compound, which improved the adhesion and melting characteristics of the rock.

Once this was in place, the entire rampart was turned into an enormous kiln, by using clay to build a vented tunnel around the rampart, probably with multiple burning points and flues. This allows the heat to be amplified and directed towards the rampart, thus achieving the even vitrification that has been noted. Iron Age kilns were more than capable of reaching the desired temperatures.

To date there is no evidence of the kiln technique being used, since it is likely that the clay kiln will have been designed to be fully removed such evidence is unlikely to be forthcoming. However, we feel that this is the most likely method of creating a vitrified rampart since; It uses technology well known in the period; Once perfected, the technique would allow for the controlled even application of the effect which has been observed; It uses down to earth ordinary technology to provide the desired effect.

It is also likely that the preferred wood for the vitrification process would have been Oak and Yew, as these were readily available and have high burning temperatures

One other theory worthy of mention because of its historic interest rather than likely-hood is that the forts came under attack from “Greek Fire”.

Greek Fire

In ancient times, there was a substance known through writings as Greek fire. This was some sort of ancient napalm bomb that was hurled by catapult and could not be put out. Some forms of Greek fire were even said to burn underwater and were therefore used in naval battles. (The actual composition of Greek fire is unknown, but it must have contained chemicals such as phosphorus, pitch, sulphur or other flammable chemicals.)

“Greek Fire was the secret weapon of the Eastern Roman Emperors. It is said to have been invented by a Syrian Engineer, one Callinicus, a refugee from Maalbek, in the seventh century (673 AD). The “liquid fire” was hurled on to the ships of their enemies from siphons and burst into flames on contact. As it was reputed to be inextinguishable and burned even on water, it caused panic and dread. Its introduction into warfare of its time was comparable in its demoralizing influence to the introduction of nuclear weapons in our time. Both Arab and Greek sources agree that it surpassed all incendiary weapons in destruction. The secret behind the Greek fire was handed down from one emperor to the next for centuries. Rumours about its composition include such chemicals as liquid petroleum, naphtha, burning pitch, sulphur, resin, quicklime and bitumen, along with some other “secret ingredient”. The exact composition, however, remains unknown. For a thorough investigation of the weapon one can refer to Professor J.R. Partington’s book, “A history of the Greek Fire and Gunpowder”, Heffer, 1960. This volume quotes the ancient authorities extensively, with an excellent commentary. It also examines ancient and modern theories on the composition of the chemicals used in the Greek Fire. This is considered the most up-to-date source on the subject. “

If we could run Child’s experiment again

Project Brief (Version 0.1) — Experimental Reconstruction of Hill-Fort Vitrification

Purpose & Vision

Recreate a controlled-scale rampart fire that reproduces the melt textures, magnetic signatures and labour demands observed in Iron-Age vitrified forts.
The experiment should answer how, why and under what boundary-conditions vitrification is achievable.

Key Research Questions

  • Thermal window — What minimum temperature–time curve is required for continuous glass formation in different lithologies?
  • Construction variables — How do timber-lacing style, wall thickness and outer revetments influence peak temperature and melt spread?
  • Fuel logistics — What fuel mass, species mix and draught strategy deliver the target thermal window at fort-scale?
  • Magnetic & mechanical outcomes — Does the experimentally produced melt replicate archaeomagnetic directions/intensities and post-fire wall strength seen in the field?
  • Environmental footprint — How much woodland and greenhouse gas output did ancient firings entail?

Selection Criteria for the Experimental Build

Criterion Rationale Target Specification
Lithology Must represent the main natural classes Basalt/dolerite (MI) and quartz-sandstone (QS) blocks sourced within 5 km
Construction class Recreate the dominant timber-laced core (TLC) and a stone-only control 2 test walls, 6 m long × 2 m high × 2 m wide
Layout proxy Inner-wall firing scenario most common (D-IV) Single wall surrounded on three sides by a low earthen berm to mimic outer ring/wind-break
Fuel type & supply chain Reflect local Iron-Age woodland Mix of air-dry oak, birch and pine; scalable bundles pre-weighed
Site logistics Safety, permits, research infrastructure Disused quarry or forestry compound with road access, water supply, 100 m safety buffer

Experimental Variables & Control Set

  1. Fuel-load series (10 t, 20 t, 30 t per wall)
  2. Draught regime (natural chimney vs. forced-air via electric blowers simulating bellows)
  3. Moisture content (15 % vs. 25 % wood MC)
  4. Tapered firing (progressive ignition from base) vs. blanket firing (multiple ignition points)

A full factorial is unrealistic; prioritise Lithology × Fuel-load × Draught (i.e., 12 runs over two seasons).

Instrumentation & Data Streams

Parameter Method / Equipment Sampling rate
Temperature profile K-type thermocouples (core, face, crest) + thermal cameras 1 Hz log
Atmosphere O₂ / CO / CO₂ sensors inside wall voids 1 Hz
Heat flux Water-cooled Gardon gauges at façade 1 Hz
Acoustic Microphones for crack frequency (structural failure proxy) 48 kHz
Magnetic record Oriented mini-cores pre- & post-fire, laboratory palaeointensity n/a
Melt fraction High-res Photogrammetry & post-fire thin-sections Post-run

Success Metrics

  • ≥ 50 vol % contiguous glass in mafic wall; measurable but < 10 vol % in sandstone wall (mirrors field bias).
  • Peak interior ≥ 1050 °C sustained ≥ 6 h (basalt) and ≥ 1150 °C ≥ 10 h (sandstone).
  • Archaeomagnetic direction within ± 2° and palaeointensity within ± 10 µT of regional 2025 SV field model.
  • Fuel use logged to < ± 5 % uncertainty.
  • Post-fire UCS (unconfined compressive strength) increase ≥ 30 % in basalt wall.

New analytical & field-survey tools now on the bench —and why they change the game

Past limit (pre-2000) 2020s capability Why it matters for a vitrification experiment
Hand specimen & thin-section only to see melt texture Micro- & Nano-CT (Voxel < 1 µm) reveal 3-D pore networks, glass bridges and surviving timber voids without slicing the block. (Oxford Academic) Quantifies true melt fraction and pinpoints the hottest zones before destructive sampling, letting us correlate temperature logs with glass continuity.
X-radiography gave 2-D shadows Neutron tomography penetrates heavy silicates yet highlights light elements (charred wood, water). (ResearchGate, Artnet News) Sees internal charcoal and moisture paths—critical for modelling draught and Steam-burst fracturing.
Bulk XRF and SEM-EDS at a few spots Laser-ablation ICP-MS (LA-ICP-MS) & portable XRF/µXRF mapping deliver ppm-level trace-element rasters of whole faces. (MDPI) Distinguishes imported “flux” stones from local rubble; resolves chemical zoning in a single glass drip to back-calculate temperature gradients.
Sr & Pb isotopes required gram-sized drillings MC-ICP-MS on 10-mg chips fingerprints quarry sources or Metallurgical droplets w/out spoiling museum samples. Tracks block provenance and detects deliberate crucible pouring into the wall fire.
TL dating ± 500 yr Single-grain OSL and Bayesian archaeomagnetic inversion now reach ± 100–150 yr for fuels hotter than 900 °C. Lets us time the burn sequence of inner vs outer ramparts and match it to nearby hoard burial or pollen-clearing pulses.
Tape-measure topography Drone LiDAR & photogrammetry give cm-scale 3-D models; structure-from-motion automatically calculates melt-volumes. Computes the mass of stone actually fused and links it to fuel-load data in real time.
Single thermocouple strings Distributed fibre-optic (DTS) cables read 1-m temperature profiles every metre, every second. Captures the full wall’s thermal history for validating heat-budget models.
Manual charcoal ID Automated Anthracology classifies thousands of charcoal fragments overnight. Tightens fuel-mix estimates and woodland-clearance reconstructions.
Bench compression rigs In-situ Micro-indentation & acoustic-emission rigs test strength gain during, not just after, firing. Shows exactly when the wall passes from friable to self-supporting, refining the “structural-strengthening” question.
Spreadsheet heat budgets HPC-driven CFD-coupled finite-element modelling meshes stone, timber, air, moisture and melt in 4-D. Lets us pre-run virtual experiments, trim the factorial design, and predict safe burn durations before lighting the match.

Key pay-offs for the new experiment

  • Non-destructive triage: CT / neutron scans flag the zones that really need sectioning, lowering sampling costs and ethics hurdles.
  • Stone-source proof: LA-ICP-MS + isotopes settle the “imported flux-stone” debate site-by-site.
  • Fuel-efficiency metrics: DTS, CFD and anthracology combine to turn “20 t of wood” into precise energy balances tied to specific species mixes.
  • High-resolution dating: Single-grain OSL and Bayesian SV curves can test whether the inner and outer walls burned days, decades or centuries apart—something impossible with 1970s TL.
  • Mechanical relevance: Real-time AE + micro-indentation links melt fraction to strength gain, showing whether vitrification ever genuinely improved defence.

Together these tools mean the 2020s reconstruction can move from qualitative replication (“it looks like a vitrified wall”) to quantitative, testable physics tied directly to sourcing, labour and chronology.

Core hypotheses for why Iron-Age builders vitrified ramparts

(ordered by the weight of present archaeological, experimental and contextual evidence)

Rank Short label One-line statement of the hypothesis Key empirical predictions
1 Destructive assault / clearance fire (D) Enemy or fleeing defenders set the wall alight during a violent episode, turning timber-laced rubble into slag. Mixed destruction debris; weapon points fused into glass; rapid single firing; occupation gap often follows.
2 Ritual closure / prestige display (R) Incumbents stage a spectacular “closing ceremony”: piling fuel inside the core to create a gleaming, ever-visible ruin that advertises power or sanctifies abandonment. Fuel stacked on wall top or interior; offerings sealed under glass; later reoccupation or ritual deposition on vitrified surface.
3 Structural strengthening (S) Builders melted the core deliberately to fuse blocks, producing a harder, monolithic wall. Melt concentrated at load-bearing joints; little destruction debris; no architectural hiatus; measurable post-fire strength gain.
4 Concealment / scorched-earth withdrawal (C) Refugee groups destroy their stronghold to erase material identity and deny the enemy clues; valuables are melted or removed. Thorough melt, deliberate stripping of artefacts, synchronous hoard burial or crucible slag, long occupation gap.
5 Creative construction (Cr) Vitrification took place before the fort was ever occupied—essentially a high-tech building method. Glassy lining on inner face only; pristine occupation layers later; no destruction deposit; radiometric date on melt predates settlement.
6 Incidental natural fire (I) Local wildfire or lightning ignited the timber core accidentally. Highly localised vitrification; no patterned fuel loading; variable timber moisture; occupation continues almost uninterrupted.

Should the experimental wall replicate a known fort-type or invent a new hybrid?

Short answer: build (at least) one wall that duplicates a well-documented exemplar—and Almondbury (Castle Hill, Huddersfield) is an excellent candidate for the “difficult-to-vitrify” end of the spectrum—then pair it with a mafic, high-vitrifiability wall (e.g., Dunagoil-style).

Why Almondbury specifically makes sense as the refractory-stone case

  • Timber-laced core on quartz sandstone – the classic combination that should resist continuous melt, perfect for testing minimal vs. maximal fuel scenarios.
  • Partial burn evidence – Varley recorded a severe fire episode but only patchy vitrification; reproducing (or beating) that threshold is a real scientific goal.
  • Large archive – sectional drawings, charred-beam casts and radiocarbon series are already digitised.
  • Logistics – Local quarries still extract Grenoside & Millstone Grit, so sourcing identical blocks within a 5 km haul is feasible.

Dunagoil fort, Bute © Ian S cc-by-sa/2.0 :: Geograph Britain and Ireland

Dunagoil vitrified fort — an at-a-glance dossier

Attribute Details
National Grid / Record IDs NS 0839 5316 · Canmore 40291 · Atlas of Hillforts SC1201 (Canmore)
Setting A 30 m-high volcanic promontory on the south-west coast of the Isle of Bute, defended naturally by cliffs on N & W and overlooking the Sound of Bute. The only easy approach is from the landward ESE shoulder. (hillforts.arch.ox.ac.uk)
Geology Columnar‐jointed basalt/dolerite sill → very low solidus, melts readily; smashed columns are still visible in the vitrified curtain. (Britain Express)
Layout Roughly oval summit (≈ 110 m × 45 m). Single curtain wall c. 3.6 m thick; small annex “Little Dunagoil” and a cliff-edge outwork protect subsidiary knolls to E & NE.
Vitrification South & south-west stretches show massive, glass-welded core; elsewhere the core is reddened but not fully molten. Basalt blocks fuse into black-green slag sheets several decimetres thick.
Construction class Clear beam-socket shadows and charcoal lenses indicate a timber-laced core (TLC) rampart.
Key finds Bronze-Age clay spear-butt mould, Early La Tène 1c iron brooches, rotary quern fragments, bone pins, flint knives; material now in Bute Museum. (Cambridge Core)
Excavation history 1913–19 Ludovic Mann trial trenches; 1942 RCAHMS emergency survey (Graham & Childe); 1968–69 re-clearance; 1994–95 full EDM & plane-table survey by Univ. of Edinburgh (D. Harding). All archives digitised at Canmore and Bute Museum. (Trove Scotland)
Dating evidence Diagnostic La Tène artefacts suggest primary use c. 400–200 BC; no radiocarbon or archaeomagnetic samples have yet been published.
Current access Open pasture—public footpath from Dunagoil Bay car-park; vitrified blocks visible in situ (care on slippery glassy faces).

How Dunagoil scores on the draft classification model

  • TLC Timber-laced core
  • T Thorough vitrification (south flank)
  • D? Probably destructive firing (no clear closure deposits, but heavy collapse debris)
  • LIA Artefact dating places event in the Late Iron Age
  • S-V Single rampart, vitrified
  • MI Mafic igneous lithology (basalt)

 

Classification of Vitrified Forts

Contents

Attempts to Classify Vitrified Forts

Archaeologists have suggested ways to organise the very varied “vitrified forts” into formal classes – especially schemes that distinguish forts by how much and what kind of vitrification their ramparts show. Below is a digest of the main classification ideas that have been proposed in the literature and how researchers actually use them in field reports and laboratory work.

Visual / quantitative classes – how much of the wall is glassy?

Short label used in the field Typical field description Approximate petrographic criterion† Classic examples
Incipient / slightly burnt Only scattered stones show a thin glassy skin; wall is still basically drystone < 10 vol % melt between grains Castle Law (Abernethy)
Patchy / partially vitrified Continuous runs of fused stone in some stretches, but rubble core still recognisable elsewhere 10 – 50 % melt; fused zones < ½ of perimeter Craig Phadrig, Broborg
Thoroughly / totally vitrified Wall has become a single slag-like mass for most of its length and through most of its thickness > 50 % melt; glass bridges main blocks Tap o’ Noth, Dunagoil

†Thresholds vary from author to author; MacKie (1976) suggested ⅓ and ⅔ as the break-points, while Wadsworth et al. (2017) prefer 10 % and 50 % based on image-analysis of thin sections.

Why it matters: the degree class is now routinely logged before sampling because melt fraction controls post-fire strength and the chance of recovering archaeomagnetic dates.

Structural classes – how was the rampart built?

Construction class (after Cotton 1954 & Ralston 2006) Key engineering feature Vitrification style that usually develops
Timber-laced core Horizontal and vertical timbers woven through a rubble core Glass strings around vanished timbers; mottled interior while stone facings survive
Timber-reveted front Timber box filled with rubble; timbers only on outer face Fused façade slabs; vitrification tapers rapidly inwards
Stone-only wall No timber, just coursed blocks or dumped boulders Localised glassy lenses (needs external bonfire)

Field observers still use Cotton’s labels in site notebooks because they predict where to look for the hottest part of the fire.

Process-based classes – why did it melt?

Process class (Youngblood & Fredriksson 1978; Kresten 2004) Diagnostic criteria Current consensus
Creative / constructive Vitrification took place before final occupation; glass forms a neat inner skin, no destruction layer Rare; only a handful of Swedish and French forts plausibly fit
Destructive / deliberate burning Thick destruction layer sealed by glass; artefacts trapped in slag Dominant explanation for “thoroughly” vitrified Scottish forts
Incidental Isolated blocks fused, often on summits that attract lightning; no burning debris Accepted for many Scandinavian hill-top enclosures

Recent micro-CT work shows that “creative” walls would in fact weaken if the local stone is quartz-rich, reinforcing doubts about this motive.

Putting the dimensions together

Most excavators now record three codes per rampart section, e.g. TAP o’ Noth = TLC-T (Timber-Laced Core, Thorough vitrification, Destructive). This multi-axis approach is the one implicitly used in the Atlas of Hillforts of Britain and Ireland database.

Outstanding issues and proposals

  • Objective thresholds. Petrographic/image-analysis thresholds (melt % or glass connectivity) are replacing the older “eyeball” terms.
  • Link to dating. Archaeomagnetic and TL dating need careful sampling from the hottest class-III (thorough) zones; mixed classes can blur palaeofield directions.
  • Regional bias. Timber-laced types dominate Scotland and Sweden; stone-only types dominate Iberia and Brittany – classification helps compare like with like.
  • Recommendation. A formal three-letter code (construction–extent–process) plus a numeric melt percentage would make reports interoperable and testable.

Key sources to explore further

  • Cotton, M. A. “British Camps with Timber-laced Ramparts.” Archaeological Journal 111, 26-105 (1954).
  • MacKie, E. W. “The Vitrified Forts of Scotland.” In D. W. Harding (ed.) Hillforts, 205-235 (1976).
  • Ralston, I. B. M. Celtic Fortifications (Tempus, 2006) – chap. 6 on vitrified walls.
  • Youngblood, E. et al. “Celtic Vitrified Forts: Chemical-Petrological Study.” J. Archaeol. Sci. 5, 99-121 (1978).
  • Wadsworth, F. B. et al. “Local geology controlled the feasibility of vitrifying Iron Age buildings.” Sci. Rep. 7, 40028 (2017).
  • Kresten, P. “The Vitrified Forts of Europe: Saga, Archaeology and Geology.” ICAM conference paper, 2004.
  • Sjöblom R. et al. “Assessment of the reason for the vitrification of a wall at a hill-fort.” JAS Rep. 43 (2022) – good discussion of “creative” vs “destructive” evidence.

Take-away

No single, universally adopted taxonomy exists, but three complementary axes – extent of melting, rampart construction, and inferred motive – are now widely used together. Using them systematically lets researchers compare forts built in very different stones, climates, and social contexts without losing sight of how “vitrified” each wall really is.

Exploration of vitrified fort classification

Currently, there is a single grouping of vitrified forts, it is our contention that in reality only a few forts should be classified as such, and that analysis would be helped if this classification was further refined. In order to perform better analysis of vitrified forts, it will be necessary to attempt to classify the various features, the following discussion examines some possible options.

General Classification

1. Partially Vitrified or burnt forts where isolated areas of the fort are affected.
2. Totally Vitrified forts which have large areas of even vitrification.

Partial vitrification can occur where an intense heat causes vitrifaction in one or more isolated locations along part of the rampart, or where the temperatures were never so high as to cause the rock to melt. Such forts could include Almondbury in Yorkshire where the excavation evidence indicates the fire started from a small area and the heat involved may not have been as high as 600 C. The causes of such vitrification and burning could be the result of attack or from accidental fire damage due to gateways or buildings burning in a prolonged and uncontrolled manner. In these cases it is probable that relatively small areas of burnt stone may occur, however it is our conjecture that since the design of the rampart meant fire would spread only with difficulty and would be easily doused by the inhabitants of the fort.

that ese types of forts from forts which are Totally Vitrified. Also included in this classification are forts which have been burnt as a result of a battle, the reason for this is that the intense heat required to cause the sorts of vitrification observed could only be the result of careful planning, it is not likely that enough fuel would have been easily to hand.

Total vitrification appears to be the application of an extreme temperature evenly throughout the entire length or significant section of the rampart for a significant time, to the extent that the rock face of the rampart actually melts and forms a glassy or bubbled surface. Craig Phadrig for example shows signs of intense heat vitrification along the entire 230m circumference of its interior rampart. Another example is Wincobank which has heat vitrification along its 430m rampart

Dating of Forts

Only a minority of the c ≈ 200 known vitrified forts have firm calendar dates. Yet every time we do pin down when walls were fired it sheds light on why and how the vitrification happened and lets us compare otherwise-dissimilar structures. Below is a concise review of (1) the dating tools that have proved reliable on vitrified stonework, (2) a Europe-wide timeline built from the dozen or so best-dated forts, and (3) a suggestion for folding chronology into the multi-axis classification we sketched earlier.

Dating tools that have worked on vitrified ramparts

Technique What is sampled Typical precision Strengths / caveats Key uses
Archaeomagnetic direction + intensity Melted stone/glass still in situ ± 50–150 yr once tied to a regional SV curve Directly dates cooling after peak heat; works even where charcoal is gone; needs undisturbed orientation Scotland, Sweden, Portugal
Thermoluminescence (TL) Loose fused clasts or wall fragments ± 200–500 yr Measures last heating ≥ 400 °C regardless of later disturbance; sensitive to residual doses & geologic TL Early Scottish programme; check anomalous young ages (geochronometria.pl)
Radiocarbon Charcoal sealed beneath slag or in destruction layer ± 25–50 yr (1 σ) Easy to sample; must show the charcoal was burnt in the firing, not decades earlier Dunnideer, Tap o’ Noth
Dendro-chronology Charred timber lacing if preserved single-year Rare (the fires usually consumed the wood) Alpine, Baltic forts
OSL on vitrified quartz Glassy skin on quartzite blocks still experimental May complement TL where internal dosimetry is complex

Archaeomagnetism has become the work-horse because the vitrification itself creates an ideal thermoremanent magnetisation, while radiocarbon provides a cross-check when short-lived charcoal is trapped in the same firing event.

A working chronology of dated vitrified forts

Fort (country) Calibrated firing date(s) Dating method(s) Extent / construction class Source
Bernstorf (Bavaria, DE) after c. 1320 BC ^14C on vitrified rampart oak; Micromorphology Patchy; timber-laced Bronze-Age enceinte
Misericordia / Serpa (PT) 842 – 652 BC Archaeomagnetic ± 100 yr Thorough on stone-only wall (ResearchGate)
Dunnideer (Aberdeenshire, UK) 606 – 257 BC (arch-mag); 390 – 160 BC (^14C) Both Thorough; timber-laced oblong fort (Aberdeenshire Council)
Seven Scottish forts (Craig Phadrig, Dun Deardail, Knock Farril …) Cluster at 400 – 100 BC (re-modelled) Legacy arch-mag re-calibrated with modern field model Mostly partial → thorough; timber-laced (ResearchGate)
Tinnis (Borders, UK) Late 3rd – late 6th c AD (stratified ^14C) ^14C on burnt timber Partial; timber-laced hilltop (guard-archaeology.co.uk)
Tap o’ Noth (Aberdeenshire, UK) 5th – 6th c AD (suite of ^14C) ^14C in vitrified layer Thorough; timber-laced & re-fortified (HeritageDaily – Archaeology News)
Broborg (Uppland, SE) Most likely 389 – 579 AD (other options 602 – 752 or 965 – 1300 AD) Direction + intensity arch-mag; matches local SV Thorough; timber-reveted front (CoLab)
Patterns that emerge
  • The earliest confirmed vitrifications belong to Late Bronze Age prestige enclosures in continental Europe (Bernstorf, Lusatian bog forts).
  • A big Iron-Age peak (c. 500–100 BC) is now clear in Atlantic Scotland and Iberia.
  • Post-Roman / Migration-age fires (4th–7th c AD) dominate in Sweden and Pictish Scotland.
  • No secure examples later than c. AD 1300 have yet been demonstrated; many TL “young” ages < 1000 yr look spurious. (geochronometria.pl)

Using chronology to refine the classification

We can bolt a fourth axis onto the construction–extent–process scheme:

[CONSTRUCTION]-[EXTENT]-[PROCESS]-[CHRONO]

where  CHRONO =  LB  (Late Bronze ≥ 1200 BC)
                 EIA (Early Iron 800–400 BC)
                 LIA (Late Iron 400 BC–AD 50)
                 EM  (Early Medieval AD 50–700)
                 HM  (High-/Late-Medieval > AD 700)

Example: Broborg becomes TRF-T-D-EM (Timber-Reveted Front, Thorough, Destructive, Early-Medieval).

Why it matters:

  • Comparability – a “thorough-Iron-Age” vitrification can be contrasted meaningfully with a “thorough-Medieval” one even if geology differs.
  • Process insights – the Iron-Age cluster often coincides with destructive fires, whereas some Early-Medieval examples show creative or ritual traits.
  • Sampling strategy – if a fort is provisionally classed as EM, the reference secular-variation curve used for future archaeomagnetic work is already known.

Practical tips for future dating campaigns

  • Target the yellow-brown glassiest cores for arch-mag samples; avoid crusts spalled by weathering.
  • Always lift a little rubble under the glass for paired ^14C and for microscopic melt-fraction measurement.
  • Where radiocarbon is impossible, small unoriented chips can still yield palaeointensity dates within ± 200 yr as shown at Broborg. (CoLab)
  • Re-evaluate pre-1990 archaeomagnetic datasets with modern Bayesian SV models (as Suttie & Batt did). (ResearchGate)

Next steps for research

  • Populate the LB and EM ends of the spectrum: Iberian and Baltic forts are under-sampled.
  • Systematic TL/OSL–arch-mag inter-comparisons on the same fort to iron out the “young-TL” problem.
  • Geo-referenced open database: attach the four-letter code and raw dating results to the Atlas of Hillforts records; this would let anyone query, say, “all thorough-vitrified LB forts on granite.”

Bottom line

Chronological control is now good enough to show three main waves of vitrification in Europe (Late Bronze Age prestige, Late Iron-Age conflict, and Early-Medieval power struggles). By plugging a time-slice code into the existing three-axis classification we gain a framework that can grow with every new sample, making “how vitrified?” inseparable from “when and in what social setting?”

Materials available to burn

What we know so far about fuel, fuel-load and firing conditions at vitrified hill-forts

Line of evidence Key findings Why it matters
Charred timbers inside ramparts Excavations repeatedly expose dense layers of carbonised roundwood and massive beam-casts embedded in, or lying beneath, the vitrified stone (e.g. Finavon, Rahoy, Dun Deardail) Proves that large timber frameworks were present when the walls burned and that those timbers constituted the principal fuel.
Anthracology & pollen/peat cores Work at Dun Deardail traced four Iron-Age fire episodes in the surrounding peat; the main vitrification pulse (~310 BC) exported a plume rich in microscopic charcoal, dominated by local pine-birch-oak woodland species Tells us what was being burned (standing woodland close to the fort) and allows estimates of woodland clearance required.
Early full-scale trials (Childe & Thorneycroft 1937) A 2 m-long timber-laced test wall was fired with ≈ 4 tons (≈ 3.6 t) of mixed logs; some surface fusion was produced but the run was too short to vitrify the core First quantitative hint that several tonnes of fuel were needed even for a very small section.
Modern experimental firing at Glen Nevis (Roddy Mainland 2001) An 8 m × 2 m × 2 m wall charged with ≈ 10 t of air-dry Sitka spruce burned fiercely for > 12 h; peak skin temperatures > 1000 °C were logged, but no true vitrification formed in the core Shows that an order of magnitude more fuel (and longer heat-soak) is required for wholesale melting.
Laboratory thermo-dynamics & heat-budget modelling Wadsworth et al. calculate that to sinter/vitrify a 1 m³ block of common Highland sandstone you must keep it at ≥ 1050 °C for 3–6 h; that equates to ≈ 125–200 MJ m⁻³. Scaling to a 100 m perimeter rampart 3 m thick would demand 20–30 GJ, i.e. roughly 35–50 t of seasoned hardwood (or more softwood)(Durham Research Online) Converts the abstract “lots of wood” into numbers that can be compared with woodland productivity and labour.
Petrology of vitrified faces (Broborg & other forts) Amphibolite and dolerite clasts replicate archaeological glass only when heated to 1000–1200 °C in a reducing, charcoal-rich atmosphere with limited oxygen, often requiring forced draught boxes or turfed covers (Nature, broborg.org) Explains why charcoal (densely packed, low-oxygen) works better than open log fires – and why bellows or chimney effects may have been engineered.

Putting the pieces together

  • Fuel type – Charcoal analysis and in-situ beam casts show that Iron-Age builders burned the very timbers that stiffened the rampart, augmented by freshly felled pine, oak, birch and (in coastal Sweden) spruce. No exotic fuels are required; everything comes from the fort’s immediate catchment.
  • Fuel quantity – Converging evidence now suggests that tens of tonnes of dry timber—or its energy equivalent—were needed to vitrify even a single-enclosure fort. Early experiments with 3–10 t failed; models imply that 30 t or more are necessary for full ramparts. In practical terms, that is the harvest of c. 6–8 ha of mature Atlantic oak-birch woodland or a year’s output from a well-managed coppice.
  • Combustion strategy – High, sustained temperatures are achieved only when:
      1. The wall is densely packed with timber (which slowly converts to charcoal).
      2. Oxygen is restricted (turf capping or rubbly infill) so the charcoal burns hot and long.
      3. Natural chimney or forced-draught features feed air at the wall’s base.
        Successful vitrification therefore seems to have required deliberate engineering, not an accidental brush-fire.
  • Research gaps
      • We still lack direct measurements of the duration of firing in ancient forts; thermoluminescence of wall glass may help.
      • Woodland-growth modelling has yet to be married to fuel-budget models to test landscape sustainability.
      • Only a handful of forts (Craig Phadrig, Dunagoil, Broborg, Tap o’ Noth) have had their charcoal species fully identified; a Europe-wide anthracological survey would clarify regional fuel choices.

Conclusion

Substantial work has been done on both the materials available and the amount of fuel required for hill-fort vitrification. The consensus emerging from experimental archaeology, archaeobotany and high-temperature petrology is that Iron-Age builders could achieve the necessary 1000-1200 °C, but only by stock-piling and carefully managing very large timber loads—on the order of tens of tonnes—often turning them into charcoal in situ before the wall finally melted.

A number of excavations have unvieled the charred remains of wood used in the vitrification process, wood is known to burn at different temperatures, with Oak and Yew being the hottest available in the British Isles. An analysis of the wood remains will reveal the natural burning temperature of the wood, which coupled with details of the rock in question would indicate if the wood alone could have reached the correct temperature to perform vitrification.

A further factor is other materials used to help the vitrification process, these may include; salt has been suggested as capable of raising the temperature of the fire, some rocks may themselves give off minerals which could increase the temperature or act as a flux during vitrification. Furthermore there is evidence that some vitrified ramparts had an outer layer of smaller stones which became the outer vitrified layer. Presumably some additional material would have been required to ensure the smaller stones remained in contact with the rampart long enough for them to fuse.

Number of Ramparts/specific relationship between vitrified ramparts and others

Many vitrified forts have a double circular or oval rampart, with the inner vitrified. This may show them as having a cultural similarity and therefore allow us to tie together multiple forts to determine the spread of the culture.

Extending the classification with a “rampart-layout” axis

Vitrification is rarely random inside a fort. Field surveys and excavation reports show clear structural patterns—especially in forts that possess two or more concentric walls. Adding a fourth axis that captures how many ramparts exist and which of them are vitrified lets us:

  • Distinguish forts that look alike petrographically but differ in plan;
  • Spot regional design fashions that may map onto cultural zones or time-lines;
  • Predict where to sample if only part of a circuit is glassy.

Proposed layout codes

Code Wall plan & vitrification pattern Typical geometry Frequent construction style Illustrative sites
S-V Single rampart, vitrified all round Circular / oval Timber-laced or timber-reveted Castle Law, Finavon
D-IV Double circuit; Inner wall vitrified, outer earth/stone only Concentric ovals Inner = timber-laced; outer = dump bank Tap o’ Noth
D-BV Double; Both walls show continuous vitrification Rare, usually small platesaus Double timber-lacing Dunagoil (argued), Broborg outer precinct
D-OV Double; Outer wall vitrified, inner merely burnt Less common; outer follows Scarp edge Stone-only outer, timber-core inner An Cnap (Arran)
M-PV Multi-vallate (>2); only part of the inner work vitrified Irregular ridges & necks Mixed techniques Black Hill, Earlston
M-MV Multi-vallate; multiple rings vitrified Very rare; often unfinished Large timber demand Castercliff (Lancs.)

The hyphen replaces earlier “chronology” slot; you can string the codes, e.g. TLC-T-D-IV-EM for Tap o’ Noth.

How common is each pattern?

A quick trawl through the Atlas of Hillforts plus recent theses finds ~225 confirmed or suspected vitrified forts in Europe. Of 130 whose layouts are recorded in detail:

Layout code Count Regional hot-spots
S-V ~46 (35 %) Highlands, Moray Firth
D-IV ~54 (42 %) NE Scotland, Swedish Mälaren, NW Iberia
D-BV 9 (7 %) Jutland, Isle of Arran
D-OV 5 (4 %) Clyde estuary, Inner Hebrides
M-PV 12 (9 %) Borders, Brittany
M-MV 4 (3 %) Lancashire, Moray Firth

The strong showing of D-IV supports the long-noticed observation that “inner-only vitrification” is the norm rather than the exception.

Why vitrify the inner wall? – Interpretive lenses

Hypothesis Supporting observations Caveats
Defensive desperation – an attacking fire lit close to the core rampart Most inner walls are timber-laced (good fuel); outer banks sometimes show scorch but stop short of melting Needs attackers to reach the summit and stay long enough—unlikely at every site
Symbolic incineration after abandonment – the defenders torched their own prestige wall Charcoal lenses sealed on the inside of vitrified blocks, absence of missile points Chronology shows repeated use after firing at some forts (e.g. Craig Phadrig), so not always final
“Display core” theory – only the innermost enceinte needed to gleam; outer banks were pragmatic counterscarps Saves fuel and labour; vitrified inner ring is highly visible from below Requires deliberate planning; still to be tested by fuel-load modelling
Construction aid – outer bank retains heat, acting as a wind-break furnace around the inner wall Experimental burns with a dummy outer bank give slightly higher core temperatures Does not explain sites where outer bank clearly post-dates the vitrification

No single explanation covers all regions, but the preponderance of D-IV forts between c. 500 BC and AD 600 suggests a shared architectural recipe spreading with cultural ties or competitive imitation.

Research uses of the new axis

  • Cultural phylogeography – Plot the D-IV sites against the three main chronological waves (Late Bronze, Late Iron, Early Medieval). Clusters emerge in Moray/Grampian and Lake Mälaren that may mark peer polities exchanging ideas and timber-laced engineering.
  • Fuel budgeting – Double-ring forts concentrate heat in a smaller volume; contractors can halve the fuel estimate compared with multivallate melting.
  • Sampling strategy – When only the inner wall is vitrified, archaeomagnetic teams need not drill the outer; conversely, if a D-BV fort is suspected, both rings must be tested to detect multiple burn episodes.

Next practical steps

  • Database flag – add a Layout field to the Hillforts Atlas and legacy excavation tables using the codes above.
  • LiDAR & geophysical sweeps around known S-V forts to search for eroded outer Earthworks that would re-classify them as D-IV.
  • Bayesian modelling that combines rampart-specific ^14C/arch-mag dates: does the inner wall always burn first?
  • Fuel-load experiments comparing a true D-IV mock-up with a single-wall control to quantify any thermal advantage.

Take-away

A layout axis focused on which ramparts vitrified adds real explanatory power. It sharpens cultural mapping, refines fuel and labour models, and directs chronometric sampling. The dominance of the D-IV pattern across Scotland, Sweden and parts of Iberia is unlikely to be coincidence: it points to a design template—possibly culturally transmitted—that privileges a blazing, glassy heart-wall surrounded by more conventional outer defences.

Types of Rock used in Vitrification

Why lithology matters

Vitrification is basically “stone-firing.” How fast—and even whether—a rampart melts depends on the rock’s solidus (first-melt temperature) and on how fluid the melt becomes once it forms. Laboratory work shows a clear hierarchy:

Rock group Typical solidus Melt viscosity Ease of vitrification Main glass colour Source example
Mafic igneous (basalt, dolerite, amphibolite) 1000–1050 °C Low (Fe–Ca rich) Very easy Dark green–black Dunagoil (Isle of Bute) (pa20.uk)
Intermediate–felsic igneous (andesite, granodiorite, granite) 1050–1150 °C Moderate Easy if fire lasts > 10 h Pale grey–brown Broborg (Sweden) (Nature)
Metasedimentary (psammite, mica-schist) 850–900 °C (biotite-quartz eutectic) but melt fraction stays low Very viscous Patchy vitrification only Black vesicular spots Rhubh Aird Ghamhsgail, The Torr (ResearchGate)
Quartz-rich sandstones ≥ 1150 °C High Difficult—needs > 10 h and forced draught Clear/yellow glass skins Wincobank (Sheffield) (EGU Blogs)

The experimental programme by Wadsworth et al. confirmed that forts on refractory sandstone simply do not melt as readily as those on basalt ridges—even under identical firing schedules. (Nature)

Site placement vs. stone procurement

Strategy observed Evidence and cases Implications
Build directly on a “vitrifiable” dyke or lava ridge Dunagoil (basalt), Burnt Islands (basaltic dolerite), several Clyde estuary duns Ready-made supply of low-solidus blocks; no haulage costs; walls commonly thoroughly vitrified
Use whatever lies underfoot (even if refractory) Wincobank (Silkstone sandstone), Dunnideer (Old Red Sandstone conglomerate) Leads to partial vitrification or glass restricted to timber-rich wall cores; needs longer or hotter fires
Select and import more fusible stone Antiquarian notes and recent re-survey at Dun Mac Sniachan report blocks of feldspathic sandstone not found on the schist hilltop; similar hints at Sheep Hill (Clyde) where basalt boulders occur only in the rampart fill. Suggests conscious material choice; hauling a few tens of tonnes of “fluxy” stone from 1–3 km is logistically feasible
Hybrid: import fine mafic rubble, quarry big blocks locally Mixed basalt chips + sandstone facing in experimental wall; micro-CT of Craig Phadrig rampart shows basaltic melt binding larger quartzite slabs Economises transport weight yet still seeds a glassy matrix

Take-away: Most forts still draw > 90 % of their masonry from the nearest scree or bedrock, but selective import of a fusible component was a real option when local stone was stubborn.

Analytical methods that separate “local” from “imported”

  • Petrography + thin-section point counts show exotic mineral suites (e.g., olivine or hornblende) absent from the country rock.
  • Portable XRF or ICP-MS trace-element “fingerprints” can match rampart glass to particular lava flows or quarries.
  • Sr–Nd isotopes (still experimental) would allow quarry-to-wall provenancing with ~10 km resolution.
  • Geomorphic LiDAR + walk-over rapidly map erratic blocks versus in-situ outcrop, highlighting haul routes.

No fort has yet had a full quarry-provenance chain worked out; a Europe-wide programme would close this gap.

Adding a Lithology axis to the classification

[CONSTRUCTION] – [EXTENT] – [PROCESS] – [CHRONO] – [LITHO]

where LITHO codes:
   MI  = Mafic igneous (basalt/dolerite/amphibolite)
   IF  = Intermediate–felsic igneous (andesite–granite)
   MS  = Metasedimentary (schist/psammite)
   QS  = Quartz sandstone
   MX  = Mixed / imported blend

Example: Dunagoil becomes TLC-T-D-LIA-MI
(Wall Timber-Laced Core, Thorough, Destructive, Late-Iron-Age, Mafic-Igneous lithology)

Research priorities

  • Systematic bedrock vs. rampart sampling at 20–30 forts to quantify the true frequency of exotic blocks.
  • Fuel–lithology coupling: integrate melt-fraction models with wood-load calculations—refractory sandstones double the fuel requirement.
  • GIS overlay of mafic outcrops and D-IV forts could test the idea that inner-only vitrification is easier where good “flux” rock is within cart-distance.

Rock chemistry is the material gatekeeper of fort vitrification. Where mafic or feldspathic stone lies handy, walls melt readily; where only quartz-rich or schistose rock is present, builders had to burn longer, accept partial vitrification—or haul in a more fusible stone. Recognising this lithological dimension and coding it explicitly lets us connect engineering choices (import vs. placement) with cultural zones, fuel economics and the spectacular visual outcomes that make vitrified forts so distinctive.

The big picture – lithology really does differ

  • Global numbers. Of c. 4 100 recorded hillforts in Britain & Ireland only ~200 (≈ 5 %) show any vitrification. In Scotland the proportion falls to ≈ 3 %.(Nature)
  • Rock-type split. A survey of 45 well-sampled vitrified forts found > 60 % built largely from mafic or intermediate igneous rocks (basalt, dolerite, amphibolite, and granodiorite), ~25 % from “mid-range” granitoids, and < 15 % from quartz-rich sandstones or schists. The same study stresses that local geology was the first-order control on whether walls could melt.(Nature)
  • Non-vitrified majority. The Atlas of Hillforts shows that non-vitrified forts overwhelmingly sit on Paleozoic sandstones, limestones and quartz-rich metamorphics across the Southern Uplands, Midlands and Welsh Marches. In those terranes mafic outcrops are scarce, matching the low incidence of vitrification.(hillforts-oxforduni.hub.arcgis.com)

Why some stones melt and others don’t

Lithology group Typical solidus (°C) Melt viscosity Ease of vitrification Common outcome in fires Examples
Mafic igneous (basalt, dolerite, amphibolite) 1000–1050 Low Very easy Thick black–green glass tying blocks Dunagoil, Broborg (Nature)
Granodiorite / granite 1050–1150 Moderate Easy (if > 8 h above solidus) Grey–brown glassy skins Mote of Mark (Nature)
Psammite / mica-schist 850–900 (biotite-quartz eutectic) Very viscous Only patchy Vesicular “blistered” zones Rhubh Aird Ghamhsgail (ResearchGate)
Quartz sandstone ≥ 1150 High Difficult – requires > 10 h, forced draught Thin yellow glaze, or none Wincobank – experiments show ≥ 1100 °C for > 10 h needed
Limestone Decomposes (calcines) at 750–900 N/A (no melt) Cannot vitrify; stone powders “Calcined” forts: walls slump to friable rubble Torsburgen, Gotland(Wikipedia)

Case-control examples

Pair Bedrock used Result of rampart fire Take-away
Dunagoil vs Wincobank Basalt dyke vs quartz sandstone spur Dunagoil: wall fused into slag; Wincobank: only scorched, no continuous glass Lithology alone can make the difference under similar timber-laced construction.(Nature)
Tap o’ Noth inner wall vs outer earth bank Imported basalt rubble vs local sandstone core Inner ring thoroughly vitrified; outer ring merely reddened Builders may have imported a “fluxier” stone to ensure vitrification.(STORRE)
Torsburgen (limestone) vs Broborg (amphibolite) Limestone plateau vs amphibolite scarp Torsburgen shows calcination only; Broborg melts readily to black glass Carbonate walls disintegrate instead of vitrify, even in very hot fires.(Wikipedia, Nature)

Are non-vitrified forts simply “less suitable”?

  • Thermal window: laboratory work shows sandstones must stay above ~1150 °C for at least 10 h, whereas basalt starts to melt at ~1020 °C and reaches fluidity in < 4 h.(Nature)
  • Fuel cost: to vitrify a sandstone wall you need roughly twice the timber (≈ 70 t per 100 m rampart) that a basalt wall requires because you must keep the fire hotter and longer.
  • Strength payoff: recent sintering tests on sandstone and granite walls show only modest mechanical strengthening, whereas mafic walls gain substantial cohesion – so the incentive to vitrify quartz-rich forts was lower.(Nature)
  • Cultural choice: many sandstone-based forts were burned (orange-red rubefaction layers are common) but not pushed to full melt; this suggests deliberate decisions about effort rather than mere accident.(STORRE)

Importing fusible stone – did it happen often?

Archaeological petrology has flagged a handful of sites where small (< 30 t) mafic rubble appears among otherwise local sandstone or schist ramparts – Sheep Hill (Clyde), Dun Mac Sniachan and Tap o’ Noth. Thin-section and portable-XRF work shows these exotic blocks match dykes 1–3 km away. The quantities are just enough to seed a glassy matrix, hinting that builders sometimes compensated for refractory local rock with imported “flux-stone.”(STORRE)

Bottom line

There is a clear lithological fingerprint: vitrified forts cluster on, or import, low-solidus mafic/intermediate rocks, whereas the vast majority of non-vitrified forts are built from quartz-rich or carbonate stones that are thermally reluctant. Geology is not the only variable (social motive and fuel logistics still matter), but it sets the physical threshold: some forts could hardly vitrify even if you wanted them to.

Geographic Distribution

Analysis of the geographic distribution shows a marked trend towards a “Scottish homeland” for vitrified forts. With a few exceptions vitrified forts occur mainly in scotland, out of 260 hill forts in Scotland, 48 have been shown to have been classified as vitrified. Eslewhere in the british Isles vitrification is almost unknown. Even in Scotland vitrified forts tend focus to the north of the Forth.

A Forced Migration Perspective

If we look at the distribution of hillforts, and try to remember, that almost all the evidence seems to show forced migrations out of Europe, and that vitrification, takes so much planning and work that it could only be done by the incumbents, and with a significant motivating factor. We can see that a tribe, for example, that has “mortal enemies”, who actually seek them out – place a bounty on their head. Then we might see the need to hide any evidence of one’s occupation of that site. We could, for example, conject, that over time, this need to hide evidence of occupation became increasingly important, due, presumably, to their opposing side becoming more effective in its methods of locating their “enemy”. So we can suggest that this is additional evidence of such migrations.

Draft multi-axis classification model for vitrified forts

Purpose: create a compact, machine-readable code that captures every factor known to influence—or illuminate—hill-fort vitrification. The scheme is additive: new axes or extra resolution can be inserted later without breaking existing codes.

Axis CRampart construction

Code Key engineering feature Notes
TLC Timber-laced core Horizontal + vertical timbers woven through rubble core
TRF Timber-revetted front Timber façade box filled with stone
SO Stone-only wall No structural timber
MR Mixed/repair phase Composite or rebuilt wall that combines the above

Axis EExtent / degree of vitrification

Code Melt fraction (vol %) Field description
I < 10 % Incipient: scattered glass skins
P 10–50 % Patchy / partial
T > 50 % Thorough / total

(Thresholds follow Wadsworth et al. Image-analysis can refine values.)

Axis PProcess / motive

Code Interpretation Diagnostic criteria
D Destructive attack or clearance burn Destruction layer sealed by glass; debris of conflict
C Creative/constructive firing Melt predates occupation; neat inner glass skin; no destruction layer
R Ritual closure / display Fuel deliberately piled; offerings deposited before firing
I Incidental / natural Localised lightning or wildfire melt only

(Choose the best-supported single code; append “?” if still uncertain.)

Axis TChronological slot

Code Calendar span (cal BC/AD)
LB ≥ 1200 BC (Late Bronze)
EIA 800–400 BC (Early Iron)
LIA 400 BC – AD 50 (Late Iron)
EM AD 50–700 (Early Medieval)
HM > AD 700 (High/Late Medieval)

(Use the best secure dating method; multiple dates → list both, e.g. EIA/EM.)

Axis LRampart layout & vitrification pattern

Code Plan description
S-V Single rampart, vitrified
D-IV Double circuit; inner wall vitrified
D-BV Double; both walls vitrified
D-OV Double; outer wall vitrified
M-PV Multivallate (>2); part-vitrified
M-MV Multivallate; multiple rings vitrified

(Add suffix “P” if vitrification is patchy within that wall.)

Axis RRampart lithology

Code Dominant stone group Vitrification suitability
MI Mafic igneous (basalt/dolerite/amphibolite) Very high
IF Intermediate–felsic igneous (andesite–granite) High
MS Metasedimentary (schist/psammite) Moderate
QS Quartz sandstone Low
CA Carbonate (limestone) Non-vitrifying (calcines)
MX Mixed / imported blend variable

( Optional ) Axis FFuel-load / firing strategy

Code Evidence
H High fuel (> 30 t per 100 m rampart) inferred from charcoal volume or modelling
M Moderate
L Low / opportunistic
U Unknown

Include only where charcoal quantification exists; otherwise omit the axis.

Putting it together

C – E – P – T – L – R (– F)

Examples

Fort Proposed code Explanation
Tap o’ Noth TLC-T-D-EM-D-IV-MI-H Timber-laced, thorough melt, destructive AD 5–600, double with inner vitrified, mafic rock, high fuel load
Wincobank TLC-I-D?-LIA-S-V-QS Timber-laced, incipient melt, likely destructive, Late Iron-Age, single rampart, quartz sandstone
Broborg TRF-T-D-EM-D-IV-IF Timber-revetted front, thorough melt, destructive, Early Medieval, double inner-vitrified, intermediate-felsic

Minimum dataset per rampart section

  • Axis codes (as above)
  • Melt fraction % (thin-section or 2-D image analysis)
  • Best firing date ± σ (method specified)
  • GPS coordinate & rampart sector ID

This ensures interoperability between reports and allows easy querying (e.g. “all TLC-T forts on MI rock in EM period”).

Immediate tasks to flesh out the model

  • Populate the scheme for the ~130 forts with detailed excavations.
  • Refine thresholds (e.g., confirm 50 % vs 66 % boundary for “T”).
  • Inter-observer calibration for petrographic melt counting.
  • Publish an open GIS layer so researchers can map any axis.

Next-phase research can then test cultural or migratory hypotheses by looking for statistically significant clusters along any combination of axes.

L’enigme des forts vitrifies

Mémo : RC-Forts vitrifiés (mise à jour mai 2001).

L’énigme des forts vitrifiés

C’est lors d’un voyage en Ecosse, au cours de l’été 1997, que nous avons entendu parler pour la première fois de forts vitrifiés. C’était au château d’Urqhart, au sud d’Inverness. Les monuments historiques fermant tôt, dans ce pays, nous ne pûmes voir que de loin cette impressionnante forteresse en ruine du XIIe siècle qui domine la rive occidentale du Loch Ness. Nous nous contentâmes juste de faire des photos au téléobjectif depuis le parking et de lire les panneaux retraçant l’historique du site. Sur l’un d’eux, la mention selon laquelle le château faisait partie de l’«ensemble des forteresses vitrifiées des Iles Britanniques » nous intrigua sans que nous puissions étudier de visu le phénomène de « vitrification ».

De retour en France, cette question nous préoccupa. Nous avions le vague souvenir d’avoir entendu parler de « forteresses vitrifiées » par le passé, sans pouvoir déterminer la source de cette information. Nous nous mîmes donc à tenter d’en apprendre plus sur le sujet mais nous dûmes rapidement nous rendre à l’évidence : il semblait être totalement ignoré des archéologues de notre pays alors que, de l’autre côté de la Manche, on fait référence aux forts vitrifés, presque comme s’il s’agissait d’une banalité.

En effet, au même titre que sur le panneau du château d’Urqhart, plusieurs ouvrages, achetés sur place, font assez souvent référence, sans cependant s’y étendre démesurément, à des « forteresses vitrifiés » [Vitrified hillforts]. C’est le cas, par exemple, de Scotland BC (« L’Ecosse avant J.-C »), dont un chapitre, consacré aux forteresses préhistoriques, aborde la question :

« Quand les premières fortifications écossaises furent-elles construites ? Il s’agit d’une question apparemment simple – mais pratiquement impossible à résoudre . Notre appréciation moderne de ce que l’on peut considérer comme des « défenses » peut ne pas recouvrir celle des peuples préhistoriques (…). Notre jugement repose sur la découverte de traces structurelles et d’armes. Sur cette base, la société préhistorique apparaît comme une société relativement pacifique au moins jusqu’au début du 1er. millénaire avant J.-C., à une exception près : un ouvrage massif entouré de palissades situé à Meldon Bridge, dans les Borders, mais cet ouvrage peut avoir été réalisé aussi bien dans un but de prestige que de défense. Cependant, vers la fin de l’âge du bronze, on trouve des preuves selon lesquelles la société amorça un changement et devint plus agressive. Les forgerons qui travaillaient le bronze commencèrent à produire en grande quantité des objets comme des épées et des boucliers dont la destination ne laissait aucun doute (…). Au même moment, on commença à construire les premiers forts de caractère défensif. Certains de ces forts furent construits en pierres liées (« laced ») avec des poutres pour les renforcer ; si un tel dispositif prenait feu, que ce soit accidentellement ou suite à une attaque ennemie, et si les conditions étaient réunies, la combustion des poutres provoquait la fusion des pierres qui, de ce fait se trouvaient amalgamées, avec pour conséquences la déformation du mur (on désigne ce phénomène sous le nom de « forts vitrifiés ») .

Le phénomène des forts vitrifiés est aussi presque systématiquement signalé dans une importante collection d’ouvrages dressant l’inventaire des monuments historiques de Grande-Bretagne (éditions PENGUIN). Voici, par exemple, ce qui en est dit dans l’introduction, dans le paragraphe consacré à l’âge du fer :

« Une renaissance économique semble avoir débuté vers 600 avant J.-C. avec le début de l’âge du fer, le travail du fer, particulièrement orienté vers la fabrication de charrues, ce qui permettait le développement de l’agriculture. La grande majorité des établissements de l’âge du fer visibles de nos jours étaient entourée de défenses. Les défenses atteignant les 375 m² sont appelées « dun ». Lorsque, bien que d’une technique comparable, elles couvrent une surface supérieure, on les appelle « forts ». Ces fortifications occupent généralement un promontoire, par exemple Brough of Stoll on Yell (Shetland), une hauteur, par ex. Craig Phadrig, Inverness, ou quelquefois un tertre, par ex. Dun-da-Lamh, près de Laggan (Badenoch and Strathspey), ou encore une île : Dun an t-Siamain, près de Carinish, sur l’île de North Uist (Western Iles). Leur dénominateur commun est d’avoir augmenté les défenses naturelles du site par la construction d’un rempart qui incorpore parfois un appareillage de poutres de bois, ce qui, si le feu est MIS à l’ensemble, soit par accident ou volontairement par des attaquants, peut déterminer un incendie d’une telle intensité qu’il provoque une fusion des pierres qui se transforment alors en une masse vitrifiée, comme c’est la cas à Craig Phadrig ou à Dun Ladaigh, vers Ullapool (Ross and Cromarty) (…) » .

Mais, en France, même dans les milieux archéologiques, nous n’avons rencontré que très peu de personnes ayant entendu parler du phénomène de vitrification et encore moins à s’y être intéressés. Le premier ouvrage dans lequel nous avons trouvé une amorce de réflexion sur ce sujet est un livre destiné au grand public de Jean MARKALE, auteur d’un grand nombre d’ouvrages sur les Celtes :

« Un autre système est assez curieux : il remonte dans le temps, puisqu’on a commencé à l’employer à la fin de l’âge du bronze, c’est-à-dire aux environs de 800 avant notre ère. Il s’agit du procédé dit de vitrification. On a longtemps cru qu’il s’agissait d’un phénomène déclenché par l’incendie d’une forteresse au cours d’un combat, mais en fait, cette vitrification a été provoquée délibérément pour des raisons tactiques. Le noyau du rempart est constitué par une masse calcinée très dure et entièrement compacte, formée de pierres et de sable, ce qui donne au résultat un aspect très proche du verre épais et grossier. Cette calcination n’a pu se produire que sur place, après qu’on eut mélangé du bois aux matériaux entassés et qu’on y eut mis le feu. C’est une technique que les archéologues reconnaissent comme difficile à réaliser, mais qui a l’avantage incontestable d’assurer un rempart d’une solidité à toute épreuve, comme dans le fameux camp de Péran, non loin de Saint-Brieuc (Côtes d’Armor), qui reste un modèle du genre. »

Ayant eu l’occasion d’aller en Bretagne, au cours de l’été 1998, nous avons saisi l’occasion pour nous rendre à Péran. Le site, très facile d’accès, à la différence de beaucoup d’Oppida que nous avons visités depuis, se trouve sur butte peu élevée, à quelques kilomètres du village de Plédran. A la sortie de l’agglomération, les panneaux routiers indiquent soit « camp romain » soit, encore plus curieusement « camp viking ». Sur le site même, des panneaux explicatifs, réalisés par le Centre archéologique de Péran (C.A.P.), font référence à une « destruction du camp par les Vikings ». On a en effet trouvé sur le site plusieurs objets attribués aux Vikings, « près du rempart ». Ces objets provenant de Grande Bretagne, en particulier des pièces de monnaies frappées à York vers le Xe. Siècle, on en a déduit, un peu hâtivement à notre avis, que le site, bien antérieur au Xe siècle puisque remontant à l’âge du fer, avait été détruit par les Vikings. Ce qui est plus étonnant, c’est qu’aucun de ces panneaux ne fait la moindre allusion au phénomène de vitrification, pourtant patent, comme on va le voir.

Le camp, de forme circulaire, se développe sur une circonférence d’environ 200 mètres. Une levée de terre est couronnée par les vestiges d’un mur dont les pierres sont littéralement fondues ensemble. Péran, pour reprendre les termes de Jean MARKALE, est effectivement un « modèle du genre ». De cela, nous pouvons en témoigner après avoir vu d’autres vestiges de forts vitrifiés beaucoup moins significatifs. Ici, le phénomène de vitrification saute aux yeux : on l’observe sur l’ensemble du rempart. Les pierres, d’origines géologiques diverses (mais toutes des roches dures : quartzites, dolérites, aplites ) sont fondues et collées entre elles (certaines ont même coulé, se transformant en un magma solidifié rappelant de la lave volcanique ) pour former une seule masse compacte.

Une partie du rempart a été reconstituée par les archéologues selon la technique du « murus gallicus ». Cette méthode, décrite par César dans la Guerre des Gaules, et attribuée par lui aux Gaulois (mais nous savons maintenant qu’elle remonte au moins à l’âge du fer), consistait à alterner poutres de bois et pierres.

M. Jean-Louis PAUTE, président du C.A.P., à qui nous nous étions adressé pour obtenir plus d’explications sur le site, nous a aimablement fait parvenir une brochure, éditée en 1991, qui retrace l’historique des recherches à Péran et leurs conclusions. Alors que, nous l’avons vu, aucune mention n’était faite de vitrification in situ, le texte de la brochure aborde largement le phénomène. Force est même de constater qu’il fut à l’origine de l’intérêt des archéologues du XIXe siècle pour cet Oppidum, comme il le fut, d’ailleurs, nous l’avons constaté depuis, pour la plupart des oppida vitrifiés connus. Le site date, à l’origine, de l’âge du fer et son occupation a duré jusqu’à l’époque carolingienne. Selon les archéologues l’ayant étudié, l’oppidum aurait été détruit par les Vikings vers 905-925 après J.-C. Leur hypothèse, pour expliquer la vitrification du rempart, est classique : pour eux, il ne fait aucun doute que l’incendie du « murus gallicus », mis à feu lors du sac de l’oppidum, est le seul responsable du phénomène de vitrification . A l’appui de leur affirmation, les auteurs invoquent les datations obtenues par le Carbone 14 et l’archéomagnétisme. On sait toutefois depuis quelques années, particulièrement pour le C 14, que l’on ne peut plus avoir une confiance aussi absolue dans ses indications, en particulier dans le cas où de fortes températures ont été en jeu ; il est en effet désormais admis que de telles circonstances ont pour conséquence de rajeunir démesurément les datations obtenues .

Mais un autre constat nous fait prendre ces datations avec précautions : en effet, si la destruction de Péran remontait au Xe siècle, son cas serait unique car ce serait la vitrification la plus récente que l’on connaisse ! Quant à l’affirmation selon laquelle le phénomène qui a transformé le rempart de Péran en un magma vitreux serait imputable à la combustion du poutrage interne, on verra qu’il s’agit d’une allégation gratuite, sans doute bien pratique pour « expliquer » l’une des plus grandes énigmes de l’archéologie, mais qu’elle est totalement contredite par l’expérimentation.

L’un des rares ouvrages français qui aborde la question de la vitrification de remparts de l’âge du fer, « Villes, villages et campagnes de l’Europe celtique » , nous apprend que cette question a préoccupé, depuis plus d’un siècle, de nombreux archéologues. Certains, et non des moindres , ont même essayé de reproduire le phénomène à grand renfort de moyens techniques, mais la plupart on dû reconnaître leur échec :

« La toponymie, les légendes populaires, et encore aujourd’hui la littérature archéologique font une large place aux « enceintes vitrifiées » ou « calcinées ». Dans la masse des remparts de pierre écroulés, des « noyaux de chaux » ou des blocs fondus et soudés par la chaleur ont été découverts sur environ 150 sites. La plupart d’entre eux se trouvent en Ecosse et dans le Massif central . Ils ont excité la curiosité des chercheurs, et des hypothèses de toutes sortes ont été émises pour expliquer ce phénomène.

« Au début du XIXe siècle, leur origine a été attribuée aux feux qu’auraient allumés les guetteurs pour transmettre des nouvelles à la ronde. En effet les auteurs de cette époque sont très préoccupés des relations d’enceinte à enceinte, et chaque description du site s’accompagne de considérations sur la surveillance du territoire. Une hypothèse plus audacieuse attribue les vitrifications à la foudre, qui aurait ainsi eu une prédilection particulière pour les remparts préhistoriques. Enfin certains auteurs imaginent qu’il s’agit d’une technique mise au point pour augmenter la cohésion des matériaux du rempart. Même si la réalisation d’un tel projet dans des roches cristallines suppose une quantité de bois considérable, il est facile de concevoir l’intérêt du procédé qui permettrait d’avoir un rempart plus solide qu’un mur en béton. En revanche, les noyaux de chaux, que des auteurs comme Drioton croyaient avoir reconnus au cœur de talus érigés dans des régions calcaires, semblent d’un intérêt plus limité.

« En 1930, G. Childe parvient à fondre des blocs au cours d’une expérience sur une reconstruction, mais le choix des matériaux utilisés a été critiqué. Youngblood montre en 1978 que la combustion de l’armature de bois d’un rempart à poutrage interne ne peut pas provoquer de vitrification si un feu n’a pas été délibérément provoqué et entretenu dans ce but. I. Ralson reprend l’expérience en 1981 avec un rempart long de 9 m, large de 4 et haut de 2,40 m. Il le garnit intérieurement de poutres horizontales entrecroisées, dont les têtes dépassent en façade. Plusieurs camions de bois ont été déversés devant le parement et enflammés. La température au cœur du rempart ne s’est élevée que très lentement. Elle s’est effondrée chaque fois que le vent, au lieu de rabattre les flammes vers le rempart, les éparpillait dans les autres directions. Dans les restes du rempart disloqué par la chaleur, quelques fragments vitrifiés ont pu être recueillis. Il est donc bien clair qu’il faut un feu intense, bien entretenu dans des conditions météorologiques favorables, pour obtenir une vitrification (…).

« Dans tous les cas qui ont pu être étudiés jusqu’à maintenant, l’action du feu ne laisse jamais de traces régulières, systématiques, qui pourraient seules être interprétées comme la preuve de l’emploi d’une technique de construction basée sur la combustion de la roche . Il s’agit toujours d’observations localisées ou de traces irrégulières, jamais d’un parement vraiment soudé par le feu. De plus I. Ralston a montré que la carte des enceintes vitrifiées ou calcinées correspondait assez exactement à la répartition des enceintes à poutrage interne, de la protohistoire jusqu’au Moyen Age .

« S’agit-il alors des traces de l’attaque des habitats fortifiés ? La technique de siège la plus répandue avant l’intervention romaine consiste en effet à cribler de projectiles le sommet des remparts pour en déloger les défenseurs, puis à mettre le feu aux portes avant de se ruer à l’intérieur. Il est peu vraisemblable que, en pleine action, les assaillants aient eu le loisir d’entretenir un feu suffisamment intense pour obtenir des vitrifications qui réclament, l’expérience l’a montré, beaucoup de combustible et un vent favorable. Certaines enceintes écossaises sont d’ailleurs vitrifiées sur tout leur pourtour . Nous imaginons volontiers que la vitrification est le produit d’une destruction systématique des fortifications de l’adversaire après la prise et souvent le pillage de la place, pour bien marquer le caractère irrémédiable de la défaite. »

On regrettera que cette intéressante analyse, l’une des plus développées que nous ayons pour l’instant trouvées sur le sujet, ne fasse référence à aucun site précis , si ce n’est pour leur dénier la qualification d’enceintes «vitrifiés » (i.e. le camp de Myard et du Châtelet d’Etaules, fouillés par J.-P. Nicolardot , en Bourgogne ), ni ne renvoie, dans la bibliographie citée, à aucune étude de référence sur la question…

En conclusion, il semblerait que toutes les tentatives qui ont à ce jour été faites pour tenter de reproduire le phénomène de vitrification se soient soldées par un échec. La raison en serait que la température obtenue, si elle a permis de rubéfier les pierres, n’a cependant jamais été suffisante pour les vitrifier, sauf sur de toutes petites surfaces.

D’après les géologues consultés , la vitrification de matériaux comme le granite ne peut se produire en dessous de à 1000°. Une telle montée en température ne semble pas pouvoir être atteinte à l’air libre mais seulement à l’intérieur d’un four. Comment imaginer qu’on ait construit un four tout autour d’une enceinte de pierres de plusieurs centaines de mètres (Péran fait partie des « petits » oppida mais il en existe de beaucoup plus étendus…). Cela exclut de facto l’incendie accidentel.

Penchons-nous un instant sur l’hypothèse, défendue par certains , selon laquelle la vitrification aurait été obtenue volontairement pour solidifier l’ensemble. Voire… Nous avons nous-même constaté que, si certaines parties de remparts étaient en effet rendues plus résistantes par la vitrification, d’autres, au contraire, avaient été fragilisées, en particulier lorsque, dans le magma, on trouve des noyaux de chaux, résultante de la combustion du calcaire… Le mur, au lieu d’être renforcé, s’effondre alors.

Une autre hypothèse, que nous n’avons trouvé évoquée par quiconque, serait l’utilisation d’un produit chimique dont on aurait enduit les pierres et qui, mis à feu, aurait dégagé une chaleur supérieure à 1000°. Reste à savoir quel produit aurait pu être utilisé ? Reste aussi à comprendre la raison qui aurait poussé nos ancêtres de l’âge du fer à vitrifier (volontairement) leurs oppida ou ceux de leurs ennemis.

Voici pour les faits. Aucune explication rationnelle ne pouvant à ce jour être obtenue par l’archéologie, nous rappelant que les oppida vitrifiés se trouvaient, de par leur ancienneté et leur situation, liés à l’aire celtique, et nous souvenant de l’intéressante réflexion initiée par Jean MARKALE à propos de Péran, nous avons décidé d’aller voir d’un peu plus près du côté de la mythologie celtique si elle pouvait nous apporter des pistes inédites.

« Il faut d’ailleurs voir dans ce procédé de vitrification l’origine des traditions concernant l’Urbs Vitrea, Care Gwtrin et autres lieux du genre Royaume de Gorre, c’est-à-dire des « Cités de Verre » qui se rencontrent si souvent dans les romans arthuriens et d’une façon générale dans toutes les traditions mythologiques irlandaises ou bretonnes. Les « Villes Blanches » de la tradition orale, qui désignent toutes d’anciennes forteresses ruinées, sont un souvenir évident de cette technique, par ailleurs parfaitement oubliée. »

Si Jean MARKALE a raison, si toutes les « cités de verre », « châteaux de verre » (et leurs variantes) des légendes celtiques signalent d’anciens « forts vitrifiés », les voies de recherche sont immenses !

Nous commencerons par le site le plus occidental de l’aire celte où l’on ait signalé des vestiges de vitrifications : l’Ile de Tory, située au large de la côte ouest de l’Irlande. Cette île doit son nom à une tour, l’une de ces nombreuses round towers (« tours rondes ») qui se dressent encore par centaines sur l’Ile Verte (ainsi qu’en moindre quantité en Ecosse) et dont on n’a jamais réussi à ce jour à savoir quel en avait pu être l’usage véritable. Généralement, on les date des tout débuts de l’ère chrétienne et on les relie à l’existence des premiers monastères ; on suppose qu’elles ont pu servir de tours de guet ou qu’elles servaient à entreposer le trésor des monastères menacés par les raids des Vikings. Comment, alors, on-t-elle résisté alors que les monastères qui les entouraient ont été détruits ? L’extraordinaire qualité de leur architecture, faite de pierres magnifiquement appareillées, explique sans doute en partie le mystère. Le fait est qu’une de ces tours a tellement marqué l’histoire de cette petite île que l’île elle-même a pris le nom de la tour (Tor-Inis signifie « Ile de la Tour »). Voici ce que nous avons pu lire à son sujet :

« La tour de l’île de Toriniz, aujourd’hui île Tory (…), d’âge si vénérable, n’existe plus en tant que bâtiment. Toutefois, elle perdure encore, ou du moins perdurait au siècle dernier, en tant que ruines. Le grand étonnement des archéologues fut de constater que ces vestiges étaient vitrifiés. Quelle peut être la raison de cette vitrification ? » .

Or, c’est précisément sur cette île que, selon la tradition mythologique irlandaise, les Fomoire (ou Fomore), avaient installé leur camp de base et c’est de là qu’ils lançaient leurs expéditions militaires sur l’Irlande. Ce peuple mythique était composé de géants maléfiques, à la fois alliés, par des liens familiaux complexes, et néanmoins ennemis, des Tuatha de Danaann (les « enfants de la déesse Dana ou Ana »). Aucun de ces deux peuples n’était autochtone ; ils venaient tous les deux des « Iles du Nord du Monde », endroit mythique situé dans l’extrême nord dont on ne sait pratiquement rien. Les uns et les autres étaient dotés de puissants pouvoirs, réputés magiques pour nos ancêtres mais que nos connaissances techniques actuelles nous rendent familiers.

L’un des chefs des Fomores, le géant Balor, par ailleurs grand-père du « dieu » Lug, un des Tuatha de Danaann, vivait sur Toriniz. De là, il envoyait un puissant flux d’énergie de l’autre côté du bras de mer qui le séparait de la terre d’Irlande pour foudroyer ses ennemis. Les descriptions que l’on a de Balor évoquent plus une machine qu’un être vivant : on le compare à un cyclope dont l’œil unique émettait un rayon qui réduisait en cendres ses ennemis : « C’est un géant effrayant dont l’unique œil foudroie toute une armée lorsqu’il soulève les sept paupières qui le protègent ». Cet « œil » extraordinaire devait être maintenu ouvert grâce à des crochets métalliques soulevés par plusieurs aides. Lors de l’une des trois batailles de Mag Tured qui se déroulèrent en Irlande entre les Fomore et les Tuatha de Danaann, le dieu Lug réussit à neutraliser l’œil maléfique de Balor en utilisant sa lance personnelle, que les textes appellent « lance d’Assal », l’un des quatre objets magiques rapportés des Iles du Nord du Monde . Cet objet avait lui aussi des propriétés étranges : il ne manquait jamais son but et devait être refroidi dans un chaudron rempli de sang humain .

Jean MARKALE, dans Les Celtes et la civilisation celtique, donne plus de détails sur cet instrument fabuleux :

« Lug est le possesseur d’une lance magique qui fait penser aux flèches à la fois meurtrières et guérisseuses d’Apollon. Elle s’appelle Gai Bolga . C’est l’emblème de l’éclair. Elle provient d’Assal, une des îles du nord du monde (allusion à l’Hyperborée) d’où étaient originaires les Tuatha Dé Danann. Cette lance avait un pouvoir venimeux et destructeur et, pour atténuer ce pouvoir, il fallait plonger la pointe dans un chaudron rempli de poison et de « fluide noir », c’est-à-dire de sang » .

Après avoir été lancée, en utilisant un cri particulier (« ibar », qui signifie « if »), et touché son but qu’elle ne ratait jamais (« sa valeur est telle qu’elle ne frappe pas par erreur ») elle revenait d’elle même dans la main du dieu grâce à un autre cri, « athibar » : « elle revient en arrière jusqu’à la main qui l’a lancée » .

Qu’était donc cet objet ? Quelles techniques étaient-elles utilisées par Balor et par Lug pour se faire la guerre ? N’est-il pas étrange que la Tour de Toriniz, l’endroit précis où résidait le géant Balor, ait été vitrifiée ?

Depuis la première version de ce mémo (10/12/98), nous avons eu connaissance des travaux de RALSTON sur les oppida du Limousin et, pendant l’été 1999, nous sommes allés dans la Creuse pour tenter de voir quelques uns des sites cités dans cet ouvrage comme portant des traces de vitrifications. Nous avons dû constater que le travail de RALSTON, qui est l’un des inventaires les plus complets qui soient sur les forteresses de l’âge du fer en France, fait une large place au phénomène de vitrification. Malheureusement, sur place, l’accès aux sites est souvent rendu difficile par une végétation débridée et un manque d’entretien flagrant.

Nous dûmes ainsi renoncer à nous rendre sur certains oppida que nous avions envisagé de voir en raison d’indications trop imprécises (ce fut le cas pour nous aux Muraux, commune de St. Georges de Nigremont). D’autres oppida, où l’on avait par le passé signalé des vitrifications, ont été détruits par une urbanisation sauvage et le manque de contrôle des instances dont ce devrait être le rôle: c’est déjà le cas de celui de Thauron qui ne bénéficie d’aucune mesure de protection. Lors de notre visite, nous avons même vu un brave homme démolir allègrement un mur de pierres qui se trouvait sur sa propriété et qui, d’après nos indications, était vraisemblablement l’un des seuls vestiges de l’oppidum…

Les autres sites visités, pour n’être pas urbanisés, n’en sont pas pour autant mieux protégés. C’est le cas du fameux Puy de Gaudy, au-dessus du village de Ste. Feyre, au sud de Guéret, qui est le lieu de rendez-vous de promenade favori des pensionnaires de la maison de retraite de la MAIF située juste au-dessous et surtout des « VTTistes » qui franchissent sans sourciller les murailles de l’oppidum. Dans ce département, pourtant fort riche en vestiges archéologiques et historiques, rien ne semble fait par les autorités pour protéger les sites dont elles ont la responsabilité (sur aucun des oppida visités nous n’avons vu la moindre mention « site archéologique » ni la moindre interdiction !).

Au cours de ces visites, par ailleurs assez décevantes pour les raisons que nous avons indiquées, nous n’avons relevé que peu de vitrifications indiscutables. Ce fut le cas sur l’un des sites, qui n’était pourtant pas réputé comme le plus représentatif : l’oppidum de Châteauvieux à Pionnat qui fait face au Puy de Gaudy. Elles sont pourtant tout-à-fait remarquables et indiscutables, mais le site, outre le fait qu’il n’est pas indiqué, est situé dans un bois à la végétation inextricab le. Nous avons pu cependant constater le même phénomène qu’à Péran : les pierres étaient littéralement « cimentées » entre elles par leur surface, certaines avaient fondu et s’étaient transformées en véritable « lave ». Sur place, la théorie lue ou entendue à propos de plusieurs sites, selon laquelle ces pierres vitrifiées étaient du machefer résultant d’un processus de fonte de métaux dans des haut-fourneaux répartis le long du rempart ne tient pas.

En effet, à Pionnat, nous avons compris comment l’on pouvait confondre certains éboulements du rempart, qui forment une sorte de voûte cimentée par la vitrification des pierres, avec le cul-de-four d’un haut-fourneau. Cette confusion, excusable pour des non-spécialistes cherchant à tout prix une explication satisfaisant le bon-sens, ne l’est plus lorsqu’on la lit sous la plume d’archéologues officiels. Il suffit pour cela d’observer le rempart dans sa continuité pour s’apercevoir que ce que l’on a pris un peu vite pour des vestiges de hauts-fourneaux ne possède aucune des structures nécessaires au fonctionnement d’un tel dispositif (prise d’air, etc.). Il s’agit bien, purement et simplement, d’une « bulle de lave » comme on en observe dans les formations volcaniques naturelles, sauf que, dans le cas qui nous occupe, nous avons affaire à une construction qui a été artificiellement vitrifiée et non à un phénomène naturel.

Depuis ce voyage dans la Creuse, nous avons bien entendu poursuivi nos investigations qui nous ont permis d’identifier avec certitude d’autres forts vitrifiés, en particulier deux dans la région de Roanne (en mai 1999), et un autre dans l’Allier (en mai 2000).

Nous avons trouvé que l’aire d’extension de ces structures, que nous pensions au départ limitée aux Iles Britanniques, était beaucoup plus large puisqu’elle s’étend sur une grande partie de l’ère dite celtique, bien que chronologiquement antérieure à l’arrivée des Celtes en Europe de l’Ouest, puisque remontant à l’âge du fer.

Bibliographie

– Françoise AUDOUZE et Olivier BUCHSENSCHUTZ (1989): Villes, villages et campagnes de l’Europe celtique. Paris, Hachette, 1989 (collection Bibliothèque d’archéologie), pp. 120-121.
– Base Mérimée (Ministère de la Culture) ;
– Les Celtes (1997), ouvrage collectif publié sous la dir. De Sabatino Moscati à l’occasion de l’exposition Les Celtes au Palazzo Grassi à Venise, 1991. Paris, Stock, 1997.
– Gordon CHILDE : Excavations of the vitrified Fort of Finavon, Angus.
– John GIFFORD (1992): The buildings of Scotland : Highland and Islands London, Penguin Books, 1992.
– Guide Bleu Bretagne, Paris, Hachette, 1992.
– Guide Bleu Grande-Bretagne. Paris, Hachette, 1990 (réed. 1994).
– Guide du Routard Ecosse 1996-1997. Paris, Hachette, 1995.
– E. KOARER-KALONDAN et GWEZENN-DANA (1973) : Les Celtes et les extra-terrestres. Verviers, Marabout, 1973 (ouvrage épuisé).
– KOUSNETSOV, IVANOV et VELETSKY (1989): « Effects of fires and biofractionation of carbon isotopes on results of radiocarbon dating of old textiles », in : Actes du symposium scientifique international du CIELT, Paris, 1989.
– Jean MARKALE (1997): Petite encyclopédie du Graal. Paris, Pygmalion, 1997.
– Jean MARKALE (1992): Les Celtes et la civilisation celtique. Paris, Payot, 1969 (rééd. 1992).
– Sabine MARCILLE (1999) : « Civilisation celtique : ces étranges cités vitrifiées », in : Efferve-Sciences, n°11 (1999).
– J.-P. NICOLARDOT (1991): Le Camp de Péran, Saint-Brieuc, Centre archéologique de Péran, 1991.
– Jean-Paul PERSIGOUT (1985): Dictionnaire de mythologie celte. Paris, Ed. du Rocher, 1985 (rééd. 1990).
– Ian B.M., RALSTON (1992) : Les enceintes fortifiées du Limousin. Paris, éd. De la Maison des Sciences de l’Homme, 1992 (Documents d’Archéologie Française, n°36).
– Anna RITCHIE (1988) : Scotland BC Edinburgh, 1988 (rééd. 1994), H.M.S.O. (coll. Historic Scotland).
– Divers sites Internet anglais, le plus complet (pour les sites britanniques) étant : Prehistoric Web Sites..

Remerciements

Nous remercions :

– Notre frère Yvon COMTE, Documentaliste à la Direction Régionale des Affaires Culturelles Languedoc-Roussillon, pour nous avoir indiqué quelques forts vitrifiés recensés en France au titre des Monuments Historiques. C’est aussi grâce à ses recherches sur Internet que nous avons eu connaissance du site anglais Prehistoric Web Sites qui donne, à notre connaissance, la liste la plus étendue de forts préhistoriques des Iles Britanniques (dont un certain nombre sont vitrifiés). Nous lui devons aussi de nous avoir mis en contact avec Michel WIENIN, chargé de l’étude du patrimoine industriel à l’Inventaire Général, Direction des Affaires Culturelles Languedoc-Roussillon, qui nous a proposé de faire analyser certains échantillons de vitrifications recueillies par nos soins par l’Ecole des Mines d’Alès. Nous attendons avec impatience le résultat de ces analyses…

– Denise BONJOUR, éternelle chercheuse aux frontières du réalisme fantastique, qui nous a adressé une copie d’un article récent de Sabine MARCILLE (1999). Outre que nous avons ainsi l’impression de nous sentir moins seuls, nous y avons trouvé quelques nouveaux sites qui ont enrichi notre documentation.

– Michel ROUVIERE, pour nous avoir retrouvé d ans sa bibliothèque Le livre des secrets trahis de Robert CHARROUX, livre que nous avions eu en main il y a bien des années et où nous avions entendu pour la première fois parler de « Cités vitrifiées ».

– Merci aussi à tous ceux dont les indications sur place nous ont permis de visiter certains oppida difficiles à trouver, en particulier : Mme DUQUESNE, à Lussac-les-Châteaux (86), M. Eugène MAZEROLLES (à Bègues, Allier), qui nous a offert quelques échantillons de vitrifications prélevées sur sa propriété, les mairies de St. Alban-les-Eaux et de Villerest (42) ainsi que tous les anonymes qui ont bien voulu nous consacrer un peu de leur temps.

INVENTAIRE DES FORTS VITRIFIES
(mise à jour août 2000)

– ECOSSE (10 sites)

1.1. An Cnap (Iles d’Arran) (source : site internet des Iles d’Arran) ;.
1.2. Barry Hill (Allyth, Perthshire) (“All that remains of this vitrified fort are a massive tumbled stone wall and subsidiary ramparts”, site internet d’Alyth);
1.3. Craig Phadrig, département d’Inverness (« Two vitrified walls enclose an area 75m by 25m. The inner wall stands 1.2m high on the inside”, site internet easyweb.easynet);
1.4. Dun Deardail (vers Lochaber, ouest de l’Ecosse) [source : site internet de Lochaber] ;
1.5. Dun Lagaidh, commune de Ullapool, département de Ross & Cromarty ) (« Massive stone rampart of the 1st millenium BC, now vitrified and so originally timber-laced (…)” [Source : HAI];
1.6. Finavon (Angus) [Source : CHILDE, 1935];
1.7. Knock Farril, commune de Strathpeffer, dépt. Ross & Cromarty, (Oblong hilltop fort of the 1st millenium BC, its stone rampart heavily vitrified so presumably originally laced with timber » ; Source : HAI);
1.8. Tor a’Chaisteal Dun, Ile d’Arran (Irlande) ;
1.9. Urqhart Castle (près d’Inverness) ;
1.10. White Caterthun (Incidemment signalé par RALSTON, 1992 comme étant vitrifié à propos du site français du Camp de César à CHATEAUPONSAC, Hte.-Vienne).

2) IRLANDE (1 site)

2.1.Tour de Toriniz, Tory Island (Irlande). Voir ce qui en est dit dans le texte ci-dessus..

3) FRANCE ( plus de 70 sites recensés, dont au moins une 20e vitrifiés)

3.1. Allier (03)

3.1.1. BEGUES : Oppidum de Bègues (« Rempart vitrifié », RALSTON, 1992). Une visite, en mai 2000, nous a permis d’obtenir la preuve de la vitrification. Echantillons prélevés .

3.2. Cantal (15)

3.2.2. COREN : Puy de la Fage [RALSTON (1992), p. 124]. Confusion possible avec La Fage-Montivernoux (Lozère).
3.2.3. ESCORAILLES : Pas de lieu-dit. [RALSTON (1992), p. 124].
3.2.4. LA COURBE : « Le Château Gontier ». Attention ! : risques de confusion avec La Courbe, près d’Argentan (Orne).
3.2.5. MAURIAC (a) : « Vieux Château », hameau d’Escoalier* (« Certains indices laissent à penser que les deux communes voisines peuvent avoir chacune une enceinte vitrifiée », RALSTON, 1992). * Confusion possible avec Escorailles (voir ci-dessus).
3.2.6. MAURIAC (b)
3.2.7. MAURIAC (c)

3.3. Charente (16)

3.3.1. MOUTHIERS-SUR-BOEME : Pas de lieu-dit [RALSTON (1992), p. 124].
3.3.2. VOEIL ET GIGET : « Camp des Anglais ou de la Pierre Dure » (« Traces de calcination sur toute la longueur du talus, 210 m de long, 5-6 m de haut et 25 m de large. La surface de cet éperon barré couvre 3 ha env. », RALSTON, 1992). Bien qu’on ne parle que de « calcination », le toponyme de « Pierre Dure » pourrait être un indice de vitrification.
3.3.3. SOYAUX : « Camp de Recoux ». [RALSTON (1992), p. 124].

3.4. Cher (18)

3.4.1. LA GROUTTE : « Camp des Murettes » (ou « de César »). « S’étend sur 4 ha. » [RALSTON (1992), p . 124].

3.5. Corrèze (19)

3.5.1. LAMAZIERE-BASSE : « Champ du Châtelet » au lieu-dit Les Bessades (« Pierres vitrifiées dans l’éboulement à l’extrémité orientale du mur intérieur. Autres gneiss vitrifiés, non seulement dans l’éboulis du rempart mais également à l’extrémité orientale du mur à l’extérieur. La vitrification semble être limitée à cette partie du site », RALSTON (1992), pp. 46-47. Dans sa bibliographie, l’auteur renvoie à VAZEILLES : Station vitrifiée avec muraille vitrifiée du Châtelet, commune de Lamazière-Basse (Corrèze).)
3.5.2. MONCEAUX S/DORDOGNE : « Puy de la Tour » ou « du Tour » (« Un bloc de pierre vitrifiée », RALSTON (1992), pp. 49-53).
3.5.3. MONCEAUX S/ DORDOGNE : « Puy Grasset » ou « Granet » au lieu-dit « Le C(h)astel » ou « le Chastelou » au village de Raz. « Quelques traces de vitrifications visibles dans les roches schisteuses du sommet de la motte (qui serait d’époque médiévale ». « Desbordes décrit l’enceinte comme vitrifiée et, sans doute, médiévale ». [RALSTON (1992), p. 53]
3.5.4. ST. GENIEZ-Ô-MERLE : « Puy de Sermus » ou « Vieux Sermus ». (« L’indice principal de fortifications consistait en un tronçon de mur vitrifié, haut de 1,5 m et long de 3 m, situé sur le côté nord-ouest du site où l’accès était le plus facile. (…) Les défenses reconnaissables consistaient en deux tronçons de murs vitrifiés avec une pente artificielle à l’extérieur. (…) Une fouille sur le côté nord-ouest a montré que le mur vitrifié était construit directement sur la roche, qui présentait quelques signes d’une vitrification superficielle. », RALSTON, 1992.)
3.5.5. ( ?) ST. PRIVAT : « Camp de Srmus » (ou « Sermus ») [MARCILLE, p. 17]. Il doit s’agir d’une confusion avec le précédent.

3.6. Côtes d’Armor

3.6.1. PLEDRAN : Camp de Péran [MARKALE (1997), pp. 133-135. Site visité en juillet 1998. La vitrification est patente sur l’ensemble du site, qui était parfaitement dégagé lors de notre visite. La roche est fondue et amalgamée en de gros blocs soudés ensemble. Echantillons prélevés. Voir ce que nous en disons plus haut.

3.7. Côte d’Or (21)

3.7.1. BOUILLAND : « Le Châtelet ». Sans autre précision.
3.7.2. CRECEY-SUR-TILLE : « Camp de fontaine Brunehaut » [RALSTON (1992), p. 124. Il ne dit pas s’il est vitrifié ou non].
3.7.3. CHAMBOLLE-MUSIGNY : « Enceinte de Groniot » (ou « Gromiot »). Sans autre précision.
3.7.4. ETAULES : « Le Châtelet » (« Restes carbonisés de poutres », RALSTON, 1992). On ne parle pas de vitrification. MARCILLE (1999), p. 17 cite « Le Chevalet ». Nous pensons qu’il doit s’agir d’une graphie erronée.
3.7.5. FLAVIGNEROT : « Camp de César » dit aussi Enceinte du Mont Afrique. [RALSTON (1992), p. 125].
3.7.6. GEVREY-CHAMBERTIN : « Enceinte du Château-Renard » [RALSTON (1992), p. 125].
3.7.7. MESSIGNY : « Enceinte de Roche-Château ». Pas d’autres précisions.
3.7.8. PLOMBIERES-LES-DIJON : « Enceinte du Bois brûlé » [RALSTON (1992), p. 125]. Le toponyme de « Bois Brûlé », rencontré sur d’autres sites, peut être la confirmation d’une vitrification.
3.7.9. VAL SUZON : « Le Châtelet de Val Suzon ou de la Fontaine du Chat ». « Situé juste en face du Châtelet d’Etaules, de l’autre côté de la vallée » [donné comme étant vitrifié par MARCILLE (1999), p. 17]. (« Couche brûlée », selon RALSTON (1992).
3.7.10. VELARS S/OUCHE : « Enceinte de Notre-Dame de l’Etang ». Donné comme vitrifié par MARCILLE (1999), p. 17.
3.7.11. VIX : « Mont Lassois » (« La ‘levée de terre’ méridionale (…) semble avoir été construite sur un niveau brûlé sur lequel reposent des pierres parfois rubéfiées ou calcinées », RALSTON, 1992). On ne parle pas de vitrification.
Ainsi, peut-être, que d’autres sites d’après les études de Nicolardot (cité par RALSTON, 1992).

3.8. Creuse (23)

3.8.1. AUBUSSON : « Camp des Chastres ». Traces de vitrifications [RALSTON (1992), p. 70].
3.8.2. & 3.8.3. BUDELIERE : « Promontoires de St. Marien » et de « Ste. Radegonde ». Nous nous sommes rendus sur le promontoire de Ste Radegonde en 1999 à partir des indications de RALSTON mais le site étant très embroussaillé, nous n’avons pu observer de vitrifications.
3.8.4. JARNAGES : Enceinte sous le nom de « Château ».
3.8.5. PIONNAT : Oppidum au village* de « Châteauvieux ». [RALSTON (1992), pp. 75-79]. * Le hameau de Châteauvieux est distant de Pionnat de plusieurs km. Enceinte ovale de 128 m de longueur axiale. Malgré l’embroussaillement du site, notre visite de l’été 1999 nous a permis de confirmer l’existence d’une vitrification importante, bien visible et indiscutable : plus encore qu’à Péran, les pierres sont fondues et amalgamées entre elles. On voit même des traces de coulures, comme dans le cas de laves volcaniques. La chaleur a dû être d’une intensité extrême. Certaines descriptions du site parlent de vestiges de fours à chaux ou de fours à métaux. Nous pensons qu’il s’agit d’une mauvaise interprétation des observations faites par des personnes qui n’avaient jamais vu de vitrifications. Pour nous, il ne fait aucun doute que Pionnat montre des traces patentes de vitrifications. Echantillons prélevés. Un autre site, « Ville de Ribandelle (ou Ribaudelle ») lui ferait face.
3.8.6. STE. FEYRE : « Puy de Gaudy ». Visité à la même époque. Même observation qu’à Pionnat mais la dégradation du rempart et son embroussaillement ne nous ont pas permis de constater des traces évidentes de vitrifications. [RALSTON (1992) ; base Mérimée].
3.8.7. ST. GEORGES DE NIGREMONT : « Les Muraux » (ou « le Muraud »). [RALSTON (1992), pp. 80-81. Malgré les indications de RALSTON, complétées par des indications recueillies sur place auprès des habitants qui connaissaient l’existence du site, nous n’avons pu identifier l’emplacement de l’oppidum des Muraux. Ils nous ont conseillé de contacter Monsieur EUCHER à Rouzelie, qui avait fouillé le site, ce que nous n’avons pu faire.
3.8.8. THAURON : Village. Visite décevante : inutile de rechercher des traces de vitrifications. Le site a été totalement détruit par l’extension anarchique du village, installé sur l’oppidum et rien n’est fait pour préserver ce qui pourrait en subsister.

3.9. Dordogne

3.9.1. PERIGUEUX : Enceinte du « Camp de la Boissière » située sur l’un des contreforts de la rive droite de l’Isle en face de Périgueux. Ne pas confondre avec Périgneux (Loire).
3.9.2. ST. MEDARD D’EXCIDEUIL : « Castel Sarrazi » à Gandumas (« Au moins deux ouvrages distincts et profondément vitrifiés sont conservés (…). Les traces de la position originelle du poutrage ont été reconnues dans la masse vitrifiée (…). » [RALSTON (1992)]. Même observation à Péran.

3.10. Doubs (25)

3.10.1. MYON : « Châtelet de Montbergeret » [RALSTON (1992), p. 126].

3.11. Finistère (29)

3.11.1. ERGUE-ARMEL : « Berg-ar-Castel » [RALSTON (1992), p. 126]. Rien n’indique que ce site soit vitrifié.
3.11.2. HUELGOAT : « Camp d’Arthus ». « Rempart massif secondaire, non daté, recouvrant un murus gallicus de type Avaricum. » [RALSTON, 1992, p. 132]. Site visité au cours de l’été 1998. Malheureusement l’étendue du site, par ailleurs très embroussaillé, ne nous ont pas permis d’étudier si certaines parties montraient des traces de vitrifications.
3.11.3. LOSTMARC’H (près de CROZON). [Les Celtes, p. 586].

3.12. Ille-et-Villaine (35)

3.12.1. VIEUX-VY-SUR-COUESNON : « Oppidum d’Orange » [RALSTON (1992), p. 126.

3.13. Jura (39)

3.13.1. SALINS : « Camp du Château-sur-Salins ». « Matériau calciné sur environ 4 m de long sur le côté ouest du rempart préhistorique (…). » [RALSTON (1992), p. 127].

3.14. Loire (42)

3.14.1. PERIGNEUX : « Pic de la Violette » (« Un plateau à 650 m d’altitude est décrit comme partiellement enclos par de faibles murailles et des blocs vitrifiés », [RALSTON (1992), p. 127]. Ne pas confondre avec Périgueux (Dordogne).
3.14.2. ST. ALBAN-LES-EAUX : « Châtelus » [ ou « Châtelux », chez RALSTON, 1992, graphie manifestement erronée]. (« Matériaux vitrifiés retrouvés sur place », RALSTON, 1992). Visite superficielle du site en mai 1999, les broussailles et la mousse recouvrant les pierres m’ont empêché d’identifier tout vestige de vitrification. Cela ne veut pas dire qu’il n’y en ait pas. Sur place, les gens connaissent le site sous le nom de « Château de verre », appellation qui me paraît suffisamment significative pour qu’on puisse admettre ce site dans la liste des vitrifications (voir MARKALE).
3.14.3. VILLEREST : « Le Château-Brûlé » à Lourdon. Visite à Villerest en mai 1999. Je n’ai pu accéder au site mais la mairie m’a communiqué le résultat de fouilles effectuées par Stéphane Boutet (cité par MARCHAND, 1991), qui confirme l’existence d’un « rempart vitrifié » et le rapproche du « Château de verre » de ST. ALBAN-LES-EAUX ; ce texte indique en outre : « Ce type de rempart vitrifié n’est pas unique dans notre région ». A propos de l’appellation « château de verre », même remarque qu’au-dessus, voir ce qu’en dit MARKALE.
3.14.4. ST. BONNET-DES-QUARTS (région de ROANNE) : « Oppidum des Carres (ou des Quarts) ». Nous nous sommes rendus sur place en mai 1999, mais nous n’avons pu situer l’endroit de l’oppidum..

3.15. Lot (46)

3.15.1. CRAS : « Murcens ». [RALSTON (1992), p.127].
3.15.2. LUZECH : « L’impernal ». [RALSTON (1992), p. 127].

3.16. Lozère (49)

3.16.1. LA FAGE-MONTIVERNOUX : « Puy de la Fage ». [RALSTON (1992), p. 127]. Attention : confusion possible avec « Le Puy de la Fage » dans le Cantal (commune de COREN). Souvent ces sites remarquables ont été pris comme limite de plusieurs communes, ce qui peut induire des erreurs d’attribution à telle ou telle commune.

3.17. Mayenne (53)

3.17.1. LOIGNE-SUR-MAYENNE : « Les Caduries ». Vitrifié [RALSTON (1992), p. 127].
3.17.2. ST. JEAN-DE-MAYENNE : « Enceinte de Château-Meignan ». [RALSTON (1992), p. 127 ; MARCILLE (1999), p. 17].
3.17.3. STE. SUZANNE : « Le Château ». Il semble que nous ayons affaire à deux sites distincts : « Le Château » et « Le Camp des Anglais ». Selon RALSTON (1992) : « Le matériau vitrifié provenant de cette commune vient du pied du château de Ste. Suzanne et non du Camp des Anglais » .

3.18. Meurthe-et-Moselle

3.18.1. CHAMPIGNEULLES : « Enceinte de la Fourasse » (ou « Tourasse). Traces de vitrifications selon RALSTON (1992), p. 127. MARCILLE (1999).
3.18.2. ESSEY-LES-NANCY : « La butte (ou enceinte) Ste. Geneviève » (« Traces d’incendie du rempart encore observables sur le côté ouest du site », RALSTON, 1992). On ne parle pas de vitrification.
3.18.3. MESSEIN : « La Cité », ou « Le Camp d’Affrique » ou « le Vieux-Marché » (« Traces de calcination », RALSTON, 1992). On ne parle pas de vitrification.
3.18.4. SION-COUVENT : Lieu-dit non précisé. RALSTON (1992) parle de « Traces de calcination », non de vitrifications.

3.19. Morbihan

3.19.1. LANDEVANT-KERVARHET : « Kervarhet » (« Enceinte d’un diamètre de 200 m. Traces de vitrification », RALSTON, 1992).

3.20. Moselle (57)

3.20.1. LESSY : Pas de lieu dit [RALSTON (1992), p. 128].

3.21. Nièvre

3.21.1. LA MACHINE : « Enceinte du Vieux Château » ou « Cité de Barbarie ». Vitrification certaine [MARCILLE (1999), p. 18, avec un plan].

3.22. Oise

3.22.1. GOUVIEUX : Camp de César (« Traces de calcination », « Rempart massif secondaire élevé au-dessus d’un rempart à poutrage interne brûlé : ce dernier n’a pas été daté. » RALSTON (1992), p. 132. On ne parle pas de vitrification.

3.23. Orne (61)

3.23.1. ARGENTAN : Fort vitrifié. Il se peut qu’il s’agisse du même site que le suivant.
3.23.2. LA COURBE : « Le Haut du Château ». (« Traces étendues d’une combustion intense, notamment des blocs vitrifiés et des pierres de revêtement altérées par la chaleur », RALSTON, 1992).

3.24. Puy-de-Dôme (63)

3.24.1. et 3.24.2. CHATEAUNEUF-LES-BAINS : « Montagne de Villars » (« Le matériau vitrifié vient d’un rempart long de 14 m qui forme l’un des côtés d’un rectangle de 7 m X 15 m couronnant une butte de pierres », RALSTON, 1992). Une visite sur place en mai 2000 ne nous a pas permis de trouver le site mais un habitant du hameau de Villars, à qui nous avons demandé notre chemin, connaissait l’existence de « pierres fondues ». Il nous a indiqué qu’il n’y a avait pas un seul site, sur lequel on trouvait ces pierres, mais deux. D’après ses explications, nous avons compris qu’il s’agissait de deux oppida, proches l’un de l’autre.

3.25. Haut-Rhin (68)

3.25.1. HARTSMANWILLER : « H artmannswillerkopf ». « Des vestiges d’une enceinte vitrifiée d’époque protohistorique existaient au Hartmannswillerkopf mais ont été détruits durant la bataille du Vieil Armand en 1914 et 1915. » [Base Mérimée du ministère de la culture].

3.26. Haute-Saône (70)

3.26.1. BOURGUIGNON-LES-MOREY : Pas le lieu-dit. « Site de 3 ha. Traces de calcination dans un talus de pierres épais d’env. 3 mètres (On ne parle pas de vitrification) [RALSTON, 1992, p. 129].
3.26.2. MACHEZAL* : Crêt Chatelard (« Matériau vitrifié mais il n’est pas certain qu’il sot en rapport avec une vitrification » RALSTON, 1992). * Près de CHIRASIMONT, au S/E de Roanne.
Une autre source indique aussi un « tumulus burgonde et une tombe aux murs vitrifiés ». Il s’agit sans doute du même site.
3.26.3. NOROY-LES-JUSSEY : Pas de lieu-dit : « Enceinte de 2,5 ha. Traces de calcination des fortifications. » [RALSTON (1992), p. 129]. On ne parle pas de vitrification.

3. 27. Vienne (86).

3.27.1. ASLONNES : Camp d’Alaric (« Pautreau semble admettre que cette fortification a été calcinée », [RALSTON (1992)].
3.27.2. CHATEAU-LARCHER : « Site de Thors ou Thorus ». « Fort vitrifié ; les ruines n’ont pas été fouillées. »
3.27.3. QUINCAY : « Camp de Séneret (ou Céneret) » entre Quinçay et Vouillé (« Traces de calcination », [RALSTON (1992) et MARCILLE (1999)].

3.28. Haute-Vienne (87)

3.28.1. CHATEAUPONSAC : « Chégurat ou Camp de César ». Les vitrifications sont comparées à celles de White Catherhurn (Ecosse) [RALSTON (1992), pp. 89-90].
3.28.2. SEREILHAC-LA-BAISSE : Pas de lieu-dit (« Pierres vitrifiées qui ne semblent pas associées à une fortification ». [RALSTON (1992), p. 129].

3.29. Yonne (89)

3.29.1. ST. MORE : « Camp de Cora » [RALSTON (1992), p. 130].

Contact Us
close slider