Mystery of Vitrified Forts

It was during a trip to.Scotland, in the Summer of 1997, that I first heard of vitrified hillforts. It was near Urqhart Castle, south of Inverness. As ancient monuments close early in that country, I could only see from a distance that impressive ruined fortress overlooking the west bank of Loch Ness. I had to be content with taking telephotographs from the car park and reading the notices on boards recounting the history of the site. I was intrigued by a sentence on one of them, saying that the castle was part of “the whole group of vitrified hillforts in the British Isles”, but I had no opportunity of studying the phenomenon of vitrification de visu .

Once back in France, the question still puzzled me. I had a vague memory of having heard of vitrified hillforts in the past, but could not find the source of the information. So I tried to learn more on the subject, but soon had to bow to facts : it seemed totally unknown to archaeologists in our country, whereas, on the other side of the Channel, vitrified hillforts are referred to as quite common.

Indeed, just as the notice board by Urqhart Castle, several books I had bought on the spot often mention “vitrified hillforts” without dwelling too much on the subject. It is the case, for instance, of Scotland B.C. , in which a chapter dealing with prehistoric fortresses tackles the question:

When were Scotland’s earliest fortifications built? This is a deceptively simple question – and virtually impossible to answer 1. .Our modern appreciation of what ranks as defences may not match the view of prehistoric people, even if we had a complete record of what they built, and our judgement must depend upon structural traces and discoveries of weapons. On that basis, prehistoric society appears to have been relatively peace-loving until the early part of the first millennium BC, with one possible exception : massively stockated enclosure built in late neolithic times at Meldon Bridge in the Borders, but even that may have been motivated by prestige rather than defence. Towards the end of the bronze age, however, there are clues to suggest that society was changing and becoming more aggressive. Bronzesmiths began to produce in large quantities items such as swords and shields that can only be weapons and defensive arms rather than equipment for hunting […] At the same time, the first defensive forts were being built. Some of the earliest forts were those built with stone walls laced with timbers to strengthen them; if such a fort were set on fire, either accidentally or by enemy attack, and if conditions were right, the burning timbers caused the stone work to melt and fuse together and the wall to become distorted (these are known as vitrified forts) 2 .

Besides, vitrified hillforts are almost systematically mentioned in an important collection of books drawing up an inventory of the historical monuments of Great Britain (Penguin). Here is, for instance, what the introduction says in the paragraph dealing with the iron age :

An economic revival seems to have begun c.600 BC, with the beginning of the iron age, the working of iron, especially for ploughs, making possible a much improved agriculture. The great majority of iron age settlements now visible have been protected by defences. Some, whose defences enclosed an area up to 375 sq.m., are classified as duns, others, generally similar but larger, as forts. These usually occupied a promontory, e.g. Brough of Stoal on Yell (Shetland), a hilltop, e.g. Craig Phadrig, Inverness, or sometimes a knoll, e.g. Dun-da-Lamh, near Laggan (Badenoch and Strathspey), or an island, e.g. Dun ant-Siamain ,nearCarinish, on North Uist (Western Isles). Their common feature is the supplementing of the site’s natural defence by a stone rampart that has sometimes incorporated a timber framework, which, if set on fire by accident or an attacker, could burn with such intensity as to fuse the stone into a vitrified mass, as at Craig Phadrig or Dun Lagaidh, near Ullapool (Ross and Cromarty) 3.

But, in France, even among archaeologists, I have met very few people who have heard of vitrification, and still fewer who have been interested by the subject. The first study in which I found the beginning of some thought on the subject was a book aimed at the general public, by Jean MARKALE, the author of many books on the Celts :

Another system is strange enough : it goes back a long way, since it was first used at the end of the bronze age, i.e. around 800 BC. It is the process of vitrification. It was long believed to be a phenomenon brought about by fire in a fortress during a battle, but in fact, this vitrification was deliberately started for tactical reasons. The core of the rampart is made of a very hard and totally solid burnt mass, made up of stones and sand, which gives the final product an aspect very close to that of thick coarse glass. This calcination could only have taken place on the spot, after some wood had been mixed with the heaped up material and set ablaze. Archaeologists admit that this is a difficult technique, but it has the unquestionable advantage of providing a thoroughly reliable rampart, as in the famous Camp de Péran, not far from Saint-Brieuc (Cötes d’Armor), which stands as a perfect example 4.

As I happened to visit Brittany during the Summer of 1998, I seized the opportunity to go to Péran. This site, unlike the numerous oppida I have visited since, stands on a small hillock, a few kilometres from Plédran village. As one leaves the centre, road signs point either to “Roman camp” , or more strangely to “Viking camp” . On the site itself, explanatory notices, drawn up by the Péran archaeological centre (C.A.P.), refer to a “destruction of the camp by the Vikings“. Actually, several objects attributed to the Vikings were found on the site, “near the rampart”. As these came from Great Britain, especially coins minted in York around the Xth century, it was inferred – a bit hastily I think – that the site had been destroyed by the Vikings in the Xth century. What is more surprising is that none of the notices alludes to vitrification, though the phenomenon is obvious, as will now be shown.

The circular camp extends about 200 metres in circumference. A levee is topped by the remains of a wall whose stones are literally melted together. Péran, in Jean MARKALE’s words, is indeed a “model of the kind”. I can testify to this after seeing other vestiges of much less characteristic hillforts. Here the vitrifying process is obvious, and is to be observed on the whole rampart. Some of the stones, of various geological origins – ‘but all of them hard (quartzites, dolerites, aplites) 5 -have even flowed, and are melted and stuck together, turning into a solidified magma recalling volcanic lava 6 , and forming a dense bulk.

One part of the rampart has been reconstructed by archaeologists using the murus gallicus technique, described by Caesar in De bello gallico, and consisting in alterning timber and stones.The roman general ascribed it to the Gauls, but we now know it dates back, at least, to the iron age M. Jean-Louis PAUTE, chairman of the C.A.P., whom I contacted to obtain further explanations about the site, kindly sent me a booklet, published in 1991, which relates the history of research in Péran and its conclusions. Whereas, as said above, no mention was made of vitrification on the spot, the booklet amply deals with the phenomenon. We must even admit that it aroused the XIXth century archaeologists’ interest in this oppidum, and also in most of the known vitrified oppida, as I have noted since. The site dates back to the iron age and was inhabited until the Carolingian period. According to the archaeologists who studied it, the oppidum may have been destroyed by the Vikings about 905-925 AD. They explain the vitrification of the rampart through a classical hypothesis : undoubtedly the fire in the murus gallicus , started during the storming of the oppidum, was the only cause of the vitrification phenomenon 7 . To support this statement, the authors refer to dating through Carbon 14 and archaeomagnetism. Yet it has been known for a few years that, particularly in the case of C14, one cannot totally rely on its information, especially when high temperatures have been at work : indeed it is now acknowledged that under such circumstances, dating goes to a much more recent period 8 .

But another observation causes me to use this dating cautiously. For, if the destroying of Péran dated back to the Xth century, it would be a unique case, indeed the latest vitrification known so far. And saying that what turned the rampart of Péran into a glassy magma was the burning of the inner timber is a groundless statement, certainly very useful to “explain” one of the greatest archaeological riddles, but wholly contradicted by experimentation.

One of the few French studies tackling the question of iron age rampart vitrification, Villes, villages et campagnes de l’Europe celtique 9 reveals that, for more than a century, many archaeologists have been concerned with the problem. Some outstanding ones 10 have even tried to reproduce the phenomenon with a copious supply of technical devices, but most of them had to admit their failure 11.

Toponymy, popular legends, and evenl to-day, archaeological literature profusely deal with “vitrified” or “calcined” walls. In the mass of crumbled stone ramparts, “lime knots” or melted blocks welded by heat have been discovered on about 150 sites. Most of them are to be found in Scotland and in the Massif central 12.They have aroused the research workers’ curiosity, and all kinds of theories have been put forward to explain the phenomenon.
At the beginning of the XIXth century, their origin was believed to be due to fire kindled by watchmen to send news around. Indeed at this period, authors are much preoccupied with the relationships between walled sites, and each description of a site contains reflexions on territory-watching 13. A bolder 14 theory attributes the vitrifications to lightning, which would have evinced a particular preference for prehistoric ramparts. Lastly, some authors imagine that a technique was worked out to increase the cohesion of materials. Even if realising such a project with crystalline rocks implies using a huge quantity of wood, it is easy to understand how profitable would be a process allowing to obtain a rampart stronger than a concrete wall. Conversely, the lime-knots authors like Drioton thought they had discovered inside banks set up in chalky areas, seem to possess a more limited interest.
In 1930, G. Childe managed to melt blocks during some experiment on a reconstruction site, but his choice of materials was criticised . In 1938,Youngblood showed that burning the internal wooden frame of a timbered rampart cannot bring about vitrification if fire has not been deliberately started and kept ablaze for that purpose. I. Ralston renewed the experiment in 1981 on a 9 metre long , 4 metre wide and 2.40 metre high rampart. He furnished the inside of it with horizontal intertwined beams, with their heads sticking out on the front. Several truckloads of wood were stacked in front of the facing wall and set on fire. The temperature inside the rampart only rose very slowly. It went down every time the wind blew the flames away from the rampart. In the remains of the rampart broken up by the heat, a few vitrified fragments could be gathered. It clearly appears that, in order to obtain vitrification, one must have a strong well-kept up fire, under favourable weather conditions […] In all the cases studied so far, the action of the fire never leaves regular systematic traces, which alone could be interpreted as proving the use of a building technique based on burning the rock 15 . Observations are always limited in space or traces are irregular. No facing wall is really welded by the fire. Moreover I. Ralston has shown that the map of vitrified or calcined ramparts quite precisely corresponded to the distribution of internally timbered ramparts, from protohistory to the Middle Ages 16 .
Then are the traces those of the attack of fortified dwellings ? The most common besieging technique before the Roman invasion consisted in riddling the rampart with projectiles to dislodge the defenders, then in setting the gates on fire before rushing in. It is hardly likely that, in the middle of the battle, the attackers could manage to keep a strong enough fire to obtain vitrifications which have been proved to require much fuel and a favourable wind. Some Scottish ramparts are even vitrified all around 17. I am inclined to imagine that vitrification is the result of a systematic destroying of the enemy’s fortification after the place had been captured and often plundered, to emphasise the irreversible nature of the defeat.

One may regret that this interesting analysis, one of the most elaborate we have found on the subject so far, refers to no precise site 18, except in order to deny them the quality of “vitrified” ramparts (e.g. the Camp de Myard and the Châtelet d’Etaules, investigated by J.P. Nicolardot 19 in Burgundy); neither does it mention, in its bibliography, any reference study on the question.

To conclude, it appears that all the attempts made so far at reproducing the phenomenon of vitrification have failed. This is probably due to the fact that the temperature could redden the stones, but was never high enough to vitrify them, except on very small areas.

According to the geologists I have consulted 20, the vitrification of materials like granite cannot be obtained under 1000° C. Such a high temperature cannot be reached in open air, but only inside a furnace. How can we imagine that a furnace was built all around a stone rampart several hundred metres long ( Péran is one of the “small” oppida, but there exist bigger ones…) This de facto excludes any accidental fire.

Let us consider for a moment the theory upheld by some authors 21, that vitrification was deliberately caused to strengthen the whole structure. Indeed… I have myself observed that, if some sections of the ramparts have really been strengthened by vitrification, others, on the contrary, have been weakened, especially when, within the magma, lime-knots are to be found, as a result of burnt chalk. Then, instead of getting stronger, the wall collapses, leaving a wide gap in the defensive structure, thus weakening the whole ensemble .

Another theory, would be the use of a chemical product, spread on the stones, and set ablaze, thus producing a heat higher than 1000°C. There would remain to discover the nature of such a product, and to understand why our iron age forefathers should have been led deliberately to vitrify their oppida or those of their enemies.

So much for facts. No rational explanation can be provided to this day by archaeology. So, remembering that vitrified oppida, through their antiquity and location, belong to the Celtic area (even if they cannot actually be attributed to the Celts), and recalling the interesting observation proposed by Jean MARKALE about Péran , I decided to see if a closer scrutiny of Celtic mythology could reveal any new prospects.

This method of vitrification must be considered as the origin of traditions concerning the Urbs Vitraea, Care Gwtrin and other places like the Kingdom of Gorre, i.e. the “Glass Cities” that are so often alluded to in Arthurian romances, and more generally in all Irish or Breton mythological traditions . The “White Cities” of oral tradition, all of them referring to old dilapidated fortresses, are obvious memories of this otherwise thoroughly forgotten technique 22 .

If Jean MARKALE is right, if all the “glass cities”, “glass castles” and such like in Celtic legends mean ancient “vitrified forts”, then research prospects become vast indeed !

Let us begin with the farthest western point of the Celtic area where vestiges of vitrifications have been located : the Isle of Tory , lying off the western coast of Ireland. This island owes its name to a tower, one of the numerous “round towers” still standing in hundreds on the Green Island (and also in smaller numbers in Scotland), and whose real use has not been determined so far. They are generally considered as dating from the earliest times AD, and referred to the creation of the first monasteries. They are supposed to have been used as watch-towers, or as safes for monastery treasures threatened by Viking raids. How, then, have they survived, when the surrounding monasteries were almost completely destroyed ? The outstanding quality of their architecture, made up of wonderfully bonded stones, could partly explain the mystery. Indeed, one of those towers so much bore upon the history of that small island that the latter was called after its name (Tor-Inis meaning “Tower Island”). Here is what I read on the subject :

The tower on Toriniz island – to-day Tory Island – so old and venerable, no longer stands as a building. However, it still survives, or at least survived during the last century, as a ruin. Archaeologists were much amazed to find that those remains were vitrified. What can have caused this vitrification ?23

Now, according to Irish mythological tradition, the Fomoire (or Fomore ) had precisely settled their quarters in that island, and from there launched their raids on Ireland. They were a mythical people of evil giants; they were both the allies of the Tuatha de Danaann (“the goddess Dana or Ana’s children”) , through complex family ties, and their enemies. Neither of those two peoples were natives. They had both come from the World’s Northern Isles, a mythical place in the far North, of which practically nothing is known. Both were endowed with huge powers which our ancestors regarded as magical, but are now familiar thanks to our technical knowledge. One of the Fomore chiefs, the giant Balor, grandfather of the “god” Lug, one of the Tuatha de Danaann lived on Toriniz. From there, he would send a powerfull beam, wich we could call nowadays a “flux of energy” across the channel between Toriniz and Ireland, to blast his enemies. Those descriptions we have of him recall a machine rather than a living creature : he is compared to a Cyclops whose only eye cast out a beam that turned his enemies into ashes. He is a frightening giant whose only eye blasts a whole army when he opens the seven eyelids protecting it 24 . That extraordinary eye had to be kept open thanks to metal hooks held up by several assistants. During one of the three battles at Mag Tured in Ireland between the Fomore and the Tuatha de Danaann, the god Lug managed to put Balor’s evil eye out of order with his own spear, which the texts call “Assal’s spear”, one of the four magical objects brought back from the World’s Northern Isles 25 . That weapon also possessed strange properties : it never missed the mark and had to be cooled down in a cauldron filled with human blood 26 .

In Les Celtes et la civilisation celtique, Jean MARKALE gives some more detail on this fabulous tool :

Lug possesses a magical spear similar to Apollo’s both deadly and healing arrows. It is called Gai Bolga 27 .It is the thunderbolt emblem. It originated from Assal, one of the World’s Northern Isles (an allusion to Hyperborea), where the Tuatha de Danaann came from. That spear possessed a venomous destroying power and, to reduce that power, its point had to be dipped into a cauldron filled with poison and “black fluid”, i.e. blood 28 . After it had been thrown – with the war-cry “ibar” (meaning “yew”) – and it had hit its target, which it never missed (“it is so worthy that it never misses the mark”), it spontaneously flew back to the god’s hand, thanks to another cry, “athibar” : “it comes back to the hand that threw it” 29.

What was that instrument ? What techniques did Balor and Lug use to wage war on each other ? Is it not strange that Toriniz Tower, the very dwelling place of the giant Balor, was vitrified ?

Since the first version of the present paper (10/12/98), I have known of RALSTON’s research on the oppida in Limousin 30 and, during the Summer of 1999, I went to the département of Creuse, to try and see several of the sites referred to in that book as bearing traces of vitrifications. I could observe that RALSTON’s work, one of the most complete inventories on the iron age fortresses in France, copiously deals with the phenomenon of vitrification. Unfortunately, on the spot, it is often difficult to reach the sites, because of overgrowing vegetation and obvious neglect. I also had to give up visiting a few oppida I had planned to see, because of inaccurate information ( as in the case of Muraux in the village of St. Georges de Nigremont). Other oppida, where years ago vitrifications had been reported, have been destroyed through wild urbanising and lack of control on the part of the authorities supposed to be in charge : it is already the case of Thauron, which remains wholly unprotected. During our visit, I even saw a fellow blithely demolishing a stone wall on his property which, I was informed, probably was one of the only remains of the oppidum. The other sites I visited, though not urbanised, are not better protected. This is true of the Puy de Gaudy, above Ste Feyre village, south of Guéret, the favourite walking spot of the MAIF old people’s home below, and even more of mountain-bike adepts, who will ride undisturbed over the oppidum walls. In this département, where archaeological and historical remains are numerous, the authorities seem to do nothing to protect the sites they are in charge of (on no oppidum have I seen any mention “archaeological site”, or any prohibition!) During those visits, which were rather disappointing for the above reasons, I have observed few unquestionable vitrifications. For instance one of the sites that was not known as the most characteristic, the oppidum of Châteauvieux at Pionnat, facing the Puy de Gaudy. Vitrifications there are remarkably genuine, but the site, besides being unmentioned, is situated in an inextricable wood. Yet I could observe the same phenomenon as at Péran : the stones were literally cemented together on their surface, some of them had melted and turned into real “lava”. On this site, the theory I read or heard about, that those vitrified stones are clinker resulting from the melting of metals in furnaces situated along the rampart, is not valid. For, at Pionnat, I understood how one can confuse certain crumbled sections of the rampart, forming a kind of vault cemented by stone vitrification, with the cul-de-four of a furnace. This confusion is forgivable in non-specialists who desperately look for an explanation that will satisfy common sense, but it is no longer so when found in the publications of professional archaeologists. One only needs to perceive that what was mistaken for the remains of furnaces possesses none of the necessary structures to make a furnace work (air intake, etc.). We actually have to deal with a “lava bubble”, such as can be found in natural volcanic structures, except that, in the present case, we have an artificially vitrified building and not a natural phenomenon.

Since this trip to Creuse, I have of course carried on investigations which allowed me to trace other genuine vitrified forts , especially two of them in the area of the city of Roanne (May 1999) and another in the Allier department (May 2000).

I found that the extension area of those structures, which I first believed to be limited to Scotland, was much broader, since it covers a great part of the so-called Celtic era, though previous to the coming of the Celts to western Europe, since it dates back to the iron age.

Acknowledgments:

Thank you for helping in this translation to Mmes BOURDEIX, BRIVET et CUVELIER

(1 )LES CELTES (1997) suggest the date of the final Tene, that started about 125-115 B.C. for the building of fortresses with the technique called murus gallicus (p.292)

(2 )RITCHIE (1988), p.61. This hypothesis to explain the vitrification of those works, seems appealing a priori . Indeed it is taken up by most writers on the subject without further critical investigation. We will see later how it should be appraised.

(3) GIFFORD (1992), p.22.

(4) MARKALE (1997), pp.133-34. The author, perhaps because he is no archaeologist, is one of the few who attempted to draw a sincere description, and above all a clear explanation of the phenomenon.

(5) NICOLARDOT (1991), p.7.

(6) Here is yet another absurd “hypothesis” to try and explain the mystery of vitrified hillforts : volcanism. Now, on the site of Péran, as on all those I could see, volcanism provides no explanation, for the geographical areas where they lie show no sign of volcanic activity.

(7) A classical assumption but, as I said, not confirmed to this day.

(8) KOUSNETSOV, IVANOV,VELETSKY (1989).

(9) F.AUDOUZE and O. BUSCHENSHUTZ (1989).

(10) Gordon CHILDE, for ex.

(11) Let us observe that this chapter is strangely headed “Glass Castles”, without any explanation in the following text. I would probably not have noticed this point had I not previously read Jean MARKALE.

(12) My emphasis.

(13) Writers do not comment on these assumptions, and one does not know what they think of them. As for me, I find it hard to believe that one can seriously contemplate such a “hypothesis”.

(14) But, to my mind, equally unlikely.

(15) My emphasis.

(16) Let me be allowed to remark that tha author himself is not as assertive (RALSTON, 1992). Moreover, it seems that no relations of cause and effect should be inferred from this observation, since if the oppida built according to the murus gallicus technique often show traces of fire, we have seen that burning the wooden structure is not enough to vitrify the stone blocks.

(17) Which is also the case of the Camp de Péran, as I could verify.

(18) The Camp de Péran, especially, is neither named nor shown on the map figuring in the book.

(19) The same archaeologist was in charge of the Péran research.

(20) Particularly Georges NAUD, chairman of the Société archéologique d’Ardèche and curator of the Musée de la terre ardéchoise, Privas. Sample analyses are in progress at the Ecole des Mines d’Alès.

(21) MARKALE, op.cit.

(22) MARKALE, op. cit.

(23) KOARER-KALONDAN (1973), p.80.

(24) PERSIGOUT (1980), p. 40.

(25)PERSIGOUT (1980), p.184

(26)MARKALE (1997), p.354.

(27) In bolga one finds the root bolg meaning “lightning” and “sun fire”.

(28) Op.cit., p.395.

(29) MARKALE (1997) quoting R.A.S. MACALISTER, Leabhar na Gabala, poem 66.

(30) RALSTON, (1992).

Roland Comte granted a Master’degree in Political Science and is a Doctor of anthropology (EHESS, University of Paris-Sorbonne ). E-mail :

rc*****@wa*****.fr