Towns earthworks and camps



During the Viking period Scandinavian society gradually reached the stage of development at which towns were founded. There were two conflicting factors at work in the creation of the Viking towns: trade and piracy – the more flourishing the former, the more lucrative the latter. Accordingly towns were located well up narrow but navigable fjords and surrounded by defensive earthworks. There was always the danger that raiding enemies would suddenly attack, but the sea-going merchants were armed and knew how to use their weapons. Even to, things often went wrong; ships were plundered and towns burnt; but stronger than fear was the need for the trade on which their whole livelihood depended.

Along  the North Sea and the Baltic coasts, like blind eyes lie the vanished towns of the Vikings, the sites of Northern Europe’s oldest trading centres, following the winding route from the mouth of the Rhine along the coast of Jutland right up to Lake Malar in the north. If you think of the traders of those times – wherever they came or wherever they were bound – a picture of trading towns, once swarming with life, but now dead, springs to mind: the Frisian town of Dorestad; Hedeby  in the south of Denmark; farther north ,still in Denmark, Lindholm hoje, on the Lim Fjord; the Latvian Grobin; the Norse-Slav Wolin; the Estonin Truso; the Swedish Birka; and, in southern Norway, Skiringssal. What do we know of these vanished towns? What traces of them remain?

Grobin has already been mentioned, and we have also dealt with Dorestad and its final destruction by natural forces, but let us look at Hedeby.


It lay at the head  of the narrow but navigable fjord, the Slie, which cuts deep into south Slesvig from the Baltic. Hedeby means ‘the town at the heaths’. On three sides – north, west, and south – it was defended by a great semi-circular rampart, but to the east it was wide open to the waters of the cove of Haddeby Nor. The grass-grown remains of the defensive earthworks are still impressive, even though they are the less formidable for the decay of their former timber revetments. Within these defences the town covered some sixty acres/24.2ha, now fertile fields, and in circumference was the largest town in the North during the Viking period. There were two gates in the wall, south and north, and probably a third in the south-west. From 1900 until the outbreak of the Second World War with some interruptions, German archaeologists dug in and around Hedeby.

Their excavations showed that the semi-circular rampart was not at the outset as high and broad as it now appears. It grew in successive stages from a relatively low and simple rampart with a stockade and ditch to a defensive complex of extraordinary dimensions. On the other hand its circumference has apparently remained constant. Either the rampart when originally built enclosed space greater than the actual area of the town, with an eye to its expansion, or else it was not built until the town had already reached a certain size. The latter seems more likely.

In Viking times a rivulet ran through the town, cutting through the west wall and flowing into Heddeby Nor; the mouth of this stream is the site of the oldest part of the town. This modest stream was about important to the Viking town because it supplied drinking water. It has also proved valuable to the archaeologists, because successive deposits of rubbish on its bed – recognizable layers lying one over another – have provided a physical index to the relative chronology of the various phases of the town’s occupation. The excavators were also able to observe the changing relationship between various of the town’s wooden buildings and the river. At times the houses, whose sites were clearly to be seen, were built close to the stream, the banks of which were supported by wooden piles; in other phases the houses were built well back from the river, separated from it by an open space. In this low-lying part of Hedeby the inhabitants of nearly every house had dug a well fitted with a wooden pipe (of excellent cooper’s work) which agrees with the Arab merchant AI-Tartushi’s observations of the town’s freshwater wells on his visit to it in the tenth century. The boggy ground at the edge of Heddeby Nor has preserved the lower foundations of the houses particularly well; and anyone who visited the site of the excavations where whole areas at a time were uncovered, must have left with the impression of a stubble-field in which the angel of death had carefully cut away with his scythe the upper nine-tenths of all human habitations. From these remains it has been established that the wooden houses included some that were stave-built (close vertical planking), others there were of framed construction with wattle-and-daub panels between the structural timbers, and others constructed like log huts. The houses had their gable-ends to the street, with barns and stables behind, and the hearths were in the centre of the floor. In the later, higher, western end of the town were found small wattle-and-daub buildings with ‘sunken’ floors dug into the ground, in which the fire-place occupied one corner.

Various crafts flourished in Hedeby. There are traces of iron-smelting, weaving, industries using bone and horn, bronze-casting, glass-making, minting of coins, and potteries Sherds of imported Rhenish earthenware among the finds belong to the early ninth century. Very few farming implements were found. Inside the semi-circular rampart were two burial places, both in the western part of the town. One contained wooden coffins (lying east-west) but very few grave-goods. The other one, farther to the south, consisted of large wooden burial  chambers, with a greater abundance of grave-goods, such as weapons and jewellery. A third, rather older, burial ground with cremation graves has recently been found close to the southern gate of the town. The animal bones found in many parts of the town were mainly food-remains, and they show that pigs and cows were the most plentiful and popular meat. Sheep and goats were also eaten, but very few horses and chickens, and no game at all. Dogs and cats had evidently been kept in the houses. Many varieties of plants and fruit were found; barley, wheat, hazelnuts, walnuts, apples, cherries, plums, sloes, elderberries, blackberries, raspberries, wild strawberries, and hops.

The cove, Haddeby Nor, provided a natural harbour for Hedeby. In 1953 divers found there a line of pallisading, 480 feet/146.3m long, which had apparently served as a harbour defence. South of this, about a hundred feet/30.4m north-east of the estuary, and at a depth of about 9 feet/2.7m the excavators came across the wreckage of a burnt clinker-built ship, made of oak and ash, about 48 feet/14.6m long and 9 feet/2.7m wide, a rather flat-bottomed coastal vessel, presumably of local construction; and in it were found the remains of a man whose face had been injured.

The question of the origin and age of Hedeby will be considered when we come to discuss its relationship with the adjacent earthwork, the famous Danevirke; but the town clearly owed its existence to the trading-route between the North-Sea and the Baltic. It was a mercantile center, sometimes in Danish, sometimes in German or Swedish possession. (1) The evidence for the final fate of Hedeby is provided not only by literary but also by archaeological sources. History testifies very plainly to the situation which arose in the middle of the eleventh century, when King Swein Estridsson of Denmark and King Harald Hardrada of Norway came to grips with each other. About 1050, while Swein was engaged in the south with the Germans Emperor, Harald seized the opportunity to fall upon Hedeby, plundering and burning it to the ground. As Swein, returning from the south, approached the place, Harald’s ships, loaded with loot, made of, Swein pursued him and caught up with him at Laso in the Kattegat, where Harald, to lighten his vessels and escape, was forced to throw his rich Hedeby plunder overboard – so that is floated on the wind-swept Jutland sea, as the skald Thorleik the Fair says in his song. Another skald, a Norwegian who was with King Harald (quoted by Snorri), celebrated the fate of Hedeby in jubilant songs: ‘Burnt in anger was Hedeby from end to end. It was a doughty deed and one from which Swein will smart. High rose the flames from the houses when, last night before dawn, I stood on the stronghold’s arm.’ This ‘arm’ is doubtless the northern extremity of the great semi-circular rampart, where it runs into Haddeby nor, and from where even today, one can command the best view of the whole settlement.

During the excavations in the area of Hedeby near the harbour, the top layer in many places proved to consist of earth thick with charcoal and ash – evidence of the catastrophe by burning which brought the town’s existence to an end. On top of this layer there was no trace of refuse, and no other indication of human activity or survival. In the upper most deposits in the river-bed were discovered the burnt remains of two men and a horse: these, like the ship in the harbour, seem to be relics of the great fire of 1050 which destroyed the town of Hedeby for ever – or at least almost for ever. The earthworks of Hedeby will presumably last as long as the earthworks of the Danevirke and it is noteworthy that when, in the 1160s, King Valdemar reinforced the Danevirke with its strong red-brick wall, a hundred years had elapsed since the razing of Hedeby. The town must have had an attenuated life after 1050, for Adam of Bremen tells the story of its plundering and burning by the Wends as late as 1066.

In Scandinavian ecclesiastical history Hedeby is famous as the scene of the first Christian mission to Denmark, where the monk Ansgar was active between 826 and 829 as the forerunner of the full introduction of Christianity to Denmark, which was to occur 150 years later. According to the Arab trader who visited Hedeby in the middle of the tenth century, the place was not quite without Christians, but the religion did not really take root in Denmark until Harald Bluetooth too converted. The oldest Christian church in Denmark was undoubtedly Ansgar’s, in or near Hedeby, but all  efforts to locate it have failed. It was doubtless built of wood and, as we have seen, the lowest parts of timber buildings were preserved in the swampy soil near the harbour; but nothing has been found there that might have been Ansgar’s church.


Near Hedeby, north of the Slie Fjord, lies the modern town of Slesvig which cannot be proved by archaeological evidence to have existed before the eleventh century. The second element -vig is from the Latin vicus (‘town’) and was embodied, in the time of the Vikings – and earlier as well – in the names of many important trading centres in north-west Germany, on the Channel coast, and in England: among these are Brunswick, Wijk-bij-Duustede (Dorestad), Quentovic (Calais), Lundenwic (London), Eoferwic (York). Slesvig, then, signifies ‘the town at the Slie (fjord)’, and Hedeby ‘the town at  the heaths’; the latter corresponding with the name aet Hathum ‘at the heaths’, given to the place in Alfred the Great’s translation of Orisius (c. 900) by two narrators, the Norwegian Ottar and the Anglo-Saxon writer, the chronicler Athelweard, that the capital of the province of Angel ‘is in the Saxon language called Slesvig, but in Danish Hedeby’. Thus there is every reason to believe that at the head of the Slie Fjord in the Viking Age lay the town of Hedeby, and that it had two names, the Saxon Slesvig and the Danish Hedeby. After the destruction of Hedeby in 1050, most of the survivors moved to the north of the fjord and founded the present Slesvig, and the former name ‘Hedeby’ was gradually supplanted by ‘Slesvig’.


The next Viking town to be considered lay at the mouth of the Oder in the Baltic, on the southern fringes of the Viking ‘spheres of influence’, It is described by Adam of Bremen, in the 1070s as ‘the well-known town of Jumne, which affords  to the barbarians and Greeks [i.e. Greek Orthodox merchants] in those parts a much used anchorage’. Adam continues;

As so many almost unbelievable tales have been told in praise of this town, I think it is worth mentioning some matters of interest: it is, for example, the largest town in Europe, inhabited by Slave and such other people as Greeks and barbarians. Even  visiting Saxons are permitted to live on equal terms as long as they do not disclose they are Christians. But though the inhabitants are still bound by their heathen delusions, they are nevertheless more honourable and hospitable than any other people. There is an abundance of merchandise from the Nordic countries, and the towns is well provided with all good and precious things.

Other writers who mention the town at the mouth of the Oder call it not Jumne but Jumneta (or, distorted: Vineta), Julin, or Wolin. Philologists have concluded that these are two names for the same place, comparable with Hedeby-Slesvig; on this basis Jummne or Jumneta is the Norse name for it, Julin or Wolin the Slav name – a conclusion all the more acceptable since the town probably had a mixed Slav-Scandinavian population. Archaeologists, on their side have tried eagerly to locate the site or sites concerned. The Oder has three outletes: Peene in the west, Swine in the middle, and Dievenow in the east. On which of these three estuaries lay the town which Adam of Bremen mentioned so favourably? Answers to this question differ, but extensive excavations in and around the little modern town of Wolin on the Dievenow suggests that it may well have been there – let us then call the Viking town Wolin.

The most significant finds were located in the very middle of the town square at Wolin; thick layers of debris of which the lower date back to later Viking times. The remains of buildings and pottery have been dated with the pretty fair accuracy; the various forms of house construction – stave, half-timber, log huts – represented on the site include Viking as well as post-Viking building (the latter from around 1200). There is a similar mixture of styles and techniques in the pottery, some of it Scandinavian and some Slav. The conclusion, then, is that here, where the present Wolin stands, on the eastern outlet of the river, the Vikings began to settle about the year 1000 and gradually became assimilated into the existing populations. There were doubtless good trading opportunities at such a place, though Adam of Bremen eulogies were probably a little exaggerated.

Another problem which has long occupied the interest of historians is whether this Wolin, or Scandinavian Jumne, is identical with the mythical  ‘Jamsborg’ mentioned in Danish historical writings of the twelfth century and in Icelandic sagas of the thirteenth century (Sven Aggesen, Saxo, the Saga of the Jumsvikings, and the Knytlinga Saga).  The Danish tradition is that when Harald Bluetooth was banished from Denmark by his son Swein, he took refuge in the land of the Wends taught them the practice of piracy, and established a base for these operations at Jomsborg. Another version is that while at the height of his power in Denmark he founded Jomsborg and gave it a Wendish garrison. placing Danish chieftains in charge. The Icelandic sagas are influenced by romantic ideas. According to them Jomsborg was a purely Scandinavian military base, an independent Viking colony, an ideal warrior community run on Spartan principles, where no women were permitted, and where great warriors and heroes were reared; figures such as Palnatoki, Earl Sigvald, Bui the Stout, Vagn Akason, Styrbjorn. Here was the famous citadel of the Joms-vikings, governed with harsh discipline; from its well-designed artificial harbour, capable of accommodating 300 longships, the Joms-Vikings set out on those historic campaigns, which included Hjorungavag in Norway, Svold in Wendland or Denmark, and Fyrisvold in Sweden. The splendour of their feats resounds through the ages unaffected by the result of their battles – for the Joms-Vikings seem to have been distinguished for suffering glorious defeat!

Behind the highly-coloured pictures the sagas present there is certainly the historical fact that, throughout the Viking period, from the time of King Godfred at the beginning of the ninth century to that of Jomsborg’s destroyer, Magnus the Good, just before the middle of the eleventh, the Danes had vital interests along Germany’s extensive Baltic coasts, where the Slavs – Obotrites and Wends – lived. There is a historical nemesis in the fact that Denmark, when its lust for plundering had ceased, itself became the victim of raids by the Wends who, in the twelfth century, showed themselves so adept in the slowly acquired art of piracy that it was not until the 1160s that the Danes and Saxons managed to check them.

The  question whether Wolin was in fact Jomsborg is not at present settled. The solution of the problem would be furthered if the town or its neighbourhood could produce archaeological evidence of a Viking fort corresponding in some degree to the account of the sagas, but no such evidence seems to be forthcoming. If the Viking relics discovered in Wolin had come to light outside the area where Jomsborg was sought for, no one would have dreamt of connecting them with that stronghold.


Our next vanished Viking town of Truso. It is first mentioned by the traveller Wulfstan in the foreword to Alfred the Great’s version of Orosius’s world history. Wulfstan’s account was re-discovered in early modern times – 1589 – by the English geographer Richard Hakluyt, who was the first to realize its significance; and since then it has been much studied by both historians and geographers. As mentioned earlier, Wulfstan describes his journey from Hedeby to Truso; seven days and nights, he says, their vessel was constantly under sail; on their right was Wendland, on their left Langeland – then Bornholm, a sovereign state with its own king, then Blekinge, More, Oland, and Gotland – all Swedish. Wendland lay all the time to starboard, that is, as far as the mouth of the Vistula. Then follows a somewhat complicated account of Truso’s position in the Vistula delta. In their endeavours to locate the place scholars have come to realize how little they could depend on Wulfstan’s vague description. Only the discovery of traces of a Viking colony could give them something to go on, and in this connexion German scholars have advanced the claims of the town of Elbing, at the head of the Gulf of Danzig. No actual town site has been uncovered there, but scattered deposits of Viking weapons have been found and, more important, a large and partly Scandinavian Viking Age cemetery has come to light near Elbing railway station. Elbing’s position fits Wulfstan’s remarks that the Vistula formed the frontier between Wendland in the west and Estland (Estonia) in the east; and the name Truso may be connected with the half-dried-out lake near Elbing, called Lake Drausen. If Truso were indeed here, it must have had a mixture of populations and excellent trading possibilities. The wide river Vistula led towards the south-east, deep into the continent, and its distant source was not far from the Dniester which flowed into the Black Sea and provided a route to Byzantium.

While we are in the Baltic it is appropriate to mention a commodity of the greatest importance in these regions many centuries before the Vikings – amber. In the Bronze Age the main source of supply was the west coast of Jutland, but by the Iron Age the situation was entirely different. It is significant than in early imperial times Pliny, when describing the quest of Roman traders for amber, does not mention Jutland or other North Sea coastal areas, but speaks only of the Baltic regions. The supplier of amber in Jutland had become exhausted by the time of the Vikings, but it was still plentiful in the Baltic. Truso must have been a very convenient mart for amber trading.


The last of the vanishing Viking towns, but certainly the most famous of them all, is Birka, the Swedish mercantile centre on the little island of Bjorko in Lake Malar in eastern Sweden. The island, like Hedeby, is well hidden from the open sea; to reach it one must first penetrate the Swedish archipelago, sail through the narrow strait where Stockholm now stands, and out into the wide expanse of eastern Lake Malar. And here, in the centre of the fairway, where the north-south and east-west sailing routes intersect, lies the island of Bjorko, lonely and secluded nowadays, but in Viking times seething with life and bustling with commerce. Birka lay on the north-western promontory of the island; and its fame as a market for furs nad other Scandinavian wares attracted foreigners from many lands: Frisians, Anglo-Saxons, Germans, Balts, Greeks, and Orientals. That this is not just exaggeration and hearsay is evidenced by the rich finds in its many graves, in which Arab silver, Byzantine silks and brocades, Rhenish glass, Frisian cloth, and Frankish weapons have been found. Superb Scandinavian merchandise, too, has come to light in these ancient graves: first and foremost remnants of cloaks made of bear, fox, marten, otter, and beaver fur; and further such valuable commodities as reindeer antlers, walrus teeth, amber, and honey. Our sources of information about Birka are partly literary and partly archaeological. The latter evidence consists not only of grave-finds, but also of the remains of the town itself which can be seen to the present day. Some description of the scene would be appropriate. A visitor approaching  the island of Bjorko from the south will first catch sight of a bare rock due south of the site of Birka, surmounted by a odern stone cross (of Irish design, oddly enough) erected in memory of St Ansgar’e mission to Birka in the ninth century. (2) On this rock was the fortress and place of refuge of the town, surrounded by a rampart of earth and stones 25 feet/7.6m to 50 feet/15.2m across, oval in plan and with three gates: one facing north, one south, and one east towards the town. Outside the northern gate there have been found relics of the garrison of the fortress. In a north-easterly direction from this gate lies an area called the ‘Black Earth’  and this is the actual site of the historic town of Birka. Excavations here have not yielded, as at Hedeby, the lower foundations of the houses, but only fragments of burnt clay formed into two different pattern-types, from which it is deduced that there were two kinds of houses in Birka, one built of wattle-and-daub, and one of timber caulked with clay. The soil on the site is mixed with charcoal, ashes, and organic materials; in short it is a black occupation-earth, and this dark thirty acres/12.1ha – or less than half that of Hedeby. The defensive rampart of Birka, 22-39 feet/6.7-11.8m wide and 6 feet/1.8m high, is considerably lower and weaker than that of Hedeby. Only the northern part of this curving earthwork, stretching from the fortress to the western harbour and protecting the town on its landward side survives, a length of about 1,500 feet/457.2m; it may have been reinforced by a row of square wooden towers. Birka possessed three harbours: an artificial one to the west, which has completely disappeared, and two natural harbours on the northern coast – kugghamm in the west (named after the Frisian type of vessel, the kagge), and Korshamm (‘cross harbour’) in the east. Still farther east, near Salvik, there seems to have been a flat-bottomed harbour for smaller boats. Salvik means ‘the place of sale and trade’, while Korshamn may be a corruption of an original ‘Kornhamn’ – ‘corn harbour’. One of the experts on Birka, the archaeologist Holgar Arbman, had drawn attention to the latter possibility, and also to the likelihood that the great markets of Birka were for preference held in the winter when the finest furs would be available. He cites in support of this the fact that many of those buried in the Birka graves had crampons on his feet, and that in the ‘Black Earth’ have been found many ice axes and skates made of cow and horse bones.

Though Birka is smaller than Hedeby it has a far larger and finer collection of graves: thousands of them which have provided the famous collection now in the museum at Stockholm. No other ancient Scandinavian town has yielded so many and such diverse objects and relics as this. to the east of the defensive wall, on gently rolling open country, there are 1,600 burial mounds, large and small, huddled close together under pine and birch trees; this is Birka’s biggest cemetery. There are several others, however: between the ‘Black Earth’ and the hilltop fort, for instance, there are a number of mounds; and in the area south and south-east of the fort 400 scattered barrows can still be counted. Farther east across the island are other isolated groups of graves. Some of the richer graves at Birka, large timber burial chambers, have been found not under mounds but beneath hollows produced by subsidence, the earth having sunk as the burial chamber rotted and collapsed. Two such burial places have been found inside the defensive wall of the town, one close to the north gate of the fort, the other south of Kugghamn. how long did Birka, the most thriving of all Swedish Viking towns, exist? not as long as Hedeby, which perished in the catastrophe of 1050 already mentioned. Judging from the remains, Birka’s decline began before 1000. it is not that Birka, like Hedeby, was destroyed by fire, leaving a thick layer of ash to mark the town’s end; it is simply that there are virtually no finds from later than the end of the tenth century. There are no examples from Birka of the coins of the English  king Athelred II, common elsewhere in Sweden, in which the last substantial amounts of danegeld were paid; and indeed, no coins minted after the middle of the century have been found in the graves. The latest datable find is a hoard of silver, discovered in the ‘Black Earth’, in which no coins is later than 963 or 967. it has been assumed, though there is no certainty, that Birka was destroyed by the Danish forces which set out to conquer Sweden at the end of the tenth century but were defeated by King Eric the Victorious at the great battle of Fyrisvold described on the rune-stnes. Whether this was so or not, Birka drops out of history about 975; its function as the centre of the Baltic trade was taken over by Gotland, and the more local trade of Lake Malar was divided among several other places on the lake: principally Sigtuna, halfway between Birka and Uppsala.


A few words about Sigtuna. The town was favourably situated on the south coast of a peninsula in the broad part of Lake Malar. Excavations there have revealed the remains of houses, and have yielded material which, in the main, is later than 1000. It is a ‘long town’ stretching about 600 yds/548.6m along the border of the lake, and about 182.8m inland. It was Christianized early – perhaps right from its foundation – and had several churches. The earliest coins certainly minted in Sweden were made at Sigtuna: Olaf Skotkonung’s coins, copies from Anglo-Saxon designs, and certainly made by an Anglo-Saxon moneyer. They bear the inscription SIDEI, which stands for (TUNE) DEI or ‘GOD”S Sigtuna’ – possibly a piece of propaganda, directed against the neighbouring heathen in Old Uppsala. Olaf’s son, King Onund Jacob, also had coins minted at Sigtuna. It is possible that coins were struck at Birka too, in its time, but there is no certainty of this. Two Swedish rune-stones of the first part of the eleventh century tell of a guild of Frisian craftsmen in Sigtuna, and Icelandic skaldic poems declare that King Magnus the Good and King Harald Hardrada stopped at Sigtuna on their way home from Russia. Adam of Bremen (c. 1070) describes the place as a large town (civitas magna), and records that it was the see of Sweden’s first bishop, Adalward the Younger. It is likely, though not certain, that the earliest Sigtuna lay at Signhildsberg, about 2 1/2 miles/4km away from the present town, and that its layout followed an Anglo-Saxon plane. Excavations in Sigtuna have revealed, as at Birka, a thick-coloured occupation soil, a ‘Black Earth’ in which were found the remains of primitive houses with clay floors and wattle-and-daub walls, as well as larger log houses.


Apart from Sigtuna there are three existing Swedish towns which have origins back in the later Viking period; there are Skara in Vastergotland, Lund in Skane, and probably Talje, the modern Sodertalje, near Stockholm.

Denmark possess one more lost town of the Viking Age: an abandoned site with a semi-circular defensive wall, at Brovold on the Isle of Als, where Viking relics establish the age of the place. There are also, however, many modern towns which have their roots in Viking times, some of them also recorded as episcocal sees. In Jutland there are Slesvig, Ribe, Aarhus, Viborg; on Fyn, Odense; and on Zealand, Roskilde and Ringsted.

In Norway the modern towns of Oslo, Bergen, and Trondheim (Nidaros) can trace their origins back to the end of the Viking Age. There is also a lost town, the harbour called Skiringssal, mentioned by the trader, Ottar. Although its site has not been located, it is probable that it was somewhere near the so-called Kaupang farms in the south of Vestfold, where many Vikings’ graves, including several ships, have been found containing imported objects of the ninth century – such as Irish and Anglo-Saxon metalware, and bits of finely-woven ‘Frisian’ cloth. The Norwegian archaeologist Charlotte Blindheim, who has been in charge of the more recent excavations (which have produced several interesting discoveries, such as the remains of a wharf, bollards, fragments of pottery and Rhenish glassware, an Anglo-Saxon and a Frankish coin (undefended) trading centre in a thriving district, in contact with several other trading centres including Hedeby and Birka. By means of the C14 method (a radioactive carbon test) all these finds have been dated to around the ninth century.

In conclusion we must mention the recent excavations of a trading settlement whose activity seems in the main to have been pre-Viking, dating in fact from the Merovingian period. Among the discoveries on the little island of Helgo in Lake Malar are gold amules,  fragments of glass, and a small gilded bronze representation of a face. There is also the top of an enamelled Irish crozier, a bronze Indian statue of Buddha, and several relics from the Viking period. These excavations were conducted by Mr W. Holmqvist.

Several causes acting together – or one if it were important enough – could lead to the foundation of a town in Viking ties. The most common, no doubt, was trade, especially maritime trade, which would bring abour the development of towns at harbours and centres of transport like Birka and Hedeby. towns came into being in the interior too, at points where the great land routes intersected and where legal decisions on commercial matters were formulated: places such as Viborg and Ringsted. Religious centres again, whether pagan shrines or Christian churches and sees, would have the effect of bringing large numbers of people together as permanent residents; and so would a royal centre. However, the most potent factor in the development of the oldest towns was the existence of a local market where merchants could buy and sell, and where the craftsmen would find a ready and stable demand for their products.





In the Frankish annals – Annales Regni Francorum – there is  described under the year 808 the action taken by the Danish King Godfred after he had sacked and destroyed the north German Baltic town of Reric. Godfred, says the chronicle,

carried the merchants off with him and sailed with his whole arm to the harbour called Sliesthorp. Here he stayed for some days and ordered that his country’s frontier with the land of the Saxons should be fortified by a rampart, stretching from the eastern bay called Ostersalt to the western ocean, protecting the entire northern shore of the River Eider and having but a single gate for carriage and horses to travel to and fro. After distributing this task among his chieftains he returned home.

This great earthwork across the neck of the Jutland peninsula from the head of the Slie Fjord right across to the west, was subsequently to be known (and still is) as the Danevirke, and a good part of it is still visible.

This protective system of earthworks, securing Denmark to the south, is the most extensive monument of the past to be found in the Nordic countries. As the sketch-map on shows, the Danevirke forms a barrier across the narrowest part of south Slesvig, from the head of the Slie Fjord in the east to the meadows by the Trene and Rheide rivers of the west. In addition to this there was the Ostervold (‘Eastern Rampart’) across the Svansen peninsular.

The most important part of the main wall (A-F on the map) is the portion which stretches with some interruptions for nearly nine miles/14.4km from the Gottorp area in the east to that of Hollingsted. One gets the impression that, apart from the isolated earthworks called Tyraborg (c) the main wall was constructed as a complete unit; and it may well be identified, therefore, with the wall referred to in the Frankish annals as reaching from sea to sea. It can be objected to this theory that the main wall does not, in fact,  reach from sea to sea. The force of the objection is reduced if we add to the main wall the (now vanished) Ostervold. Further, one must not expect such an account as that given on the Frankish chronicles to correspond exactly with the actual monument. Moreover, during spring and summer at least, a wall linking meadow to meadow and swamp to swamp would have the effect of an unbroken barrier from sea to sea. Another contradiction must also be noticed: the expression in the annals about the wall protecting ‘the entire northern shore of the River Eider’ does not strictly apply to the main wall. On the other hand, however, the gate for traffic is doubtless the so-called ‘Kalegat” (D) where the ancient route (known as the ‘Army Road’ or the ‘Ox Road’) from south to north through Holstein, Slesvig, and Jutland intersects the main wall. From C to D the main wall covers a length of some 1,850 feet/563.8m, and is about 18 feet/5.4m high and 28.9m wide. The middle part of the main wall (D- E) continues for over 2 3/4 mile/4.4km in a south-westerly direction, to Kurbotg (E). This part of the wall is of very considerable dimensions, and, according to German archaeologists, was built during several different periods. It was strengthened by palisades and a shallow moat towards the south-east and its rear sloped down to an inner road well protected by the wall along its whole length. At some period this middle esction of the main wall (D-E) was reinforced by a stone parapet 9 feet/2.7m broad and 9 feet/2.7m high: the celebrated ‘graniteboulder wall’ with a facing of stones set in a zigzag pattern and bound with clay. In front of this parapet there was a U-shpaed ditch. A further and final reinforcement of the middle section was supplied by a brick wall 6 feet/1.8m thick and 19-22 feet/5.7-6.7m high, which was equipped with buttresses and probably with battlements and gallery as well. This addition, made in the 1060s by King Valdemar the Great, is still preserved in parts, and it is the only part of the Danevirke which can safely be dated from literary sources- i.e., by the writings of Sven Aggesen and Saxo and by the inscription on a lead tablets in Ringsted church in Zealand. in the final 1100 feet/335.2m before the bend near Kurborg (E) the middle section again becomes comparatively low – 6-9 feet/1.8-2.7m high – and no more than 65 feet/19.8m wide. The western section of the main wall (E-F), which traverses the Hollingsted meadows, has the same dimensions. Examination by archaeologists of sections of this part of the main wall led to the conclusion that it was constructed at up to four  different periods, and that her, too, there was a road along the inner side of the wall. This part of the main wall is called the ‘Crooked Rampart’, and here just south of Ellingsted, Soren Telling discovered, in 1959, an 800 yds/731.5m long double rampart enclosing a water pit.

Between the main wall of the Danevirke and the semi-circular defensive rampart of Hedeby runs a connecting earthwork (B-H), almost two miles/3.2km long. It is about 15 feet/4.5m high for most of the way, and its western end consists of two parallel walls. It traverses the low ground south of Bustrup and joins the semi-circular wall at the place where the tivulet cuts in. German archeaologists are of the opinion that this connecting wall was built at three different periods, and that from the very first it was protected on the south side of a moat.

Hedeby’s defensive semi-circle of earthworks (G-I) is over three-quarters of a mile/1.2km in length, while its height varies between 12 and 25 feet/3.6-7.6 m. Near its northern gate, according to German experts, there is evidence of no fewer than nine periods of construction, but this figure is perhaps an overestimate. Outside its south-western stretch there was an advanced earthwork as well, of which only slight traces remain. North of Hedebt lies the ‘Hojborg’, a steep hill with a flat top on which there are a number of shallow burial mounds surrounded by a low wall; this whole system, however, seems to date from before the Viking period. About 1 1/4 miles/2km south and south-west of Hedeby there used to run, from the head of the Selk cove to the meadows by the river Rhiede, a completely straight wall (N-O) about 3 3.4 miles/6km in length and 6 feet/1.8m high, with a V-shaped moat to the south; it is called the ‘Kovirke’ but has now almost disappeared. Stretching across the peninsula of Svansen from the Slie Fjord to the Vindeby cove, was the Ostervold, some two miles/3.2km long; but this too has almost wholly vanished.

The Danevirke as a whole is an impressive feat of engineering, covering 350 years of Danish history, from its beginning in Viking times to the addition of King Valdemar’s brick wall in the 1160s – a most fitting memorial of days gone by. Long stretches of these broad earthworks still remains a testimony to Denmark’s ancient history and in the spring of 1945 – several sections of the Danevirke were badly damaged during German preparations against an anticipated attack by British forces. Admittedly most of this damage has been made good, but it will never be possible to repair the demolition by the Germans of large parts of the Kovirke (N-O) to make way for their airfields.

The scientific investigation and dating of the different parts of the Danevirke has been going on now for almost a century. It has led to much discussion, but the various points of view cannot be enlarged upon here. Instead we shall give only a summary of the probable course of development.

King Godfred’s wall, mentioned in the Frankish annals of 808, can apparently be identified with the main wall (A-F) of the Danevirke. The location of the harbour of Sliesthorp which, according to the annals, was used by Godfred, is uncertain, but as it must have been protected by his new wall near the present Gottorp or Slesvig. Shortly afterwards, in the ninth century, there grew up the trading centre of Hedeby-Slesvig at the cove Haddeby Nor. Around the year 900 this place fell into Swedish hands and was protected by a semi-circular earthwork later connected with the main wall of the Danevirke. (this, of course, presupposes that this section of Danevirke was also in Swedish hands). After the collapse of Swedish power the main wall was reinforced by the ‘granite wall’ and at the end of the tenth century the Kovirke was built. In 1050 Hedeby was destroyed and most of the surviving inhabitants moved to Slesvig on the north side of the Slie Fjord; soon afterwards the name of Hedeby was lost. This outline of development is full of uncertainty. In particular the problem of Sliesthorp presents difficulties. It should be kept in mind that the absences of archaeological evidence from before the eleventh century in modern Slesvig is hardly conclusive evidence for doubting the existence of the town in Viking times; yet the absence is undeniably striking.


Elsewhere in the three Viking countries there are no other extensive system of defence comparable to the Danevirke. One would not expect it to be otherwise, for defences of this elaborate kind, built of earth stone, and timber, were only required at the place where Viking territory met that of considerable and hostile foreign powers. The Danevirke was to protect the north’s only port of entry against Frankish and Saxon powers coming from the south. The construction of similar defences in the other northern countries was quite superfluous, not because the Vikings were at peace with each other – far from it – but because militarily they were maritime powers and for preference campaigned by sea. There are, then, no other Danevirkes in Scandinavia. There are, of course, many smaller earthworks scattered everywhere, but these are generally attributed to tribal conflicts in pre-viking times.

It might be expected that in the Viking colonies in England and France earthworks would be used as protection against attacks by the natives, but there is no certain example for us to point to. It is very difficult to fix the date of any earthwork unless datable objects are found within it or literary evidence certainly referring to it exists. There are in England many ‘burhs’ and ‘camps’ and ‘dykes’ associated philologically with Danes but so far we have no positive knowledge of their origin – though perhaps future research will bring fresh evidence to light. This is more or less the case in northern France too; but attention must be drawn in this connexion to the Dano-Swedish excavations in 1951-2 of the remarkable earthwork which cuts off the northern end of the Cotentin Peninsula in Normandy and is apparently directed against attack from the south. In its east-west course to sea it takes in two small harbours, and was apparently built by sea warriors to guard against their southern foes. It is named ‘Hague Dike’ – that is the rampart (dik) which cuts off the tongue of land (hagi) – a pure Old Norse place-name. This fortification may very well have been constructed by the Vikings in the ninth or tenth century. The excavations have proved its military character: on top of the earthwork, for example, were found extensive traces of fire, indicating that long trenches were filled with a fiercely burning fire as a means of defence. No datable objects have yet been discovered, and Hague Dike so far remains undated. Perhaps the C14 method will bring results. (C14 tests, recently carried out on some Hague Dike charcoal material, point to the Hallstatt period, 800-900 B. C. Yet the place-names show a Viking settlement. it is probable, therefore, that this very old rampart was re-used in Viking times)



In a special category must be placed the one kind of Viking monument so far found only in Denmark; the military fortress. Of these, four have so far come to light on Danish soil – one in Zealand, one in Fyn, and two in Jutland. The earliest to be found, and the only one fully excavated (between 1934 and 1941), is Trelleborg on Zealand.


Trelleborg was excavated by Dr Paul Norlund, who in 1948 published a detailed account of his work. it is situated in western Zealand, two and half miles/4km west of the town of Slaglse, between this town and sea (the Store Baelt). The actual site is on a headland formed by two rivers as they join to run into the Store Baelt north of Korsor. Here, in the later part of the Viking period, the fortress was built apparently on top of an old heathen place of sacrifice comprising sacrificial pits and remains of houses. The work seems to have been very thoroughly carried out: the site was levelled and filled in, and all was built according to carefully made plans by experienced engineers working with mathematical precision. The fortress consisted of two sections: a main and an outer enclosure.

The main part of the fortress is circular in plan, surrounded by a strong rampart which still exists. On it landward side – south and east – it is further protected by a wide deep, U-shaped ditch. In the rampart there were four gates at the four main compass points. These were linked by two streets, constructed with wood and crossing at right angles at the centre of the site, thus dividing the circular ares into four quadrants. In each of these there stood four houses of equal size, arranged in squares, making sixteen in all. In plan, these buildings – with curving long-walls and straight ends – resemble ships with their bows cut off. Each of the houses is divided into three sections, the centre one being the largest (58 feet/17.6. long). Although the timber from which all these buildings were made has vanished, the plan of the houses can be traced from the remaining post-holes in the ground. As a rule the houses had doors in both gable walls, with corresponding doors between the three inner rooms. The large centre room also had side doors diagonally placed in each of the long sides of the house. The end rooms sometimes had cellar-like pits underneath them, perhaps for stores, rubbish, or even prisoners. The floor of the centre room was of planks or clay, with a fireplace in the centre-flanked, it is assumed, by wide settles on which the Vikings used to sit or lie. In the roof there was, presumably, a ventilation hole (Old Norse ljori). Besides the sixteen houses there were a number of small detached ones: guard houses at two of the gates, officers’ houses in the centre of two of the squares, and a boat-shaped house of smaller dimensions to the north of the north-eastern square. there appears also to have been a street or footpath running inside the whole length of the circular rampart. This ring-wall round the main camp was strengthened and retained on both sides by palisades, and reinforced internally by lacing with transverse and longitudinal timbers. Its four gates, defended by palisades covering heavy stone packing, probably had wooden roofs which would have given the effect of tunnels. Two folding doors barred entry from the outside: iron rings and massive keys were found near the gates. The outer surface of the ring-wall facing the land was faced with thick clay held together by stout sticks and branches; and where it lay towards the swampy areas to the north, west, and south-west, the wall rested on a foundation of stones and piling, and was well protected higher up by vertical palisades.

The outer defence-work served as a reinforcement on the landward side: its southernmost section is curved, concentrically with the shape of the main fortress, but to the north the enclosed area is rectangular, widening to the east; throughout, the defensive earthowrk consists of a low rampart and a shallow ditch. Within the curved section were set radially thirteen long-houses of the same elliptical plan as the sixteen in the main fortress, though of smaller size. Near the main eastern gate (in the rectangular part of the outer work) were two similar houses set parallel to each other; and farther east in the same square was the stronghold’s cemetery (comprising about 150 graves), probably a continuation of the ancient burial ground which had belonged to the sacrificial place existing before Trelleborg was built. Dr Norlund surmises convincingly from the two parallel houses that the engineers originally intended to build a set of house-squares in the outer camp, but that, in view of its curved shape, they altered their plans and built instead the radially-set series of houses which has been revealed. The principal approach both to main ad outer camps was from the south. We are justified in calling the men who built Trelleborg engineers, so mathematically precise was their method of planning. Their standard unit of measurement was a modification of the Roman foot; that is, 29.33 cm/11 1/2 inches. compared with the Roman 29.57 cm/11.64 inches. On this scale the houses of the squares are 100 feet/30m, the houses in the outer work 90 feet/27.4m long, the circular rampart 60 feet/18.2m wide. The smaller houses in the middle of the two of the squares are 30 feet/ by 15 feet/9.1m by 4.5m . The radius from the centre of the main enclosed area to the inner side of its circular rampart is 234 feet/71.3m, the distance between the ditches of the two enclosures is for some distance also 234 feet/71.3m, and the distance between the centre of the main enclosure and the  nearer gables of the houses in the outer work is exactly double this: 468 feet/142.6m. The engineers began by marking the centre of the fortress and from it striking  the circles of the ramparts and ditches. The same central datum is also the point of intersection of the two transverse axes, cutting each other at right angles, which divide the area of the main fortress into its four quadrants, linking the gates and continuing outwards through them. The constructors’ exactitude is everywhere apparent. Dr Norlund points out that the curved sides of the long-houses are always symmetrical about central axes and that their construction was based on ellipses. Two houses at right-angles to each other had the same focus, and to construct the four ellipses of a house-square the engineer only had to fix four focuses constituting the corners of a square whose side was about 36.45cm/14.3 inches. or 124 Roman feet/37m. The construction of the whole house system within the circular earthwork of the main fortress is consequently based upon squares.

Trelleborg is splendidly situated in wide meadows backing onto the higher land. The holes in the ground which, in ancient times timbers, have been clearly marked by the excavators with cement; ramparts and ditches have been cleared and partly reconstructed, so that the visitor nowadays gets a graphic impression of what a Viking base looked like, particularly if he has enough imagination to picture the former appearance of the square groups of elliptical houses.

The Trelleborg cemetery reveals, as we would expect in a military garrison, that most of those buried there were youngish men – between twenty and forty years old. There were also a number of women, but few children or old people. Grave goods were very scarce, but there was no specific evidence of Christian practices.

Inhumation-graves orientated east-west were not unknown in Denmark before the advent of Christianity. Three mass graves were found at Trelleborg, the largest containing ten bodies. The objects from the graves included few weapons, although one notable find was a silver-inlaid battle-axe with a very broad edge but narrow blade. There was a quantity of tools such as scythe-blades, and coulters, ornaments, earthware, and objects used in spinning and weaving. From these discoveries it is possible to give an approximate date to Trelleborg, within the period from the later part of the tenth century to the first part of the eleventh, say 975-1050. The camp did not have a very long life, therefore, but it evidently existed at the time when Swein Forkbeard conquered England and when Cnut the Great fought the ‘battle of the three Kings’ at the River Helge in 1026.

It stands to reason that Trelleborg was primarily a military and naval bases: it had theclassic location for this purpose, providing easy access to the sea and yet protected against attacks and sheltered from storms. it is presumed that vessels were towed up the river to the camp, and that each of the boat-shaped houses afforded quarters for a ship’s crew. The contents of graves (coulters, scythe-blades, etc) suggest that the garrison were able to keep themselves supplied with certain commodities, so that foraging raids upon adjacent farms need not have been too frequent. The building of the Trelleborg houses must have made heavy demands upon the near-by forces.

It has been suggested by some scholars that each house at Trelleborg has as its roof a ship turned down, and that in this way the camp provided winter quarters for the fleet, but this thesis is scarcely tenable. Another supposition-acceptable as long as Trelleborg was the only such camp known – was that here, at long here, at long last, was the elusive and perhaps fabulous Jomsborg, for the ground plan did indeed conform to the legends about Jomsborg. however, this theory was completely shattered when four Trelleborg-type camps were unearthed in Denmark alone, and was invalidated also by the fact that (as the graves proved) there had been women in the camp, whereas women were prohibited from entering Jomsborg. One thing is certain: Trelleborg betokens a powerful organization; only a king would have had the means and power to carry out such a large project. That Trelleborg (and its equals in Denmark) should have been built by enemy invaders is very unlikely; it is not supported by the archaeological evidence, and still less by the historical records, which show that Danish power was at its height during this period. Dr  Norlund contends, with great justification, that Trelleborg seems to point to Swein Forkbeard’s and Cnut the Great’s powerful military force, their hirth or housecarls. Such conquering monarchs as these would be likely to use effective garrisons to maintain peace in the homeland. Trelleborg must have provided accommodation for about 1,200 men. How and when it ceased to be a base is not known: there are certainly no signs of destruction by fire.


The next military fortress of the Trelleborg type is at Aggersborg in north Jutland, and it was partly excavated by C.G. Schultz between 1945 and 1949. It is situated almost in the centre of Denmark’s largest fjord system, the Lim Fjord, a long fairway which cuts across north Jutland from sea to sea and was as in early Viking times the starting-point of many raids on England. The location of this fortress is in its own way as significant as that of Trelleborg. Lying between the Aggersborg estate on the coast and the church and churchyard on top of the hill, it occupies sloping ground commanding an extensive view of the fjord. Nothing of the old camp is visible on the ground, which is now just ordinary cultivated land, but anyone who knows what he is looking for can just about trace outlines of the circular wall which surrounded the camp. It was similar in layout to Trelleborg, about the same age, but much bigger. A conjecture can be made about the date of its destruction. In 1086 the rebellious peasants of north Jutland rose against the King, St Cnut, and sacked his stronghold. (The revolt ended with the killing of the king in St. Albans church at Odense in the same year). The monk Alnoth gives a vivid description of the destruction of the king’s stronghold – could this have been Aggersborg?

The site has not been completely excavated, but enough had been done to establish the shape and size of the fortress. There was no outer annexe as at Trelleborg. It consisted simply of one large circular as at Trelleborg. It consisted simply of one large circular area; but within its rampart there were no fewer than forty-eight houses in twelve squares- compared with sixteen houses in four squares at Trelleborg. Moreover, the houses are 110 Roman feet/33.5cm approx long – that is one-tenth longer than those found at Trelleborg. The two axial roads across the camp had on each side a shorter upper and lower road. The inside radius of the Aggersborg camp was 407 Roman feet/124m approx (cf. Trelleborg’s 234), its extreme diameter no less than 960 feet/292.6m. It has a central square 72 feet/21.9m by 72 feet/21.9m, and there seem to have been wooden towers over its gates. The encircling rampart was timbered inside and out more strongly on the outside – and was presumably topped with a timber parapet and doubtless strengthened with transverse timbers. The houses at Aggersborg had walls with a less pronounced curve than those at Trelleborg, but their gable ends were much broader. The excavator reached the conclusion that the walls were not constructed of close-fitting upright planks, but of heavy upright structural members with a filling of wattle-and-daub between them, and that the gable-ends of the houses were built of horizontal planks.

A most interesting discovery was that the fortress had been built on the site of an earlier settlement; a village with much smaller boat-shaped houses built on the same general pattern as those of the later stronghold. The village also included many outbuildings laid out east-west, without fire-places. Large post-holes at the ends of these buildings must have held supports for a horizontal member, probably the roof-ridge. scattered finds on the site indicate that this village must have been there a couple of centuries before the military camp was constructed in the middle of the seventh century; so that the boat-shaped type of house was evidently a much-used design throughout the Viking period – in Denmark certainly, and probably throughout the North. It is likely, indeed, that the design originated from the primitive shelter made by dragging a boat ashore and upending it across standing posts, either to repair or to store it. This explanation is more likely than the somewhat far-fetched theory that the early builders had made the technical discovery that a curved wall resists the wind better than a flat one! Later discoveries in Denmark have definitely confirmed that these ‘boat-shaped houses’ were known before the Viking period.


The third Viking camp of the Trelleborg type was found in north-east Jutland at the head of the narrow Mariager-Hobro Fjord a few miles/km south-west of the modern town of Hobro. It is called Fyrkat, a comparatively modern name which has been explained by the excavator of the camp, C. G. Schultz, (An architect as well as an archaeologist, and fully familar with Trelleborg and Aggersborg), as follows: the word kat (English ‘cat’) is a seventeenth century term used in the art of fortification for a work built on top of a bastion or rampart, and when, during the Renaissance, there was nothing visible of this old fortress except the ruins of four walls between four collapsed gateways, it was known as de fire kattle – the four kats. Fyrkat, not yet fully excavated, closely resembles Trelleborg, although it has no outer annexe. It stands on a low promontory, and it is fair to suppose that vessels could have been towed up the little river Onsild from the fjord. The site, once more, is militarily correct – providing free access to the sea and natural protection. The encircling rampart was about 10 feet/3m high: a solid wooden affair filled with earth, reinforced internally with timbers, and doubtless crowned with a parapet. Its four gates, like those at Trelleborg, elaborate timber-lined tunnels; and within the circle of the camp were sixteen houses arranged in four squares. The boat-shaped houses were rather shorter (96 feet/29m) than those at Trelleborg; they had double clay-built walls, and were equipped with tambours. The axial streets of Fyrkat were of wooden construction, and so was the road round the inside of the circular wall.

In the post-holes of one of the gateways the remains of the heavy oak timbers which supported the gate still survived. Various objects have been found on the site – fragments of a gold filigree brooch, a pair of silver bracelets, a silver finger-ring, an iron axe, a couple of whetstones, some vessels made of soap-stone – and from these finds the cap has been dated, like Trelleborg, to about the end of the tenth century. Fyrkat appears to have been destroyed by fire during a south-westerly storm, but it is not know when. It may well be that it was sacked and burnt on 1086 during the rising of peasants which brought about the fall of Aggersborg; but we do not know. Quite close to and north-east of the camp lay its cemetery.


Finally we come to the fourth and, for the time being at any rate the last of these Viking camps in Denmark. This one was situated right in the middle of the present town of Odense, the capital of Fyn, on a hill called Nonnebjerg. The existence of this camp was demonstrated partly by the discovery by the artist Ernst Hansen of an map of Odense on which it was mentioned and parlty by trial excavations on the spot – sufficient being uncovered to confirm the existence here of a fortress of the type now familiar. Further discoveries of various silver hoards of the Viking Age in the same vicinity have made it possible to date the Nonnebjerg fortification to the end of the tenth century, contemporary with Trelleborg and Fyrkat. it is a reasonable conjecture that this fortress, which lay quite close to the church where King Cnut met his death in 1086, was destroyed on that occasion. Nonnebjerg, like the other three military sites, was located in the usual way at the head of a long fjord, in this case that of Odense.

Perhaps the future will bring to light more of these circular Viking strongholds in Denmark, and so prove that Swein Forkbeard and Cnut the Great secured the loyalty of their home country by means of a network of bases manned by their housecarls and mercenaries. These camps are strategically placed for communication by land and water, and could serve an army no less than a fleet.



Where did the Vikings learn the science of constructing such fortress? The question is far from easy to answer. To some extent this mastery of the science of fortification might be said to have a native Scandinavian origin, as even before Viking times the northerners could build circular places of refuge of earth and stone, as, for example, the fifth-century fort at Ismantorp, on Oland, Sweden, with its nine gates and its houses radially placed on the inner area. Yet this primitive work affords no real comparison with the viking feats of engineering;

nor do the circular defensible works of a later date found in north-west Germany and in England help us in our inquiry. What we need for comparison are duplicates of Trelleborg in other countries, not merely more or less primitive planned forts. Even if such were found we should still have to determine whether they preceded the Danish examples or were imitations of the Danish prototype. The Anglo-Saxons, for instance, had no more particularly a native technical or traditional basis for constructing a Trelleborg than had the Danes, and a Trelleborg in England would appear just as foreign a phenomenon as are the Danish ones in Denmark. If we inquire which of the two peoples – the Anglo-Saxons or the Danes – would have been more likely, around the year 1000, to learn the science of engineering from a native source, the answer must be the engineering from a foreign source, the answer must be the Danes – because of the wider horizons given them by their Viking campaigns. Engineering as skilled as that revealed as Trelleborg was not native: it must have originated either in Roman tradition, via the western Roman Empire; or in Byzantium, whence it could have reached Denmark along the routes through Russia; or else in the Near East ,dervied perhaps, from the Arab Empire, and spread along the eastern ot the western ways (Russia or Spain), to Denmark.

If, however, , we assume that military engineering of this sort began in Rome, and penetrated north through the country of the Franks, we at once run into a problem: the Roman camps and fortresses were invariably rectangular, not circular; the Frankish forts too, were never circular. The unit of measurement, the Roman foot, is the only link between the Trelleborg and the Roman camps, but this has no significance considering the radically different building designs of the two types. That the basis of the engineering of Trelleborg derives from Byzantium seems much more likely, but before enlarging on this I must mention the opinion which sees resemblances between Trelleborg and the Araband Oriental (Persian-Sassanid) circular forts, towns, and holy places; resemblances and comparisons which are much too vague, and far-fetched.

The Byzantine world, as we know, was not very remote from the Scandinavian countries, especially Sweden. Without repeating what has been said earlier on this matter, I refer to the salient points of the connexion. At the end of the tenth century the Greek Emperor in Constantinople engaged Vikings as mercenaries in his household guard. The contract between Byzantium and the North was a lively one at this time; about half-way between the Greek and the Scandinavian worlds lay the important Norse town of Kiev. Some archaeologists have compared the huge earth, wood, and stone ramparts of Hedeby with the famous semi-circular wall which shielded Constantinople from attack by land. It is not unlikely that the highly developed mathematical skill revealed in the construction of the north Danish Trelleborg-type fortresses was brought to the north from Byzantium; perhaps by Byzantine engineers who had been taken prisoner and who applied the science they had practiced in stone to similar engineering projects in timber. However, this imported skill did not last long in the North. It could be employed only in the service of a powerful employer: that is to say, a strong king, for no common chieftain would be able to take the task of building himself these imposing military bases. for that reason alone it seems quite likely that no other Trelleborgs will be found in Denmark. The four we know of would amply suffice for the security of the kingdoms, though perhaps a further one in the south of Jutland would be warrented. The Danish writer Palle Lauring has ventured the suggestion that a fifth stronghold – now vanished – once existed at Egernborg near Eckenforde. There would also seem to have scope for one in Skane.

If a Trelleborg-type camp should ever be found in England it seems highly probable that it would owe its existence to Danish influence.

Though scholars are on the whole agreed that Trelleborg, Fyrkat, and Nonnebjerg were built in the late tenth century, there is some doubt about Aggersborg. The supposition most favoured is that it dates from about the middle of the eleventh century, the reigns of Kings Hardacnut (1035-42), Magnus the Good (1042-7), and Swein Estridsson (1047-76). it is relevant here to point out a clear Byzantine influence on Danish coins minted in this period, an influence which began with Hardacnut and was increased in the reigns of his two successors. Coins from Lund, for example, show Swein Estridsson in the style of a Byzantine Emperor: at full length, holding sceptre and orb. This phenomenon is generally explained by the fact that the Varangians (and not least Harald Hardrada) bought back a lot of Byzantine coins from their journeys, coins which were then copied by local moneyers. Should it prove correct that Aggersborg was built as late as the middle of the eleventh century, we can point to the fact that at this very period Byzantine influence was affecting another field of Danish design, that of coins.