52 (224)

98 The Viking Age in Denmark

wali is of about the same size as the one at Aggersborg, measuring 240 metres, while Fyrkat mcasures 120 metres at the innerside and Nonnebakken probably the same. Trelleborg’s diameter is 15 metres longer (and of the same size as Oost-Souburg), but herc we also have a low^Ter fortress circle wali on one side.

The circle area of the three fortresses is divided by streets, running north-south and east-west through the gates; in addition, a plank-covcred Street runs along the innerside of the w’all, as in Arhus. The wali itself is strengthened by palisades, galleries and ditches. At Aggcrsborg an additional four streets are known, dividing the space _% into tw^relve, instead of the usual four, blocks. The blocks are filled in with a square of four halls of the Trelleborg-type w^rith one major room, and two smaller ones at the ends. The long walls are curved and supported by oblique outer posts. In addition, minor, almost all rectangular, constructions are found, including a ^ktown-type' pit-house, at Fyrkat, with parallels in Arhus and Hedeby. The smali buildings lie in the middle of the blocks, at the gates, etc. (Fig. 29). The lower fortress at Trelleborg is madę up of halls and other long-houses lying in a semicircle from the east to the south gates, outside the main ditch. A bend of the outer wali and ditch leaves room for the cemetery. Also at Fyrkat a cemetery was found outside the main fortress.

The large Aggersborg fortress is situated at a Crossing by a narrow part of the middle Limljord inlet.⁸² In the Viking Age the Limfjord was open to the west, constituting a protected sea-route w^rhich was probably preferred to the navigation around Jylland. Fyrkat is in the river valley leading into the Mariager inlet, a few kilometres Irom the

Figurę 29 Reconstruction of the Fyrkat fortress (north Jylland). (Atter Olsen et al.)

opcn water. The wet areas helped to protect the fortress,controlling also the several fords and bridges in the area. Trelleborg lies in the fork between the confłuence of two strearns a few kilometres from the Storę Baelt. This site has by far the strongest ramparts and the broadest ditch, which powerfully reflects the need for protection. Ali the fortresses seem to lie at important routes of transportation, and especially at junctions. Odense also, incidentally, has such a position.

Traditional historiography has connected the fortresses with King Sven and his son Knud, in about 1000, and their campaigns and sway over England. In this context the barracks would be winter quarters and training canips for the Danish army. This preconception has hampered an objective dating of the artefacts from Trelleborg, which was the first to be excavated, and Aggersborg. The investigation of Fyrkat has shown that it was in use only for a short period, sińce repairs to it are few. It was apparently not rebuilt after a fire and has therefore given all the sites an unequivocal tenth-century datę, and, in addition, demands a new view of the function of the fortresses.

C^I4-dates, and especially a detailed analysis of the objects from Fyrkat, point to a construction at the mid-tenth century, before the campaigns in England, and a short-lived period of occupation in the second half of the tenth century. At Trelleborg there are traces of a smaller, ‘civiT, tenth-century settlement, perhaps a magnate farm to judge from a hall, prior to the fortress.⁸³ The artefacts have good parallels on Fyrkat and in other tenth-century finds, while they clearly differ from the picture at Lund at the beginning of the eleventh century. Furthermore the period in which it was functioning must have covered at least the last quarter of the tenth century, while it is unlikely that it goes back to the first half of the century. This is in accordance with dendrochronological data suggesting a datę around 980 for at least repairs of the site.

Aggersborg is preceded by a rural settlement of the eighth and ninth centuries, continuing into the tenth century. Dating the fortress is difficult, but it should be noted that elear eleventh-century artefacts are unknown within the rampart. On the other hand, the structures of Fyrkat and Aggersborg are so similar in detail that we expect contemporeneity between the two fortresses. All three sites are most probably constructed by the same architect and owner, the Jelling King Fłarald, and at approximately the same datę.

The fortresses are at almost the same distance from the Jelling centre and form, with the Danevirke walls to the south, the mentioned belt of defences around the centre of the State (Fig. 1). As noted below, the same belt contains a number of cavalry graves belonging in generał to the same settlement (Chapter 6). In this context, the heavily defeneed walls find an explanation, and the fortresses a function, not as training camps, but smali royal ‘burhs’ at the perimeter of the readily defined realm of the kingdom, controlling the border provinces and regu-