landscape features, the goodness of the soil and, of course, the historical development of the estate. With regard to the close relationship between the late prehistoric settlement and the earliest Medieval settlement (Callmer 1986) it should be possible to find traces of these patterns in a diverse source material including early historical settlement data, archaeological prospection and excavation data and place names. It should be characteristic of these traces of estate structures that the centre after its heyday is much reduced in relative size and qualities. What could remain is above all the structure i.e. an eclipsed solar system. We may also set up the possibility of a complete reduction of the centre. This could result in a gap in the settled landscape - a black hole -but still visible.
The later division of the territory of the centre would leave distinct traces in boundaries and perhaps in names. Only a reestablishment of settlement after a hiatus on the same spot and with a new name would cause serious problems.
Fig. 1. The surroundings of the Uppåkra settlement with phosphate data according to Arrhenius 1934. Only values of 100 P0 and more are indicated.
Some minor areas are not included in the survey of Arrhenius. They are not mapped here for reasons of the readability of the map. Contours are
given for 5 m, 10 m, 20 m etc a.s.l.
- """
but the divergence cannot be great. What we can conclude is that we are confronted with an agglomerated settlement of unusual size. The question is if we have positive evidence of a comparable settlement in Southern Scandinavia in the Late Iran Age.
This exceptionally large settlement was reduced and restructured in the late 10th or early 11th century resulting in two villages: a larger one counting at the end of the Middle Ages 16 farms anda smaller one with 7 farms (the hideage 1660 is however 10.5 and 2.5 respectively).We must here add that rnid 17th century farm numbers and hideage normally diverge little from figures in the late Middle Ages when we first have ac-cess to them. lf we campare conditions ca. 1660 the combined hideage of these two villages of 13 hides is not exceptional for other villages in the neighbourhood (the territory mapped in fig 2).
No less than fifteen villages have hideages between 10 and 18.5. Also the number of farms taken together for the Uppåkra villages is by no means exceptional. Six villages have considera-bly higher numbers (24 to 37 farms).This means that there is an obvious discrepancy between the ratio of hideage and number of farms of the 17th century ( and probably as well for the late Middle Ages) and size relationships in the prehistoric period between Uppåkra and settle-ments in the neighbourhood. Uppåkra was thus not only divided but also dramatically reduced with regard to hideage and the number of farms. At this point we must also introduce the phosphate map which covers a very considerable part of the surroundings as an important source of complementary information. A rather safe indicator for settlement sites is a value of over 100 P degrees. It is however essential to consider the natural phosphate level of an area. In the eastern part of the plain between Malmö and Lund we have a zone with unusually high natu-ra] values beginning just to the southeast of Lund in the parish ofBjällerup, widening towards the east and there joining up with another wedge-shaped area beginning at the contemporary po-pulation centre of Staffanstorp. This zone with few exceptions covers the innermost part of the plain bounded to the northeast by the horst of Romeleåsen and to the south by the hummocky
landscape of inner southern Scania. In the closer surroundings of Uppåkra (within a radius of 6.5 km) this anomaly is not disturbing the picture and the interference in the defined zone in the inner part of the plain is only moderate. The historical village sites are well visible as nuclei with elevated phosphates. Additional areas of 2 to 6 ha are often found in the immediate neighbourhood of the sites and frequently we can observe that the site nucleus extends con-siderably outside the historical plot boundary.
The big problem when we use the phosphate map is the question of dating . In this part of Scania we can draw on a considerable number of excavations which could be compared with phosphate data.
Phosphate areas in the neighbourhood of histo-rical village sites and extensions of phosphate nuclei outside the village boundaries often must be given Late Iran Age dates (including the Viking Period; Callmer 1986). Uppåkra is sur-rounded by a number of often rather small phos-phate nuclei sometimes with closely adjoining small phosphate areas. In details we will come back to these indications.
We must however first return to the histori-cal settlement around Uppåkra where we can observe that the two historical Uppåkra villages are surrounded by a few small-sized, but pre-dominantly middle-sized and even big villages both when we consider the hideage and the number of farms. Very big villages are only found at a distance i.e. to the north of the Höje River and on the northern fringe of the hummocky landscape to the south of the Ton-eberga peat bog. The names of the villages surrounding Uppåkra are special when we remember that we are in the centre of a South-Scandinavian plain landscape which has been densely settled since the Neolithic. The usual pattern in Eastern Jutland, on the Danish Isles and in Scania is a dominance ofnames ending in -hem, -löv, -lösa, -stad, -inge in the central plain districts. These names are on linguistic grounds thought to represent a chronological stratum generally with a pre-Viking Period date. When names with a classification as definitely younger types occur in this cultural landscape with a long continuity they can rather easily be shown to have a secondary status and the mother unit can often
be identified. Among the younger names those ending with -to11J, -ryd, -böle, -hult are most imporianl. They ai·e generall.y thought to be of Late Viking Period date or even later. Villages with these names are mostly found in the periphery of the plains and river valleys, in now completely transformed ancient woodland tracts and in still existing woodland ( outer colonisation) and, as already stressed, in central settlement areas as secondary settlements (inner coloni-sation). An intermediate group is composed of names with endings in -by and -tofta. They are generally dated to the Viking Period but were formative probably also earlier. A special pro-blem is connected with the torp-names. Some scholars have maintained a longer formative pe-riod also for this ending. The remarkable condition prevails that in the wide plain between the Höje and the Sege rivers we have hardly any names of the old stratum except along the northern bank of the Sege river and at some distance to the east along the Höje river (the villages of Väsum and Kvärlöv). In fäet the two Uppåkra villages are surrounded by a corona of nine villages ending in -torp. Only to the southeast is there the vil-lage of Gullåkra with a deviating name. Further-more there is an outer row of four -torp villages along the coast of the Sound. Also to the east there are an additional three villages with this ending. Along the northern bank of the Sege river and directly adjoining we find four villages with -löv names indicating a completely different structure of settlement. Elsewhere in this area between the Höje and Sege rivers there are no traces of an earlier settlement structure connected with the old stratum of place names. It is however unlikely that this entire territory has been restructured after the Viking Period and thus veiling another earlier and "normal" settlement structure. Consequently we can advance the hypothesis that the structure of the centre with a corona of secondary settlements is contemporary with one or more phases of the large settlement atUppåkra.
This hypothesis can in fäet be further corroborated by excavation data and stray finds.
Recent excavations at Hunnerup on the southern side of the Höje river immediately to the north of Uppåkra have revealed settlement traces ofLate
I ron ge date Jacob on 2000: 164). There i a pho. pba1e pa1ch I e by. At Hjärup ca. 2 km L the we L of Uppåkra a duster of three . mall pho. phate nuclei indi ate the slighLly changing po ili n of a ·ettlemenl going back to the 6th -7lh century and by ca. AD 1000 more or less fixed at the village site (Jacobsson 2000: 199, 201). The 'situation is similar at Åkarp where several stray finds indicate a continuity within a small duster from the Roman Iron Age to the historical village (Strömberg 1961 ). It is important to point out that the hideage in the 17th century for both these villages is considerable (7 for Hjärup and 9,5 for Åkarp) indicating real growth.
At Brågarp to the east of Uppåkra we have a similar picture with a small duster of phosphate patches. An excavation at one of these yielded early Viking Period material (Jacobsson 2000:
189).The hideage 1660 for Brågarp was 3 5/8 and the number of farms 5. We also have data from two of four torp-villages along the coast of the Sound. Excavations slightly outside the histori-cal village site of Karstorp yielded Late Iron Age and Viking period material (Ambrosiani &
Magnusson 1972). The hideage ofthe historical viUage of Kar torp wa. four and the number of farms eight. r m Alnarp there is a stray find of an 8th century fibula immediately to the west of the park (private collection). No reliable early historical data are available for Alnarp. Also in these two cases is there a close connection between the archaeological observations and the phosphate picture. The village of Knästorp is situated ca. 1.5km north-east of Uppåkra on the southern bank of the Höje river. From the opposite northern bank of the river there is a find of a bead of the Middle or Late Iron age (SHM Stockholm). There is a very distinct phosphate nudeus on this side of the river. It is also of interest that the northern part of the parish i.e. to the north of the river at least from the 16th century on was distinguished as a separate unit, but at that date without any farms. It is thus probable that a part or the entire settlement was situated to the north of the river in a phase preceding the early historical village (12th century parish church). All together we have to our disposal a considerable supporting archaeological material. Very important are the chronological
data, which show us that many (probably all) of these settlements around Uppåkra are of Viking Period date or even earlier.
The prereform agricultural landscape of the Reformation period and the 17th century has played an important role for our considerations concerning the structure of the late prehistoric cultural landscape. It is then very important to remember that the later village boundaries have little to tel1 about earlier division of territory. It is the location ofthe settlement which is important.
The sometimes very considerable hideage and the number of farms of the discussed torp-villa-ges indicate that some of the villatorp-villa-ges have grown very much, presumably above all in the Early Medieval Period (before the 14th century). Other settlements have remained quite small like Tirup, Tottarp and Stanstorp. From the beginning the torp- settlements have been small units of a few farms or justa single farm. They have been esta-blished not only in the Viking Period but certainly also earl i er. It is however very uncertain if these settlements are of the same age. A successive development of the domain seems to be more likely. This question can however not be discussed in detail. The very short distance e.g. between the historical villages Vinstorp and Karstorp, in the row along the coast, of only ca. 250m could indicate a late split. On the other hand we may contemplate the possibility that the closest sett-lement to the south of Karstorp at Alnarp could be only a little more than 500m to the south ( cf.
above). We must also consider the possibility that other torp-settlements have been abandoned already in the prehistoric period or in the Middle Ages without leaving any traces. After all these abandoned settlements can not be numerous since they should, given an existence of a few generations, tum up on the phosphate map.
We can now specify our hypothesis some-what further. Uppåkra was not just a centre of political, econornical and religious importance, it was also the centre of a vast and complex estate extending from the coast of the Sound and covering the majority of the land between the rivers Höje and Sege to a depth of ca. 10 km. The estate certainly comprised Lomma, the closest potential harbour ofUppåkra. A more complicated question is the position of Värp inge and
Trolle-berg ( earlier Värpinge gård).It is not to be excluded that Värpinge with a strong aristocratic tradition throughout the historical period should be understood as a separate domain adjoining the central estate. Another intriguing question is to what extent the estate extended to the northern side of the Höje river. An indication is already given by the special situation at Knästorp. There are at least two late prehistoric settlements on the north side ofthe river between Värpinge and Knästorp. At Källby graves and settlement finds indicate a continuity from the Late Roman period to the Early Middle Ages (Callmer 1986; Wilson 1955).The settlement at Klostergården/ St. Lars Hospital appears as a considerable phosphate nucleus and has yielded a 6th century inhumation grave (LUHM). It seems very likely that these two settlements directly on the river bank were included in the estate. The extension of a strip of the estate on the northern bank of the river would also solve the old problem how the village of Rå.by (the village at the boundary) related to a boundary. It could then be that the situation of the village on the no1thern boundary of the estate was the reason for the name rather than the relationship to the urban centre of Lund (late 10th century) in the west. Continuous settle-ment at Stora Rå.by goes back to the 7th-8th centuries atleast (Callmer 1986). This outline of a boundary of the estate to the north of the river gives a new possibility to approach the problem of the foundation of Lund. In a situation when large parts of Scania were taken over directly by the Danish king (probably a development in the 10th century) it was natural that the estate at Uppåkra became his domain. If this king was Harald Gormsson it is likely that he was accompanied by his ambitious son Sven. It could have been interesting for Sven to build up a power basis of his own close to the traditional centre of Western Scania (the Torn-estate).
Further upri ver the situation is more complicated with the historical villages ofVäsum and K värlöv, i.e. with names indicating another structure and possibly a higher age. It is however clear that we have to reckon with considerable rearrangements of settlement here. At K värlöv with 6th-7th century finds (Strömberg 1961:61) we have probably a subsequent reduction, in fäet to a
status as a single farm, whereas settlement on the other side of the river at Stora and Lilla Bjällerup had a strong development. Settlement at Stora Bjällerup begins in the 11th century if not earlier. The Bjällerup villages have a hideage of 11,5 and 11.25 and the numberof farms is 13 and 14 respectively. This means that the pre-conditions for expansion in the Medieval Period were very favourable or that farms were relocated from Kvärlöv to the new torp-villages. On the east side ofUppåkra the situation is agairnrather clear with the hypothetical boundary of the estate running east of Brågarp and Stanstorp (today Staffanstorp which however has nothing to do with St. Stephen). Hemmestorp lies on the east side of Brågarp. It forms a kind of corridor between the estate and an intriguing block of three villa-ges with the name
*
Heddinge to the southeast (cf. below the comparable Heddinge complex on Stevns, Sealand). The historical territory of Hem-mestorp is however certainly partly a result of the already noted reduction of K värlöv. Along the southern boundary there is as pointed out above a single row of villages from Arlöv to Djurslöv which certainly were not part of the estate structure. Only Burlöv is somewhat remo-ved from the Sege river.Let us recall the strong concentration of po-pulation in the central settlement at Uppåkra.
Any agricultural strategy to solve the problem of supporting so many people must go for differentiation. It is then very likely that the torp-settlements were specialised. To decide the character of specialisation is mainly a question for future research. It may however be noted that there were excellent conditions for pasture along the coast of the Sound and in the river valley of the Höje river.
There are in addition to the Uppåkra and Värpinge estates two other anomalies which probably are related to it. To the south-east ofthe estate a wide stretch of the fertile plain was occupied in the Late Middle Ages by three villa-ges: Kyrkheddinge, Koddheddinge and Moss-heddinge. Since there is no distinct landscape formation which possibly could have contributed to the naming of the whole area in question the peculiar name giving must have another explan-ation. The reason for the uniform naming must
rather be sought in its being a special unit later divided. Under these circumstances it is temp-ting to inlerprel Heddinge a another separate or adjoining domain. The p.ho phate map ha some very !arge p.h sphate nuclei indicating sub-sequent minor changes of location. In the case of the village of Kyrkheddinge excavation data suggest the establishment of settlement on the historical village site in the Viking Period but settlement nearby is earlier. An extensive pho pl1ate nucleus stretches along the Höje river nearby. There is not place to discuss lhis prob-lem further here but we must not forget to point at the similar, impressive Heddinge complex on the peninsula of Stevns (Sealand) clearly a royal estate turning real town in the 13th century or even earlier.
When we consider the historical settlement of the region at the transition from the Middle Ages to the Postreformation Period the size of Dalby in the north-eastern periphery ofthe plain calls for our attention. The hideage in the 17th century is 18,5, a value only reached by one other village in the area (Stora Råby ). The number of farms, 37, is extremely high but probably the resµlt of egalisation during the late Middle Ages when Dalby still in its totality was a monastery domain. The premonastery history of Dalby is very interesting since we must reckon with an early royal estate here with a manor, which was the basis for the erection of the pfalz-inspired palace and church complex by king Sven Estridsen shortly after the middle of the 11 th century. The early palace complex of king Sven has recently been reconstructed by Andren (1999:387f.). The complex consisted of the palace and the church centre, an adjoining village, possibly including a production and trade cen-tre and a deer park. Andren calls to our attention the Christian symbolic background for the complex. This is certainly correct but the high probability that there was some sort of pre-decessor suggests that there may be pre-Chris-tian roots for an estate here at the fringe of the plain and within sight of Uppåkra. The 11th century complex of Dalby rather tums its back to the horst of Romeleåsen . This dominant forma-tion runs 30 km to the Southeast. The importance of an early estate at Dalby may have been the
Fig. 2. The surroundings of the Järrestad settlement with phosphate data according to Arrhenius 1934. Only values of 100 P0 and more are indicated. Some minor areas are not included in the survey of Arrhenius. They are not mapped here for reasons of the readability of the map. Contours are given for 5 m, 10 m, 20 m etc a.s.1.
direct access to wide hunting grounds. The his-tory of deer parks in Anglo-Saxon England and on the continent suggests a pre-Christian back-ground for them. They are documented from the Merowingian and Carolingian periods on the continent (Fenske 1997). There is also an intriguing indirect argument. This is exactly the period when antler in great amount becomes available fora specialised comb production. The most likely background for the high proportion of shed antler in workshop finds from Uppåkra (5th-6th centuries; cf. Lindell this volume) Ribe, Åhus (both 8th-9th centuries) and Hedeby (9th-10th centuries) is the keeping of deer under controlled conditions when the shed antler can be easily collected. When deer roam freely in unbounded woodland the likeliness that intact antler is found in numbers is very slight. A med-ieval deer park at Dalby may have pre-Christian roots and may have been an important part of the royal domain in western Scania.