As settlements became more clearly bounded and fixed in the landscape, so too did territories based on landed production, which became increasingly intensive and politically controlled (as we shall see in Chapter 5). These territories became formalized when leaders were able to exercise authority within them by protecting clients through juridical and/or military means, and by extracting surplus from, and controlling access to, landed resources. The identification of communities and individuals with a particular territory or region, whether this was defined by shared markets, dialect, military allegiances, or other commonalities, must also have grown in importance in this period, as ties of ethnicity and kinship began to give ground to bonds of clientship and rank. The formalization of territories was of course key to the formation of early kingdoms. What can archaeology tell us about the effects of territorialization and estate formation on rural communities? Certain regular features govern territorial formation in pre-industrial societies. In particular, universal ‘push–pull’ factors underlie the territorial structure and settlement pattern of agrarian communities. Briefly stated, every community needs to establish a territory in order to keep neighbouring communities at a distance and preserve its resources (‘push’ factors), but the necessity of maintaining certain social ties between communities, such as marriage, trade, and shared defence (‘pull’ factors), will act to minimize the distance between them (Heidinga 1987, 157). For example, the distribution of settlement in the Veluwe district of the central Netherlands shows that the northeast and the southwest regions were largely empty in the seventh century, even though their soils were suitable for farming and they were occupied both before and after this period. They lay outside the core area of the seventh-century resettlement of the Veluwe, however, and it appears that communities chose not to spread out thinly across the entire territory, but rather to remain relatively close to one another (Heidinga 1987, 162). In the Netherlands, Germany, and England, early territories could, under certain circumstances, be remarkably stable and survive to be detected in much later boundaries (e.g. Waterbolk 1982 and 1991a; Cunliffe 1973; Janssen 1976). In view of this stability and the behavioural ‘rules’ which appear to govern territorial formation, some archaeologists have attempted to reconstruct proto-historic territories. Several presuppositions underlie such reconstruction. The first is that the ‘push–pull’ factors already mentioned invariably operate between neighbouring communities.