Journal of Historical Geography, 22, 1 (1996) 3–15 Multiple estates and the origins of the manorial structure of the northern Danelaw Dawn M. Hadley Large multi-vill estates have long been recognized as the basic unit of exploitation in the early-medieval period. Although many have sought to distance themselves from the “multiple estate” model developed by Glanville Jones (largely on the grounds of the lateness of the evidence he used), most would still accept that early-medieval England was everywhere divided into large multi-vill territories which fragmented in the later pre-Conquest period to form the manorial units recorded by Domesday Book. In the northern Danelaw large estates were, however, still common in the eleventh century and these have been explained as relict features of early-medieval organization. This paper considers some of the flaws which remain with the “multiple estate” model and its variants, and examines the problems inherent in attempts to uncover early territorial organization. It also argues that although some sokes may be a surviving feature from pre-Viking organization, there is evidence to suggest that other sokes were much later creations and that estate amalgamation as well as estate fragmentation were features of the tenth and eleventh centuries in the Danelaw.  1996 Academic Press Limited The complex manorial structure of the northern Danelaw [1] was characterized in Domesday Book by large estates (or sokes) consisting of a central manor with outlying berewicks and sokeland, and it stands in marked contrast to that of regions to the south and west where manor and vill more commonly coincided. [2] It has been claimed that the sokes of the Danelaw were remnants of pre-Viking organization, given their similarity to the “multiple estates” believed to have been the archetypal unit of land exploitation in the early medieval period. [3] The details of the “multiple estate” model, as developed by G. R. J. Jones, have been subject to criticism, which need not be reiterated here at length. [4] It is su fficient to note that many have rejected the model chiefly on the grounds that the evidence used by Jones is very late, and they are unable to accept that the social structure and forms of exploitation he describes could ever have characterized the early-medieval period. Nonetheless, although many studies would seek to distance themselves from that particular model, many have accepted the basic principle that early-medieval society was commonly organised in multi-vill estates, and numerous studies have sought to identify a network of such estates through the use of documentary, cartographic and place-name evidence. [5] These have come to be known as either “multiple” or “federative” estates or by their local name of shire, soke, lathe and so on, and one historian has commented that it is really only the choice of local terminology that varies in the pattern of local organization that has been identified. [6] This paper considers whether or not the territorial sokes of the northern Danelaw really 3 0305–7488/96/010003+13 $12.00/0  1996 Academic Press Limited