Stengrunder och gränser en studie av kontinuitet från äldre järnålderns stengrundsbygd till nutid

This thesis studies settlement continuity on Gotland between approximately AD 200 and AD 1700. The method used is to calculate correlation between the geographical distribution of all known Iron Age stone-wall-houses (on Gotland known as "kämpagravar") and all farms included in the detailed taxation maps from approximately 1700. The number of remaining house foundations is between 1800 and 1900. A model to estimate the number of removed foundations is presented. It is based on the assumption that the settlement density was proportional to land use around 1700, and that the rate of removal is related to the current land use. Based on similarity with contemporary farms on Öland and in Denmark, it is also proposed that Iron Age farms on Gotland were less dispersed. The common view in archaeological literature is that Iron Age farms on Gotland often had their buildings spread more than 200 meter apart. It is estimated that the remaining 1800+ foundations represents more than 2700 Iron Age farms with a total of more than 4700 houses. Compared to previous studies, the fluctuation in number of farms between maximum expansion during Late Roman Iron Age and Viking Age, and maximum contractions during the Migration Period and Late Middle Ages, also is much larger. By parish, the number of farms per km 2 and the average farm size have been calculated, both for the stone-wall houses and for the farms as they were represented around 1700. Using regression analysis, the correlation between the two datasets was estimated. No significant correlations were identified. The distribution of the remaining stone foundations and/or the estimated distribution before removal have also been compared to known borders and corresponding administrative districts – since the Middle Ages, Gotland has been divided in tredingar (third parts), settingar (sixth parts), 20 thing/court districts and 95 parishes. The tredingar and partly the settingar correlate with the stone foundation distribution, but not the thing districts or parishes. The correlation and lack of correlation might be explained by input errors or confounding factors, but the historical records supports the interpretation that the tredingar and settingar has been in continuous use since before the Middle Ages, and that the thing districts and the parishes were introduced during the Middle Ages. At least 24 farms with the name "Stenstugu" are known in Gotland. "Stenstugu" distribution coincides with centrums of stone foundati