Originally, my sole intention with this publication was to give an account of the excavations I carried out in 1945-46 and 1948 on behalf of the Danish National Museum at the presumed Benedictine convent at Narsarsuaq in Uunartoq Fjord (Site No. Ø 149). To prove that this locality (discovered in 1932 by Poul Nørlund, and already then identified by him with certainty as the Benedictine convent known from Ivar Bardarson's fourteenth-century description of Greenland) really is the true site of the convent, I found it necessary to start with a close study of the church topography of the whole Eastern Settlement (including the Middle Settlement), with the emphasis on the area around Uunartoq Fjord. I have attempted to prove that the Benedictine convent actually was established at Narsarsuaq in Uunartoq Fjord, and that the church I found in 1946 at Narsaq in the same fjord is the Vagar Church of the sources. All the Norse churches of the Eastern Settlement identified up to 1946 are enumerated, and I offer some proposals as to where we might find those parish churches that are still unlocated, especially in the Middle Settlement.

Part II is entirely devoted to the initial intentions of this work: the archaeological excavations at Narsarsuaq. The excavation of the church and part of the surrounding churchyard is described in detail on the basis of a very comprehensive body of plan drawings, photos and my notebooks.

I have demonstrated that there were at least two Norse settlements at Narsarsuaq, the oldest of which may date back to the landnama period, and that there existed at least one church before the one excavated, which may be dated about 1300. I have also described the many very interesting finds of all kinds made at the site, especially from the oldest phase of habitation. Marie Stoklund and Søren Thirslund have each contributed with special chapters, the former on runic inscriptions, the latter on a unique wooden artefact thought to be a compass. The skeletal material has been studied by the anthropologists N. Lynnerup, B. Brølich, V. Alexandersen and J. P. Hart Hansen, who will be publishing a separate account of this material; but N. Lynnerup has kindly informed me of the main results of their studies, published here in sections 5 and 5a.