A long, narrow fireplace usually extended through the middle of the principal room, and an essential feature was the cooking fireplace, which was about one metre square. These were either paved or surrounded by upright stones. The plan is of the ruin of the house of Eric the Red in Haukadalr, Iceland. It shows the different forms of fireplace, and that the walls, which were built of turf, were one and a half metres thick. Outhouses were often dug into the hillside, and were sometimes walled up on the inside with stone and turf. Ruins of such old settlements in Iceland are usually low, grass-grown ridges and hollows. When Professor Horsford first visited the site which his study of maps and literature had led him to believe was Vinland, he found a few hollows in the hillside and also some broad, low ridges on the level ground, indicating that a building about twenty metres long by five metres broad had once stood there. There was also a mound some distance away which has since proved to be of modern construction. [Illustration: PLAN OF SUPPOSED NORSE RUIN IN MASSACHUSETTS.] No digging was done here until after Professor Horsford's death, with the exception of a few trenches across the supposed site of Leif Erikson's house on the other side of the creek. In 1896, during a visit of Dr. Valtyr Gudmundsson and Mr. Thorsteinn Erlingsson, of Copenhagen and Iceland, extensive excavations were made, leaving practically nothing unexamined at this site. Three kinds of earth were revealed. The upper layer was of black loam from thirty to forty centimetres deep; below this was a yellow soil of sand and clay thirty centimetres deep; and below that again the sand and gravel which had remained undisturbed since the close of the Glacial epoch.