526 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1956 



lowed us to camp overnight at Prairie Point and examine an old 

 Sadlermiut village site of 15 stone and sod house ruins. 



At Native Point we had another busy and productive summer. Ad- 

 ditional excavations were made at the Sadlermiut site, at the early or 

 proto-Dorset site, T 1, the later Dorset site, T 2, and at a third site, 

 T 3, which appeared to be slightly younger than T 1. A large body of 

 material was excavated which strengthened and rounded out the 

 archeological picture obtained the previous year. 



We planned to make another reconnaissance trip to Coats Island, 

 this time toward the southwest end, and on July 20 we set out in the 

 Nayavak for that purpose. As it was not much out of the way we 

 decided to stop briefly at Walrus Island where there were several old 

 house ruins we wanted to examine. 



Walrus Island is a small granite islet 25 miles off the south coast 

 of Southampton. The six house ruins lay in a valley extending east 

 and west across the south end of the island. The three oldest-looking 

 houses (Nos. 1, 3, and 5) consisted of a single room round to oval in 

 shape (pi. 14, A). Another (No. 4), more recent in appearance, had 

 two oval-rectangular rooms. The two remaining houses (Nos. 2 and 

 6), also recent looking, were cloverleaf in shape, with three rooms 

 (pi. 14, B). The house walls had been made of massive blocks of 

 granite piled one above the other. Most of the houses were deep and 

 all had entrance passages from 5 to 10 feet in length. In some cases 

 natural rock ledges and huge boulders in situ had been incorporated 

 into the house structure to serve as parts of floors, walls, or sleeping 

 platforms. House No. 6, the best preserved of the group, had upright 

 stone pillars — roof supports — still in place, and fallen slabs indicated 

 that the roof itself had been made of stones as in the case of the 

 Sadlermiut houses on Southampton and Coats Islands. The absence 

 of roofing slabs and supports in the other houses suggested that the 

 roofs had been made of skins. 



When we began to excavate we had naturally assumed that these 

 well-preserved house ruins were of Sadlermiut origin. They were 

 similar in general structure and two of them, Nos. 2 and 6, had the 

 cloverleaf shape characteristic of many Sadlermiut houses. More- 

 over, some of these Walrus Island houses had been partially exca- 

 vated in 1936 by the British Canadian- Arctic Expedition (Manning, 

 1942) and found to contain material described as resembling Sadler- 

 miut, with only a few Dorset artifacts which might easily have been 

 explained as relics. We had not been digging long, however, before 

 we began to suspect that the houses were Dorset rather than Sadler- 

 miut. With this unexpected development, we stayed at Walrus Island 

 for five days instead of going on to Coats Island as originally planned. 



Our excavations in and around five of the houses produced over 100 



