70 BUREAU OF AMERICAN ETHNOLOGY [Bull. 192 



Tsimshian and Haida, not from the Tlingit.) Another informant 

 reported that rafters, about 6 inches in diameter, were laid across the 

 ridgepoles and eaves beams on each side of the house. The eaves 

 beams were notched or grooved at an angle to hold the ends of these 

 rafters. On top were a series of longitudinal planks, parallel to the 

 sidewalls of the house, to support the roof planks proper. Another 

 informant denied such rafters, although his account was not clear, 

 and it is likely that the longitudinal planks or poles (corresponding to 

 intermediate beams) rested either on the wall planks at the ends of 

 the house, or on the gable rafters. 



The gable roof itself was made of two layers of planks, those on top 

 laid to cover the cracks between the planks below. These boards are 

 said to have been about 12 inches wide and 2 inches thick. On a 

 large house three sets might be necessary to reach from the ridge to 

 the eaves; on a small house perhaps a single set would be enough. 

 These planlcs were held down by horizontal logs or poles, laid parallel 

 to the eaves and ridge. One informant said that they were lashed 

 on by spruce root ropes. Another insisted that no lashings were used, 

 but that they were secured by wooden pegs set into drilled holes that 

 penetrated both the poles and the roof planks below. If no pegs were 

 used, the logs were held in position by short poles that ran down the 

 roof between them, with a longer set of poles running between the 

 lower log and the ground at the sides. At the ridge, the roof planks 

 on one side projected above those of the other to shed the rain. 



A small house had a gable roof, but this was supported by a single 

 ridgepole. Again, informants disagreed as to whether the latter 

 rested on a post at each end or on the wall planks, and also as to 

 whether or not extra beams and posts were used between the ridgepole 

 and the beams at the eaves. 



While one informant denied the existence of a smoke hole in the roof 

 and said that the cracks between the roof planks provided sufficient 

 ventilation, others insisted that all houses had a smoke hole in the 

 middle of the roof. This was equipped with a movable screen or 

 shutter, which could be tilted from one side to the other, depending on 

 the direction of the wind. Since two names for the windscreen were 

 given, there may have been two types. The more modern (?) type, 

 observed on an abandoned smokehouse, was a fiat, square plank affair, 

 fastened to a pole that ran longitudinally down the middle of the smoke 

 hole like a ridgepole. This pole could be rotated and the screen tilted 

 by pulling on one of a pair of ropes that hung down inside the house. 

 The older (?) type of screen was said to have lacked ropes. Instead, 

 someone had to climb onto the roof by means of a notched log ladder 

 and push the screen by hand. 



The main room of the house was heated and illuminated by a large 



