My post today is a modest contribution to some work that Kostis Kourelis is doing over at his blog. On Monday, he offered a brief post on Byzantine roof construction. My colleagues and I at the Pyla-Koutsopetria Archaeological Project have been thinking a good bit about roof construction also.
Our interests in roofs derives from our study of a 6th century A.D. annex room associated with an early Christian basilica. It seems clear that the room was two storeys and had a heavy roof covered with thick, flat, “Kopetra” type tiles. The second floor had collapsed into the first and there is no evidence for whether the room had an internal support. We did, however, find some large plaster fragments that we have associate with the tops of the walls.
The most interesting fragment (and please excuse my sketch) looks like this:
The chunk of plaster preserves the impression of the beam that ran along the top of the wall. It also had the impression of the perpendicular rafter which sat atop the beam. On the top of the piece of plaster we discovered the impressions of reeds or small sticks. This must have been the layer of rushes, beanstalks or rushes described in the 4th century B.C. inscription cited in Kostis’ post. The tiles would have sat atop the reeds preserved in the plaster impression.
The plaster impression preserve some evidence for the construction process. It would seem that the wall beam and rafters were set into place and the gap between the rafters was filled with plaster (or a kind of mortar) immediately before the reeds were put in place atop of the rafters. The plaster or mortar would have had to be still wet for the small reeds to make an imprint. In effect, this piece of plaster preserves the roof as it was being built. This makes some sense, I suppose, because it ensured a strong seal between the roof and the wall to prevent water from entering the wall and weakening the rather humble used to bond the stones.
David Pettegrew and I have documented a series of roofs at the opposite end of their life cycle at the early to mid 20th century rural site of Lakka Skoutara in the Corinthia (follow this link to an archive of over 600 photographs taken over 10 years at the site).
The roof of House 5 showed no evidence of the mud-and-reed layer between the tiles and roof structure although it is possible that this layer eroded away quite quickly as the roof deteriorated. The roof of house 5 appears to have been supported by some well-cut timber that set atop walls reinforced with cinder block suggesting that a fairly recent date of construction.
The roof of House 3 shows a similar construction style with the use of more rustic rafters. The mud plaster interior walls stood until early in this century (2001), but deteriorated rapidly after the roof collapsed.
House 14 was the only house in the settlement that preserved the mud-and-reed packing between the tiles and the rafters. This photograph is from 2001 and the packing remains barely visible. By 2009, the entire roof had collapsed and evidence for the mud layer under the roof tiles was lost.
The temptation to recycle the precious roof tiles even in our century manifest itself in the roof of our house 2. The first photo is from 2001 and the second from 2002.