Frankish Roof Tiles: Morea

Medieval roof tiles tend to be thrown away in excavations because they are both plentiful and undiagnostic. In contrast to beautifully shaped stone or clay tiles from antiquity, medieval tiles seem vernacular, produced locally with less regulation or form-work. Luckily, the discipline of medieval tiles in the Peloponnese is witnessing a renaissance. My experience with medieval roof tiles began in 1991 at the Morea Project. One of the few things that makes a medieval settlement recognizable on the landscape is the scatter of roof tiles. If tiles is all you got, then you start taking them seriously. While surveying countless medieval settlements in the Peloponnese and looking at thousands of tile fragments, the eye began to recognize certain patterns. A few analytical drawings and reconstructions produced a distinctive tile system that Fred Cooper articulated most succinctly in our project monograph, Houses of the Morea (Athens, 2002), pp. 23-35. I will quote Fred's analysis:"The distinctive Frankish roof tile system has a different shape altogether for the pan and the cover. The pans are somewhat wider than the Byzantine (ca. 0.25-0.30 m.) and the flanks turn upwards sharply to for an edge for insertion underneath the closing edge of the incumbent cover tile. The cover is quite broad (0.20-0.30 m.) in comparison to older and later examples. In section, the arch forms one ogive or, at least, the arch tightly curves at the peak. Cover and pan are relatively thick, ca. 0.025 m. (figure 1).

Another distinctive feature of the Frankish variety is the interlocking system of covers above pans. Adjoining pans are separated by the span of the cover, rather than the reverse that is typical of other types of roof tile systems, in which the lower or pan tile has flank sides abutting, while a narrower cover tile spans a width no more than necessary to close the joint. In the Byzantine system, bothpans and covers abut.

Still another characteristic of the Frankish type is the fabric: course and mottled black to red by uneven firing. The underside or unexposed surfaces are quite crumbly, often having straw embedded, a result of nested stacking. Exposed surfaces are more evenly fired and usually bear finger impressed loops on both covers and sometimes on pans.

In an exploratory trip to Burgundy, France, I found the same type of tile system in use on extant Romanesque church roofs, such as Vezeley, as that used for the churches of Glarenza and Andravida and at our kastro sites. This is to say, the technology, but probably not the tiles themselves, came directly from the Frankish homeland of the Morea occupiers. The topic of Frankish roof tiles is one of several which I will pursue further. In any event, the presence of Frankish tile fragments on the ground at sites is taken as evidence for a medieval date."

Fred Cooper has, indeed, done a lot more fieldwork in Burgundy since this publication, but he has not published his conclusions. One more thing I'd like to add. Right around the time that the Morea Project began, Charles Williams had been excavating a Frankish complex in Corinth. In his early explorations, he thought that he might have identified a distinctively Frankish (versus Byzantine) roof tile. But in the course of the excavations, he seems to have grown more and more uncertain about the diagnostic potential of the tiles.

Cooper had encountered the Frankish tiles in two additional projects, in the Minnesota excavations of the Frankish Cathedral in Andravida and in his excavation of Karl Blegen's dump in Pylos. Yes, you read correctly. Nestor's Palace in Pylos contained Frankish material. Blegen excavated Pylos in 1939 and the 1950s. As was common, he tossed all the medieval roof tiles in his excavation dump. Cooper talks about finding newspapers that provided termini for the layered dump that he re-excavated in the 1980s.

Cooper's Frankish tile thesis needs obvious refinements and has been a topic of intense debate among the Morea Project staff. Even at its preliminary articulation, the roof tiles of the Morea Project thesis presents invaluable evidence in roof tile scholarship. It should also be noted that the Morea Project was an architectural survey and not a pedestrian surface survey. We were, thus, forbidden from collecting any archaeological material from the ground. We simply noted it, sketched, and photographed it. This means that we, unfortunately, do not have a sample study collection. Luckily, there is no shortage of these roof tiles on the ground. Here is an example from Kastro tes Orias (pp. 116-118), photographed in 1994.