Macellum: August 12-25, 2007

Excavation of the Macellum continued for two more weeks, under the direction of Julian Richard (KULeuven, Belgium), Firat Kesim (Istanbul Technical University), and Marijke Van Looy (KULeuven, Belgium), work ending on August 23. During these two last weeks of the campaign, we focused on the lower level of the portico bordering the central courtyard to the south, an area we started excavating in the fourth and fifth weeks of the campaign. After examination of the debris fallen from the portico, we could excavate deeper to reach the level of the foundations. Documentation of both the collapse process and the structural characteristics of the south portico is one this campaign's most important results, resulting in a much better understanding of the Macellum's architectural organization.

The southern tunnel during excavation



View of the northern tunnel

The south terrace wall terrace supporting the complex--described in the previous report--consisted of a 1.75 (west) to 1.90 m (east) wide rubble wall preserved to a height of 2.35 m. The wall's southern face is made of large rubble blocks, but the northern is a more carefully built wall composed of smaller rubble stones. The reason for this is the presence at the back of the terrace wall of one of the two tunnels mentioned two weeks ago.

We were able to examine the tunnel system under the southern part of the Macellum's courtyard in more detail. The hypothesis presented two weeks ago--that they were conceived as substructure tunnels supporting the terrace--should most probably be re-evaluated. Indeed, the two tunnels do not run parallel: the southern one follows the orientation of the south portico, but the northern one clearly shows a divergent orientation. We explored its course over ca. 20 m: it runs first in an east-northeast direction for ca. 5 m before turning in the direction of the central tholos. A partial collapse of the tunnel prohibited further exploration, but we could observe that the tunnel clearly goes further towards the northeast.

The structure of both tunnels is similar. Their size is remarkable: the northern tunnel, whose elevation is entirely preserved, is 2.24 m high, with a width of ca. 0.60 m. The walls are made of carefully assembled rubble stone, with some sections made of brick masonry. The upper part of the elevation is made of a barrel vault also built in rubble masonry. The architectural configuration and measurements of the southern tunnel are similar. A striking element is that they do not show any traces of waterproofing, such as a hydraulic mortar revetment or a tile floor, expected in the case of tunnels dealing with continuously running water. However, small quadrangular, secondary channels ending in the northern tunnel just under the level of its vault should be identified as tributaries emptying into the main tunnel. The current hypothesis is that both worked as emergency drainage system.

This hypothesis is supported by some structures observed at the tunnels' endpoint. The hypothesis of a "collection chamber" could unfortunately not be confirmed, as the whole area where both tunnels meet has collapsed. Farther south, however, some structures excavated along the main terrace wall can throw light on the functioning of the system: it appears that the massive terrace wall was originally provided with an opening shaped as an arch made of brick masonry (preserved height: 0.60 m; with: 0.54 m) whose left pier only was preserved. It should most probably be identified as the original outlet of the tunnels. However, we could observe that the system was partially blocked at a later date: the right pier of the arch was removed and replaced by a fill of rubble, combined with a wall oriented to the south. The function of the wall was clearly to divert water: its western part is indeed carefully coated with mortar whereas its eastern face is simply made of piled rubble. A clear distinction in the stratigraphy could also be observed: to the west--where water was supposed to be collected--we found a very fine-grained clayish deposit containing a substantial amount of ceramic sherds. The nature of the deposit and the fact that some of the potsherds were rolled and eroded by water indicate that we are facing a deposit induced by water, most probably originating from the tunnel system and diverted by the "dam" wall. The reason for the modification of the evacuation facilities and the reduction of their scale was probably because of the structural weakness of the area: three brick buttress walls added at a later date at regular intervals along the south terrace wall indicate that it needed to be reinforced at some point.

We also focused on the eastern part of the courtyard, where we exposed the original pavement in two sectors. We noticed a repaired zone (2.40 x 1.55 m) along the staircase of the eastern portico: the original pavement was emptied and partially refilled with rubble stones and re-used blocks, such as one of the molded parapet blocks once standing between the column pedestals of the tholos. The presence in the topsoil of column fragments of another type than those composing the west and south porticoes--i.e. fluted columns in marble and unfluted column shafts in limestone--indicate that the area must have undergo substantial repairs at a later date. This hypothesis is supported by the discovery in the third week of the excavation of a later dedicatory inscription (see web report 2), preliminarily assigned to the fourth century A.D. by our epigraphists.