During the first half of the 1980s the Ecole Francais de Rome undertook various campaigns of archaeological investigation of the site of Monte Castellaccio, which has been identified as the city of Vaccarizza. In 1999 attention was focused on the mound where the Byzantine fortified citadel ( _pretorion_ ) stood. The chronology of the archaeological deposit runs from 10th to the 13th century. In 1999 exploration of the eastern sector began, where several structures and occupation levels relating to the final phase of the citadel in the Byzantine period were identified. A number of rooms were uncovered belonging to a craft workshop. These abutted a massive wall with irregular facings, built in several phases using different construction techniques, which determined the morphology of the entire mound. The perimeter wall seems to have been used for defensive purposes until the beginning of the 11th century, a period which coincides with its first mention in a written source in occasion of the battle between Melo and the Byzantines in 1017. Moreover, the entire perimeter of a production area was uncovered, inside which there was a trapezoidal kiln formed by two parallel walls on a north-south alignment which abutted the citadel’s defensive wall. This workshop had no internal dividing walls. A second walled structure connected the front of the kiln to the west wall and the north side. In the south-west corner, in the space between the west side of the kiln and the wall closing the room, a tile collapse and several post holes for the roof supports were found. It is probable that at least part of the area was covered by tiles and imbrices, perhaps in order to guarantee the conservation of the fuel that was necessary to fire the kiln. The latter had a horizontal draught and the structure showed at least two phases of use. In the last occupation phase, the kiln was fed through an opening on the southern exterior wall of the structure. Inside the structure there may have been a workshop or small storeroom used for the preparation the artifacts for firing. The circular firing chamber was directly connected to a narrow corridor into which the combustible material was inserted. The roof, _in situ_, was formed of tiles bonded with clay and covered on the outside by a layer of unbaked clay. On the kiln’s southern side, in the interior of workshop/storeroom, a blocked opening was found 1 m above the kiln floor and the workshop paving: whether it functioned as an air vent or flue for the kiln, that was later blocked is not clear. The most likely hypothesis for the kiln’s use is that it was conceived for the production of clay, although no waste products relating to this activity were found. West of the workshop several trapezoidal rooms belonging to the residential sector were identified. They were buried below the mound and partially demolished by a ditch. The domestic coarse ware pottery recovered from them dates to the mid 11th century. The eastern sector of the settlement was raised, covering the existing structures and adjacent production area relating to the Byzantine citadel. The tower was probably on this side. On the opposite side a large rectangular area was uncovered. In the same sector another walled structure and several post holes for the wall of a defensive structure were found, providing further evidence that the mound’s structures in the first Norman phase were mainly constructed in wood and earth on a stone footing. In a later phase, datable to the second half of the 12th century, these structures were replaced by stone walls bonded with lime mortar, built on the same alignment. The abandonment seems to have been a slow, gradual process, and from the fragments of proto-maiolica found in the layers preceding the collapse, seems datable to around the 13th century.