The structure B/3/2 is located a short distance from the Painted Residence (B/3/1), at the north-west corner of an open court that links these two structures with others on the west. To the southwest of B/3/2 (due west of B/3/1) is yet another structure similar in plan to B/3/2. The overall dimensions of the four-roomed building are 11.2m x 8.95m. It actually comprises two separate structures adjacent to one another, each consisting of two rooms on a north-south axis. The eastern structure (8.95m x 6m) is built from reddish mud brick and has rounded corners; the western structure (8.95m x 5.2m), is in greyish mud brick, and was built against the eastern one. Both were originally barrel-vaulted with an upper storey but the vaults were clearly breached in the past and the two northern vaults did not survive particularly well. During excavations it was necessary to remove portions of these vaults in order to avoid their collapse. Some external clearance was undertaken and the mass amount of pigeon nesting pots collected from the rubble collapse around structure B/3/2 permits identification of it as a columbarium or dovecote. The presence within this rubble of fragments of mud-plaster still attached to brick that preserved the impression of such pots provided clear evidence that the structure of the upper storey incorporated these vessels.
Room 3 (2.8m x 2.6m) and Room 4 (2.8m x 2.7m) were excavated during 2006. They are connected by a door at the eastern end of their common wall. Both were cleared down to floor level. Room 3 had been filled in by windblown sand and structural collapse. At levels 30-40cm above the original floor, evidence of stabling was found in conjunction with a secondary partition wall that divided the room in half and remains of palm roofing which had covered over the western part of the room. The vault of Room 4 remained relatively intact except for small breaches. It was found to be filled with predominantly windblown sand. Stable use also became apparent approximately 30-40cm above the original floor, as was the case in Room 3. At these occupation levels varied artefacts were encountered including several baskets, pieces of leather and a few animal straps/harnesses. Diagnostic ceramics from these levels indicate a 2nd to 3rd century CE date. 37 inscribed jar dockets, either still set within mud-jar seals or detached from their sealings, were also found, especially below the stable level in Room 4. The orthography of the writing indicates a date within the 2nd or early 3rd century for this material.
Externally, an almost complete stairway abutting over half the length of the west wall was revealed, as well as an unusual floored surface situated above straw layers extending westward well above the surface (i.e., basal clay) upon which the structure was built. The stairway met ground level in line with a mastaba along the front (north) of the building and would appear to have provided access to the upper storey, from which much of the surrounding rubble likely derived. The mastaba continued along the west of the columbarium, behind the stairway, where ground level was also reached. It is probable that the walls contained cavities into which the nesting pots found in large quantities around the building were once set. It was likely open to the sky to enable easy access for the birds, which would have provided a convenient source of food and fertiliser. Similar structures are found in various parts of the oasis, especially in the western half. It is of interest to note that the only two dovecotes noted at Ismant el-Kharab are in close proximity to the complex of painted buildings of which B/3/1 is part.
During the 2007 field season, Rooms 1 and 2 on the east of the building were examined. The dimensions of the two rooms appear identical at 3m x 3m and both incorporated barrel vaulting (the vault of Room 2 is 90% intact). Room 1 contained much windblown sand that was deposited on rubble from structural collapse and this covered the entire room for over one meter depth. Below the rubble in Room 1, stable matter, indicating the presence of donkey and camel, and flooring was revealed. Within it, pockets of ash as well as a pocket of yellow-red ochre mixed in with the stable matter were also encountered. This particular room fill at these levels was artefact-rich, containing much ceramic, complete and semi-complete pots, and items of glass, faience, textile, worked wood, flora, fauna, leather, basketry and footwear. Importantly, 19 fragments of Greek papyri, 6 Greek ostraka, 1 jar sealing with docket, a child's silver ring decorated with an image of Nefertum, and a wellpreserved mud sealing with and impression representing Herakles and the Nemean lion were also discovered. Diagnostic ceramics included 2nd-3rd century and 4th century material. Below this deposit, similar secondary flooring occurred and a further 8 fragments of Greek papyri and three Greek ostraka were found amongst items of glass, faience, worked wood, leather, textile, basketry, flora and fauna. Underneath these deposits, the original mud-plastered floor remained in very good condition across the room. A test cut revealed it to be approximately 4cm thick, laid directly on the basal clay surface of the site.
Excavation of Room 2 could not begin until a section of the wall and doorway between both rooms was secured and rebuilt. Once this was completed work continued and Room 2 was found to contain windblown sand for approximately 2m. Below the sand, stable flooring was encountered as well as a significant coverage of ash and soot indicating that fires had been set. These deposits, occurring 25-30cm above the remnants of the floor level, contained ceramic, glass, faience, textile, fauna, leather and basketry artefacts. Diagnostic ceramics indicate a 2nd-3rd century date. Only one Greek ostrakon was found in the room, yet surprisingly, three deposits of a yellow-green mineral (possibly jarosite) were encountered within the stabling deposit. The traces of original flooring found in conjunction with the wall foundations revealed that the height of the room reaches 3.3m at the vault peak.
An examination of the foundations in both rooms indicated that the walls sat directly upon the basal clay surface of the site in a few places, but in others clearly were constructed upon foundation cuts and foundation coursing. A mastaba was found along the north outer face of the structure. This appears to be consistent with the previous season's find around the western side. The external rubble here appeared to cease at a point level with the threshold and doorway into Room 1. Below this point hard compacted layering mixed with clay occurred. The composition of the doorway itself (i.e., door jambs and plastering) suggests that it underwent secondary modification. It is also evident that the outer east wall is a separate construction which butts the south and inner east walls. Moreover, it consists of predominately grey brick as opposed to the red brick used throughout the south, west and north walls.
An extensive list of publications relating to Ismant el-Kharab is available for consultation.