Phase IV: Claudian Renovations

During the middle and later years of the first century A.D., several areas of the house underwent modification (fig. 19). The social and historical context of these modifications are discussed elsewhere; here, discussion will be limited to their stratigraphic aspects.

Fig. 19: The House of Diana, 50-60 A.D. (EF).

Little changed in the main part of the house. In shop D, a small and inexplicable hole was dug into the beaten-earth surface in the area of the former cesspit and then filled (cut 175, filled by 176). In the atrium, a low wall made of tile and stones bonded in an earth and mortar mixture and covered with plaster was crudely superimposed on the pavement around the impluvium (28). Additionally, at some point after the Augustan occupation, cubiculum F was divided by a roughly built dry-stone wall, 37, running on a slight diagonal from the southwest wall of the room to about 0.75m. from the northeast wall. Between the wall end and the northeast wall was placed a large, roughly circular stone with a smooth, shallow rounded depression in its upper surface. Both the small size of cubiculum F and the relatively careless masonry of the new wall suggest that the wall served as the foundation of some piece of furniture rather than a true divider, as does a posthole cut in the pavement to its northwest. The stone with the round depression may have had some as-yet-unidentified structural or industrial use. Finally, the pavement of this room was mended with beaten earth during the last phase of its use. This beaten-earth patching (171) contained a fragment of a Republican funerary inscription (part IV, pp. 266-267).

Unlike the main living quarters, the rear of the house underwent substantial changes at some point around the middle of the 1st c. A.D. Most of these changes focused on the garden, which seems to have developed a new function during this period. A new wall (264) was built along the edge of room Q from the corner of 231 to about 1meter past the southeastern column, partially engaging that column and creating a door where there had previously been an open loggia. The new extension was less substantial than 231: it was built of tile fragments set in a matrix of yellow, friable mortar and was significantly thinner than the Augustan wall. Where 231 completely absorbed the pre-existing northwestern column, the southeastern column remained visible on both sides of the Julio-Claudian addition. This wall was probably a pluteus. It was certainly not structural; it seems to have been intended to screen room Q off from the garden, completing the separation of the living rooms of the house from the open space at the rear.

Fig. 20: Axonometric reconstruction of the garden with the shrine to Diana.

In this increasingly independent space, the ornamental garden that seems to have been cultivated during the Augustan period gave way to a very different arrangement (fig. 20; pl. 7 and pl. 58). Plantings seem to have been removed and replaced with bare earth or perhaps a grass lawn. Although we were unable to identify a clear surface associated with this phase of use, the plain earth surface still in use in the final phase of the area seems to indicate the arrangement of the Claudian garden²⁸. Along the northwest wall of the garden, about halfway between the exterior wall of the house and the rear of room K, a large aedicula, about 2 x 4m. and more than 2m. high, was constructed in the same brick/tile and yellow, friable mortar used to close off room Q and to raise the height of the impluvium. The shrine, like the wall closing Q, shows a certain hastiness of construction: most notably, the building seems to lack a proper foundation and now sags visibly. It has been only partially preserved, but it is clear that the structure consisted of a podium with frontal steps and a small cella. Originally, it may have presented a distyle prostyle facade. The exterior was plastered and decorated with a red reticulate pattern on a white ground; inside the cella, marble plaques in the shape of a square and a circle were set into a plaster floor and a trapezoidal platform of earth revetted with marble (241) stood against the back wall. The narrow entry to the cella was equipped with a worked stone doorsill (298) with cuttings that may indicate a locking door. While the walls of the cella (230) were built entirely of courses of tiles and stone in yellow mortar, the podium (299) consisted of a casing of stone, tile and mortar around a core of pebbles set in a sandy mortar matrix. This technique caused the steps and pronaos to crumble more rapidly than the cella, and although the front corners of the podium remained intact, only a single step was preserved. This step may give a further indication of the somewhat expedient nature of the construction of the aedicula: it was a well-worked rectangular marble block, broken jaggedly at one end, that seems clearly to have been cannibalized from some other setting. In front of the aedicula was set the broken base of a column. Although its upper surface was rough and uneven, it may have functioned as an altar or a statue base in association with the shrine.

In the west corner of the garden a substantial stone staircase (296) was built, covering the wall plaster of the earlier phase. Large, roughly worked stones of various sizes and proportions were cemented with the same yellow mortar found in the aedicula and the wall of Q; four steps rose to a square platform from which more steps probably sprang. Oddly, although the walls of the house were preserved to 0.5m. or more above the level of the existing top of the stair, there was no evidence of a door. Nevertheless, it seems obvious that the presence of the stair requires the existence of a door, and the ground level of both the street to the west and the Ïforum piscariumÓ to the northwest allow the possibility of an entrance to the garden via the stair from either of these directions.

[ 1 2 next ]

28. Excavation was complicated by the fact that an apparent surface was reached during the 1996 season, then left exposed for a year; subsequent surfaces may have been the product of weathering rather than ancient activity. The ceramic record for the surface we originally understood to be in phase with the shrine shows that it must be dated to the later life of the area, and it will be discussed in phase V. All of the soil above the Augustan garden surface, however, was strikingly uniform, and there is no evidence that the bare earth or grass surface at the end of the life of the garden differed from the surface in use during the Claudian period.