This season the work concentrated on five units (OPB 9,10,11,13, and 14) and the cataloging of stored materials that remain from previous excavations. We cataloged some 600 amphora, and the boxes also yielded fresco fragments, plaster casts, part of an inscription, and_tegulae_ _mammatae .

OPB 9
The primary aim of this unit was to understand the entrance on the eastern side of the peristyle. We recovered two pavement levels, a possible mechanism for a gate closing the peristyle, and wheel ruts resulting from cart traffic. The wheel ruts were so distinct that their presence almost seems purposeful. Inside them we distinguished a thick incrustation of crushed pottery; it may represent a re-paving or may be the result of usage and accumulation.

OPB 10
The aim of this unit was to determine the presence of a well in the north western corner of the peristyle. Here the foundations of the colonnade were deliberately cut to follow the wellhead. During the final phase of the building the well was clearly filled in as the structure went out of use perhaps in conjunction with the construction of the water system. Its depth and the confined space impeded us from reaching the bottom. The well did not preserve any sort of lining suggesting that it was robbed out in antiquity or decayed.

OPB 11
This unit investigated the western side of the peristyle. Room 22 revealed a shallow floor (B11003) covering a drain (B11011, B11012) that had collapsed and was abandoned in antiquity. The drain ran north-south and likely connected to the one running east-west which we previously recovered in units OPB 2 (B02010) and 3 (B03014) on the south side of the peristyle. Room 22 displays a walled up entrance on the western end suggesting that it was reorganized after the drain collapsed. In the southern corner of the unit excavations recovered the remains of the drain that ran east-west through the peristyle (B11028, B03014, B02010) and a small settling tank that was completely filled with amphora fragments. Its bottom was broken through, suggesting that it was deliberately put out of use in antiquity. We also discovered the remains of two pilasters (B11005) that once supported the second floor.

OPB 13 and 14
These units recorded the full extent of the holes in the floor of rooms 26 (unit 13) and 41 (unit 14) in antiquity. The clean-up revealed an abandoned cistern at least 2.5 meters high x 1.5 wide and some 15 meters long. Unit 14 preserved the eastern edge of the cistern which contained with a broken amphora neck once channeled water into the cistern (B14036). To the south are the remains of a small wellhead (B14004) which allowed water to be drawn from the cistern. The wall separating spaces 17 and 41 (B14007) truncates the structure indicating that the cistern was abandoned well before the eruption.

Conclusion
The sum of the evidence points to at least three distinct phases associated with the complex. The storage rooms on the southern side of the building are a late addition. They truncate a cistern that likely functioned together with the well in unit OPB 10. The well and the cistern probably ceased to function with the introduction of the water channels recovered in units OPB 2, 3, 9, and 11. In a further phase part of the water system went out of use. In the courtyard we recovered at least two pavements and a thick incrustation stratum which suggests long period of accumulation associated with the final phase of the building.