In the third sector, under excavation since 2014, we focused on the trench that we had opened in 2015 perpendicularly to the south side of the temple (Trench 11) and where we had exposed the north wall of the south side of the pi-shaped complex (Wall 2558) and a few more architectural remains along the south limit of the trench (Contexts 2596 and 2597). The continuation of the excavation here this year aimed at further exploring these remains as well as at locating the southern (external) wall of the pi-shaped complex that crossed this side of the agora. For this reason we extended the trench by 2 m towards the south. We removed the top layer of this extension (Context 2599), for the most part devoid of artifacts, with a bobcat. We continued manually down to bedrock which appeared at a depth of ca. 1.30 m below ground surface without distinguishing any obvious change in the fabric of the layers. The pottery recovered from the lowest levels (Contexts 3504 and 3505) continued to be mixed, dating from the Hellenistic period to the 7th century CE.
In order to continue the excavation of the space between Wall 2558 and Walls 2596-2597 we had to remove the terracotta pipe (Context 2595) found in 2016 after we had mapped it photogrammetrically. At a depth of about half a meter, the lowest course of the external (southern) wall of the south side of the p-shaped complex (Wall 3500), ca. 0.90 m wide, came to light. This wall is built with rubble and mortar same as the northern (inner) wall (Wall 2558). The distance between the two walls is 6 m, the same that we measured at the excavated section of the northern side of the pi-shaped complex. This space must have been roofed but we found very few tile fragments and no trace of a destruction layer, same as in the section that we had dug close to the northeastern corner of the complex. The pottery found between the two walls (Context 2598) dates as late as the first half of the 5th century CE but with a fair number of Hellenistic and Early Roman sherds in it. The lower layers (Contexts 3506 and 3512) contained mostly Late Hellenistic pottery with a few sherds of the 1st century CE. By removing these layers the two bottom courses of an unknown building came to light near the western edge of the trench, as well as the foundation trench of Wall 3500.
The two courses of limestone ashlars (Context 3509) were uncovered to a length of just 0.84 m and seem to continue westwards beyond the edge of the trench. The bottom course, 1.10 m wide, is made of two parallel and adjoining ashlars, 0.23 m thick. The second course had a single row of ashlars, 0.71 m wide and 0.23 m high. The placement of these courses required digging a foundation trench on the conglomerate. Directly east of the exposed courses we identified a robbing trench (Context 3510), which clearly continues to the east but have excavated only to a length of 1.5 m. The fill of this trench (Context 3507) contained many tile fragments and pottery dating from the mid 1st century BCE to the mid 1st century CE. To this period then we should place the robbing action of the stone courses of this solid foundation of east-west orientation, consequently the building to which it belonged is most likely of pre-Roman date.
The foundation trench of the southern, exterior wall (Wall 3500) of the pi-shaped complex was identified along its north side. Its width comes to 0.32 to 0.37 m but from the soil found in it (Context 3514) we retrieved only a handful of diagnostic sherds of the Hellenistic period.
Our last efforts in this area were directed around the temple, especially along its western side where we took one more pass so that the upper part of the second course of the toichobate is exposed throughout. Most of the pottery found here (Context 3516) is Late Hellenistic but the presence of even a few (less than a dozen) Late Roman sherds indicates that this level too did not remain undisturbed. Three coins were found at this level (inv. no. 32-34) which await their conservation and identification.