Iron Age and Persian Period. The excavation of the lower, southern unit in Area D2 (Sqs AO/11–12) added some information about Persian-period and Late Iron Age features. At the lowest point, the excavation reached a kurkar bedding of the Ir2a late pavement, found previously north of this unit. In all likelihood, this floor is associated with the Phase D2/7 ashlar building (the so-called ‘Taphat’s palace’) excavated in previous seasons further to the north. A robber trench running north–south along the eastern edge of the kurkar bedding may be the result of robbing one of the walls associated with this floor, possibly the continuation of W10606. To the earliest Persian remains (Phase D2/5b) belong a pit and additional patches of the red surfaces uncovered in previous seasons, which were found south of the facade wall of the northern insula (W09D2-381).
Hellenistic Period. The balk between Areas D1 and D2 (Sqs AP/11–14) yielded four architectural stages of the Hellenistic and Roman periods. Most of the uncovered features are extensions of features known from previous excavations on either or both sides of the balk. Still, the excavation in this area revealed important relations between the features of Phases 1 and 2 in Areas D1 and D2.
The Hellenistic remains (Phases D2/3–4; Fig. 2) reached in the western part of the southern unit include a facade (W26119) built in the method of interlocked squares (Sharon 1987) that ran along the northern margin of an east–west street, as well as a series of north–south walls. The facade and W26246, perpendicular to it, form the southeastern corner of a room or a building that extends into Area D1 to the west. In the eastern corner of the unit there were another facade wall (W5207) and two north–south walls that might have been connected with it. The southern wall (W5207) of the insula previously excavated in Area D2 continues eastward c. 15 m along the street. Since the Roman street pavement and facade walls were not removed this season, we do not know whether the Hellenistic facades were connected in the same zigzag plan as the Roman ones (see below), or if there was a passage between the western (Area D1) and the eastern (Area D2) buildings. The second option is the more favored at the moment, because W26246 and the two eastern walls are not parallel, thus it is more likely that they belong to two different structures.
The only Hellenistic feature found in the northern part of the area is a segment of W26000, running east–west (the so-called ‘Big Mama’ wall; W17562 in Area D2). This extremely wide compartment wall belongs to the eastern wing of the ‘Monumental Hellenistic Complex’ (MHC), excavated earlier in Areas D1 and D2 (see Area D4).
Roman Period. In the southern square, immediately below topsoil, we found the pavement (L11D2-632) of the Roman street running east–west along the southern slope of the tell (Street 2). At this spot, the facade of the Roman building to the north of the street, partly built on top of the Hellenistic facade, creates a zigzag and continues eastward. The street pavement was adjusted with a wedge-shaped segment to accommodate this strange corner (Fig. 3). Under the street was a stone-built drain, excavated previously in Areas D2 and D1 (L5104).
The northern squares of the balk were heavily disturbed, and our aim here was to remove the Roman remains and prepare the area for the excavation of clean Hellenistic deposits next season. The major features found here relate to the Industrial building (Sharon, Gilboa and Karasik 2006
) and to its southwestern annex: a robber trench of the southern wall of the building (W17591) with a cement floor to its north, and remains of hypocaust tiles and a plaster basin south of the wall.
Persian period. In the lower part of Area D4, under the debris associated with the compartment walls of the MHC (see below), the excavation reached remains of the Persian period structures, namely a set of trenches that mark robbed-out walls, and a sequence of floors in the spaces created by these walls (Fig. 4). Fragments of mud-brick material found on these surfaces and the matrix of the fills right above them suggest intensive use of mud-bricks as building material, chiefly in wall superstructures, as opposed to the overlying Hellenistic architecture, composed only of stones.
Hellenistic period. The removal the balk between Areas D2 and D4 (see Fig. 1) prevented us from digging in the southern part of Area D4. The balk (Sqs AO–AP/15; D4 South), excavated from topsoil, revealed a stratigraphic sequence of Roman and Hellenistic deposits and architectural features.
Both the northern and the southern units revealed several features that proved to be crucial for our understanding of the stratigraphy of the Hellenistic remains (Phases D4/3 and D4/4), both for this specific area and for the entire Hellenistic sequence on southern slope of the tell (Areas D1, D2 and D4). This was especially useful for understanding the relation between the ‘Monumental Hellenistic Structure’, constructed in the compartment-like technique, and the ‘Thin Headers Building’ (Gilboa, Sharon and Shalev 2010).
An ashlar-constructed rectangular feature (W11D4-812) was uncovered in the southern unit (Sqs AN–AO/17) under the foundation level of the Phase D4/2 Roman walls. It is almost identical in size (1.5 × 2.0 m) and shape to two other features (W08D4-363, L09D4-584) found in previous seasons in the central part of the area and identified as piers (Sharon, Gilboa and Shalev 2011
). This similarity suggests that all three belonged to a set of four, or most probably, six piers, that were arranged in two rows within the central space of a building. Several possible relations between the piers and the two Hellenistic buildings were suggested in the past; currently, the most probable one is that they were constructed together with the later set of walls, those of the ‘Thin Headers Building’ (see below).
The puzzling stratigraphic relations between the compartment walls of the ‘Monumental Hellenistic Structure’ and the thin header walls were clarified this season with the following findings (Fig. 4). While deepening and cleaning a narrow space between the northern compartment wall and a thin header wall, two ashlars that dovetailed with the compartment wall were found. These two ashlars belonged to a north–south wall of the ‘Monumental Hellenistic Structure’, which was largely robbed out by the builders of the ‘Thin Headers Building’. The continuation of this compartment wall (W11D4-802) was uncovered in the southern unit, under the eastern thin header wall (W09D4-554). The superposition of these walls leaves no doubts regarding the order in which the two complexes were built: the thin header walls are later than the compartment walls.
The renewed examination of the stratigraphic relations between the walls led to the conclusion that the floor fragments found along the northern compartment wall and the thin header wall were not cut by the header wall, but reached it both within the building and in the space between the walls. Another surface fragment, approximately at the same level, reached the northwestern pier, on top of which a Phase D4/2 wall was built. This evidence, although indirect, supports our suggestion that the piers were part of the ‘Thin Headers Building’, i.e. belonged to Phase D4/3. Another, even more circumstantial clue, is that the southern pier was superimposed by the Phase D4/2 installation built of a secondary-used column drum. It seems likely that the Phase D4/2 builders used an existing feature of the preceding phase. Taking all the above into consideration, we assign the compartment building to Phase D4/4, which represents the early Hellenistic period in Area D (Fig. 5; Nitschke, Martin and Shalev 2011
The southern part of the ‘Monumental Hellenistic Structure’ (the so-called ‘Big Mama’ wall and walls perpendicular to it) were excavated in the 1999–2000 seasons in the northern part of Area D2 (Sharon et al. 2009
), where it belongs to Phase D2/4. A short section of the same wall was cleared this season when excavating the area D2 balk. Several header walls, found in the same units, as well as in the balk, definitely continue the ‘Thin Headers Building’ of Phase D4/3. These two sets of walls clearly establish a correlation between the stratigraphic sequences in these two areas. The Monumental Hellenistic (compartment) Structure continues further west, beyond the D1/D2 section balk, where it is abutted by the eastern wall of another monumental structure (referred to in early publications as the ‘Persian Palace’ [Stern et al. 2000
; Gilboa, Sharon and Shalev 2009
; Sharon, Gilboa and Shalev 2009
; Sharon et al. 2009
], but clearly Hellenistic in date). The two buildings comprise the so-called ‘Monumental Hellenistic Complex’, the focal point in the southern part of the city (Nitschke, Martin and Shalev 2011).
With the question of the sequence of these structures resolved, we could examine their dating and interpretations. Unfortunately, the floors mentioned above were very flimsy and neither carried any primary ceramic assemblage nor were they sealed by debris that could date the construction of the later structure. Preliminary pottery analysis shows that the assemblage associated with the later, ‘Thin Headers Building’, is of an early Hellenistic date (the last quarter of the third century BCE). Nothing definitive can be said about the date of the compartment building, as it has no preserved floors. Constructional fills within this structure go down to the Phase D4/5a floors. They yielded Persian-period assemblages still mixed with some Hellenistic pieces. A terminus post quem for this structure in the second quarter of the third century BCE is provided by a coin of Ptolemy II (284–246 BCE) and a pottery assemblage found in a pit under the so-called ‘Big Mama’ wall.
Among the questions to be answered by future research in this area are as follows: what were the height and the roofing materials (if there were any) of the buildings; why different styles of masonry were used; why are the floors of the ‘Monumental Hellenistic Complex' almost entirely missing; did the Phase D4/3 building utilize some of the walls of its preceding building; why were these architectural programs initiated, and whether they were ever finished; and of course, what functions did these buildings performed, as they were obviously public or palatial in nature.
The uppermost part of the revealed sequence comprised a series of cement floors with some traces of plain mosaic. These floors belong to a large Late Roman building of either a public or industrial character, excavated in Area D4 during the 2004–2008 seasons (Sharon, Gilboa and Karasik 2006
; Gilboa, Sharon and Shalev 2009
; Sharon, Gilboa and Shalev 2009
; Gilboa, Sharon and Shalev 2010
). Several food-processing installations and a drainage system found in this structure indicate that employing cement, plaster and mosaic tesserae
for the floors and, in some cases, for wall makeup and coating as well, served a waterproof function rather than an esthetic one. A column drum in secondary use was embedded in the southern part of the earliest floor—probably an installation of some sort. A southern wall that enclosed either the room or the building (W11D4-751) seems to continue, albeit in a zigzag course, a wall excavated in Area D2 in the 1990s (W17624).
The Iron Age
. As in the previous season (Gilboa, Sharon and Shalev 2014
), only the western and the eastern higher parts of the area were excavated, with the purpose of leveling them with the deep trench that was cut through the area in the 1980s and 1990s. Both subareas revealed a continuation of the Ir2a ‘Courtyard Building’ (Phase D5/9), whose central courtyard was uncovered in Area D5 Center in the 1990s and in recent seasons (Gilboa, Sharon and Shalev 2009
; Gilboa, Sharon and Shalev 2010
). As the area deepened, the excavated squares on the slope of the tell were extended further south, towards the Iron Age city wall.
Over the past years, this segment of the slope has been cleaned during excavation seasons, as has been further exposed by severe winter storms, making several lines of fortification walls traceable. An attempt to map the wall lines (Fig. 6) revealed three superimposed walls, just slightly shifted from one another and separated from each other by earthen deposits. The highest wall is the Phase D5/9 offset-inset city wall (W05D1-548; Ir2a; Fig. 6, red), built mainly of ashlars, which encloses the ‘Courtyard Building’ on the south. The three construction stages of this wall correspond to the changes identified in previous seasons within the building (Gilboa, Sharon and Shalev 2014
). As new parts of the ‘Courtyard Building’ have been uncovered (see below), it has become clear that the stratigraphic relations between the building and the various phases of the ashlar city wall are more complex than previously assumed; we hope to clarify the situation in future seasons.
Running grosso modo
below W05D1-548, at the foot of Area D5, is a rubble wall (W09D5-811; Fig. 6, blue), which was initially attributed to Phase D5/12 (Ir1a; Gilboa, Sharon and Shalev 2011
). The wall has not yet been reached in Area D5, but is visible along the southern margin of the tell. Now that the line of this wall has been mapped, it seems to curve northward, and therefore can be connected by a virtual line to the so-called ‘Bastion’ wall of Area D2 (Fig. 7; Gilboa, Sharon and Shalev 2009
; Sharon, Gilboa and Shalev 2009
; Sharon et al. 2009
), also constructed during Ir1a. Thus, it is quite plausible that wall W09D5-811 and the ‘Bastion’ wall in Area D2 were built concurrently and belong to one fortification system, which was constructed at the inception of the Iron Age town at Dor (the Ir1a horizon), between the late twelfth and late eleventh centuries BCE. Another option is that this fortification wall extended from the ‘Bastion’ wall westward along the ridge above the southern lagoon, south of the offset-inset wall (W05D1-548) and of W09D5-811, and was washed away by the surge; indeed, it is evident that many architectural remains on the southern slope of the tell were damaged by the sea. In this case, W09D5-811 is not a city wall at all. We hope that future excavation in the vicinity of these walls on the acropolis and intensive cleaning of the southwestern slope will enable us to verify or discard this theory. The lowermost wall (Fig. 6, purple) is too unclear to be discussed in detail at this point, but it does indicate that W09D5-811 is not the earliest substantial construction in this part of the tell.
Remains of Phase D5/10 were found this year on the lower terrace of the western part of Area D5. In several spots, the excavators reached superimposed phytolith layers of either an Ir1|2 or Ir2a date, cut by the walls of the ‘Courtyard Building’. The same exact situation was observed in the courtyard of the Phase D5/9 house in the central part of Area D5 (Gilboa, Sharon and Shalev 2014
). These layers signify a period for which no construction is in evidence in Area D5.
The so-called ‘Courtyard Building’, constructed in Ir2a and known to us mainly from its central, large space (the ‘courtyard’), had side rooms east and west of it; on the west the addition might have extended to a complete building (Figs. 8, 9). Prior to this season, only a short segment of a rubble wall (WD510-918) that belongs to the western extension was visible, and its relation to the courtyard building was even doubtful. As the southern part of this wall was found (W11D5-620=W11D5-640), it became clear that it indeed belonged to the long wall delimiting an elongated space (Room 1; c. 2.5 × 10.0 m). The room might have been divided in two (as was the eastern wing; see below), in which case the partition wall would have been disturbed by or buried under the southeastern corner of the Phase D5/8 structure (see below). Two more walls that belonged to the same Phase D5/9 complex (W11D5-665, W11D5-666) ran further west, parallel to WD510-918. These north–south walls, built in an identical manner to those of the ‘Courtyard building’, were found only in the northern part of the area (Sq AX/11); their southern segment disappeared due to erosion on the south slope of the tell. Two additional spaces created by these walls (Rooms 2, 3) differ in their width (c. 3 m and 1 m, correspondingly). The western part of the northern courtyard wall (W10825) extended beyond the northwestern corner of the courtyard. Its stratigraphic relations with the three long walls—essential to our understanding of the plan of the building— are currently unknown. One of the main questions is whether W10825, in the west, is one continuous wall or has a gap, which may indicate an alley that ran between the building and a house to its west. Another option is that the long wall(s) met W10825 and continued even further north; as we know from the central part of the area, the building consists of at least three more rooms to the north of the courtyard (Fig. 8). Currently, the western part of W10825 is under the boulder ‘tower’ (Phase D5/8, see below), and its upper courses were most probably destroyed during the construction of the ‘tower’. If the foundations of the Phase D5/9 building are as deep here as in the central part of the building, then we can hope to reach the lower courses, which will clarify the picture. This building seems to have a more complex plan than hitherto assumed, and we are not sure now whether it belongs to the ‘Canaanite Courtyard’ type or is a regular dwelling; it may be part of a public complex of some kind.
The stratigraphic relations of the western wing of the ‘Courtyard Building’ and the city wall in the south are not entirely clear either. The southern segment of W10D5-918 (W11D5-620) extends beyond the course of the city wall (W05D1-518b) as we know it west of W11D5-620. One possibility is that the western part of the ‘Courtyard Building’ extended further south, cutting this stage of the city wall. Another option is that the city wall turned here southward, and the ‘Courtyard Building’ wall reached a city wall offset, similar to the one found enclosing the southwestern corner of the courtyard.
and plaster floors related to the long walls were found in all three ‘rooms’ and west of W11D5-666 (Fig. 9), yielding Ir2a ceramic assemblages. Correlation between the elevations of these floors and the floors of the courtyard suggests that they belong to the two last stages of the building (Phase D5/9b–c; Gilboa, Sharon and Shalev 2014
In the eastern part of the area, Phase D5/9 became clearer due to cleaning, thus clarifying the stratigraphic relations between features uncovered back in the 1990s. The eastern wall of the Phase D5/9 ‘courtyard’ (W5562) was traced further northward up to the where it would have met the northern wall (W10825); however, the corner was robbed out by a large pit. In the southern squares, the excavators dismantled a pier-like feature (W5554; Phase D5/8), and found that W5562 reaches the ashlar city wall (here W5603) and that an east–west wall (W5563) reaches the ‘Courtyard Building’ wall, forming two additional spaces east of the courtyard. Several surfaces found in these rooms presumably belonged to the latest stage of the building (Phase D5/9a). It is still unclear how far east the building extends, but we suspect that a line of stones in the northeastern corner of the area will be revealed to be another long wall parallel to the eastern wall of the courtyard (W5562).
The excavation of a balk between the central and the eastern parts of Area D5 revealed a set of surfaces within the courtyard, reaching W5562 from the west. These are fragments of the upper courtyard floors (Phase D5/9a) excavated in the 1990s and the scant pottery on them dates to Ir2a.
The southern city wall (W09D5-811=W5603) was traced in the easternmost square of the area (and called W11D5-400). The lower stage of this wall continues the same course of large elongated ashlars as Stage B of the wall in the central and western subareas. From the upper stage (W11D5-326) only a short fragment remained. These two stages are correlated with Phases D5/9b and D5/9a respectively.
A building constructed of boulders and large ashlars in the western subarea was further exposed (Phase D5/8). This square structure was preserved only one (foundation) course high. Its northern wall and the northwestern corner had an ashlar facade. The southern and the western walls were partly robbed out and partly destroyed due to slope erosion. Only the eastern wall of the structure’s southern extension was preserved. The building had a ‘pier’ in the center, suggesting that it was a tower facing north or northwest, with a staircase that surrounded the central ‘pier’; however, it might have been part of a larger structure extending northward, beyond the confines of Area D5. Ceramics from the kurkar surfaces associated with the boulder walls are attributed to the Ir2b–c horizon. Scant remains of a massive boulder structure, uncovered in Area D5 East, have been initially attributed to the same stage. These buildings are correlated with thick white floors excavated in the central part of the area in the 1990s.
Hellenistic and Persian periods. The main goal in excavating the higher parts of this area was to fully unearth the Hellenistic- and Persian-period remains in all the units. However, it was more difficult to correlate between the post-Iron Age architectural remains in the eastern and western parts of the area than it was to correlate between the ealier layers. Numerous pits that mark the transitional Persian–Hellenistic layers were excavated in both parts of the area (Phase D5/5 in the west and D5/5c–6a in the east). The stratigraphic relations of the Persian features excavated in previous seasons (Phase D5/6) were sorted out, and a much clearer plan of the area during the earlier stages of Phase D5/6 was drawn.