The twenty second season of excavations was conducted in July 2004 at Tel Dor (the second season of the renewed expedition to the site; License No. G-56/2004). The excavation, on behalf of the Institute of Archaeology of the Hebrew University, the Zinman Institute of Archaeology of the University of Haifa and the Israel Exploration Society, was directed by I. Sharon and A. Gilboa, with the participation of c. 50 students and volunteers from abroad, 15 students from the University of Haifa and 30 staff members. Participating institutions were the University of California at Berkeley, headed by A. Stewart; University of Washington at Seattle, headed by S. Stroup; the Kimmel Center for Archaeological Science at the Weizmann Institute of Science, headed by S. Weiner, the University of South Africa headed by W. Boshoff and the Center for Nautical and Regional Archaeology, Dor, headed by I. Hirshberg, where the offices of the expedition are located. The excavation’s field staff included J. Berg (senior field archaeologist), A. Estes (Area D1 supervision), E. Bloch-Smith of Villanova University and A. Brody of the Pacific School of Religion in Berkeley (sub-area supervision in Area D2), S. Matskevich (architectural drafting and electronic infrastructure), R. Lin (conservation), S.R. Martin (sub-area supervision in Area D1 and in charge of late pottery), A. Karasik (Area G supervision), Y. Shalev, J. Nager, E. Dintino, J. Chesley, M. Thornton and R. Karberg (sub-area supervision in Area D1). Experts and support staff that actively participated in the excavation included U. Smilansky and H. Mara of the Technical University of Vienna (computerized typology), A. Killebrew of Pennsylvania State University (cultural heritage), E. Boaretto and E. Mintz (14C dating), R. Shachak-Gross (micromorphology), T. Goldman (electronic infrastructure and computerized graphics), R. M. Alberts (phytolith analyses), F. Berna and A. Behar (sediment chemistry), S. Silshtein and N. Yahalom (archaeometallurgy), N. Raban (archaeozoology), N. Kruger and E. Cruywagen (recorders in Area D2), A. Haiem (administration), T. Ortballs (photography) and V. Rosen (find drawing). The expedition was lodging at the guest house of Qibbutz Nahsholim.
A practical course on architectural conservation was conducted, for the first time, within the framework of the excavation, on behalf of the University of Haifa.
Area G – The Early Iron Age
During the 2002 season an expedition of the Weizman Institute of Science began working at the site. Its aim was to examine the feasibility of large-scale sediment sampling and its on-line analysis during the excavation, which would facilitate an immediate feedback between the mineralogical examination results and the excavation process. In 2002–2003, the samples were mainly taken from existing soil exposures, both vertical (balks) and horizontal. This year, a single excavation square in Area G was selected for a slow-moving excavation, whose main purpose was the implementation of the mineralogical analyses.
The excavation was focused in one room, on the western side of the courtyard building in Area G, which was the principal architectural complex in the area during the Early Iron Age. The excavation exposed the white layer that had previously been discovered in this room at the end of the last season (Fig. 1). This layer, which, at first glance, was considered a white lime/chalk floor, was determined to be a horizontal cluster of phytolith layers (the mineral skeleton of plants, composed mainly of the mineral opal) that ranged in thickness from several centimeters to fractions of a millimeter (total thickness of the accumulation c. 20 cm). The phytoliths originated in different species of seasonal grasses and contained also pytoliths of the bloom, indicating that the grasses were harvested in the spring time. The original volume of the floral matter in the room could be calculated based on experiments with modern floral matter. It turns out that following the disappearance of the organic material from the floral matter, the volume of the remaining (mainly inorganic) fraction is c. 3% of the original volume. Therefore, a layer of a few centimeters of phytoliths represents an accumulation of floral matter c. 1 m high. Hence, the current horizontal phytolith surfaces were large heaps of floral matter in the past. The ‘chalk floor’ then was not a floor at all, but rather an accumulation of floral matter (on a habitation level of soil), evidenced by the flagstones and the pottery vessels (including an intact bowl) that were found between and inside the phytolith horizons. Consequently, it seems that the numerous ‘thin white’ floors at Dor and at other sites are not floors but the remains of floral matter accumulations.
In one of the corners of the room, above the layers of phytoliths, another horizontal accumulation of phytoliths was discerned. A microscopic examination revealed that the phytoliths of this layer originated only from barley, which was apparently bound together in small bales.
There are several possible explanations for the phytolith phenomenon and additional tests are required to choose between them. It is possible that this room was used for storage of animal feed. In this case, each layer may represent a single seasonal harvest. Alternatively, the room may have served as an animal pen and the pytoliths were derived from the accumulated dung. A third possibility is that the room was used for storing dung for construction purposes (plastering for example), or for other uses (such as fuel)––all of which are well-known from ethnographic studies. The upper phytoliths layers may have originated from a collapse of a thatched roof over the contents of the room, whatever it was.
A fieldstone-paved floor was exposed below the phytolith accumulations (Fig. 1). Stone pavements in the Early Iron Age at Dor were a rare phenomenon (including in the large public buildings). The only known pavement, so far, was in the courtyard of the courtyard building in Area G, implying that the pavement below the phytoliths was also an unroofed space. Surprisingly, another layer of phytoliths was discovered in a probe below the pavement.
Based on the stratigraphy and the pottery vessels, the accumulation of phytoliths should be ascribed to Phase 9 in this area (the Iron Age 1a [late] according to the Dor terminology). This phase ended in a destruction, which was evidenced in this room by a few pottery vessels discovered in situ (fierce conflagration was also evident in the southern parts of the area). The stone pavement abutted the walls of Phase 10 (the Iron Age 1a [early]).
While sifting the accumulation above the lower phytolith layer dozens of tiny disc-like beads (diam. c. 1 mm) were discovered. Some beads were made of metal and others––of a silicate material (not yet identified). Stone beads (c. 3 mm in size) were also found.
The large quantity of phytoliths (here and in other areas at Dor) is being used, among other things, for a new 14C dating technique that was developed in the Weizmann Institute.
Preparing the area for excavation, above these remains, revealed a small section of a floor from Phase 6 (Iron Age IIA), which was part of a room that had previously been excavated and was the principal complex of this phase in Area G. Among the finds discovered was an in situ jar embedded in the floor, a ‘Black on Red’ Cypriot vessel and a basalt grinding stone.
Area D2 – The Late Iron Age and the Persian-Hellenistic Periods (Fig. 2)
The excavation was focused in five squares north and west of the area’s center that had been excavated in the past (down to bedrock). The goal in excavating these squares was to lower down and terrace the high balks along the edges of the main area, as well as to examine issues relevant to the late Iron Age and the transition to the Persian period.
The Late Iron Age
The excavation headed by E. Stern in the western center and deep part of Area D2 exposed the outer corner of an ashlar structure (stone dimensions 1 × 0.5 × 0.5 m), preserved only two courses high, which were probably the foundation courses. To its east, above the Iron Age IIA stratum (Phase 8b) was a large white floor (Phase 7), severed by numerous pits (Phase 6). Only a few potsherds, probably dating to the eighth century BCE, were discovered on the floor and at least some of the pits were ascribed to the seventh century BCE. Yet, the chronology of the ashlar building and the large floor was not clarified and the stratigraphic relationship between them was not ascertained. Squares were opened both inside and outside the ashlar building in this season. Several kurkar floors were uncovered, some of them above stone pavements. The relation between the floors and the ashlar building had not been definitively established. The floors may postdate the building, since in two points it seems they overlay the ashlars. An installation that contained a large quantity of carbonized seeds was also found. Insufficient datable finds were discovered on the floors, although the few pottery fragments recovered from unsealed loci seem to belong to Iron Age IIA/B. This matter and the further exposure of the ashlar building will be pursued in the coming seasons.
The aforementioned stone pavement was cut by a pit, which contained a jar, in situ, dating to the seventh century BCE. This pit was one of a large series of pits from the eighth and seventh centuries BCE that were probably the remains of an industrial area. The largest pit in this series was enormous (diam. c. 4 m, depth in excess of 3 m) and its northern part was excavated this season, revealing, as in previous seasons, very large quantities of pottery vessels dating to the seventh century BCE, mainly Phoenician commercial jars and jugs and a few Assyrian-style vessels. Other artifacts in the pit included slag that was probably copper, nodules of vitrified quartz and other industrial waste, as well as an antler and worked ivory.
The activity in the industrial zone seems to have ceased in about the middle of the seventh century BCE or somewhat thereafter. Since the transition between the Iron Age and the Persian period at Dor is a debated issue, an attempt to define the nature of the transition was undertaken here. Some time after the industrial area was no longer in use, a residential insula was built (Phase 5, see below). It is not clear, however, when the function of the area changed. Mixed pottery from the end of the Iron Age and the Persian period occurred in several loci, but it was impossible to specifically identify a transition phase, as none of the loci were sealed.
The Persian and Hellenistic Periods
Many robber and foundation trenches of Phase 5 from the Persian period, excavated this season, enabled to corroborate the basic stratigraphy of three sub-phases (5a–5c). Several wall remains indicated that an orthogonally planned city was established on the tell during Phase 5. Based on the robber trenches of Phase 5 walls we could conclude that some walls were erected utilizing the Phoenician building technique of alternating segments of ashlar and rubble construction.
One of the most important finds, discovered sealed below one of the Phase 5 floors, was a silver Athenian tetradrachm that probably dates to the middle of the fifth century BCE. It is unclear if this was the earliest floor of Phase 5, but the coin undoubtedly belonged to a relatively early phase in the stratigraphic continuum of the Persian period at Dor. This further validates the impression that the Persian settlement at Dor developed relatively late (middle of the fifth century BCE?) and the apparent hiatus between the Iron Age and the Persian period is thus expanded to c. 200 years.
The dismantling of one of Phase 4b massive ashlar walls continued this season. This wall was stratigraphically connected to Phase 4 in Area D1. A public building, formerly referred to as the ‘Persian Palace’ (see below), was built in Area D1in this phase. It was suspected some time ago that the ‘palace’ was not Persian, but rather Hellenistic. The dismantling of the wall in Area D2 revealed Hellenistic potsherds that corroborate this.
Area D1 – The Hellenistic and Roman Periods
This area was expanded eastward and northward, with nine new excavation squares that were intended to connect Areas D1 and D2. Archaeological soundings were continued in the squares that were opened in the past.
The Hellenistic Period
The clarification of the stratigraphic relationship between the thick, east–west ashlar-built wall from Phase 4b in Area D2 (above), and the large Hellenistic public building in Area D1 (the so-called ‘Persian Palace’), was continued. It became clear beyond the shadow of a doubt that the northeastern corner of the large building abutted the ashlar wall. Therefore, the ‘Palace’ postdated that wall, at least in terms of construction (Fig. 3). Not only is the ‘Palace’ not Persian, it is not even particularly early within the Hellenistic sequence.
The southwestern corner of a large Roman building (below) was uncovered above the interface between the eastern part of the ‘Persian palace’ and the western part of the large ashlar wall. Most of its walls were plundered, although its boundaries were still obvious, which made it possible to discern small service rooms built on its exterior (Fig. 4).
The Roman Period
A large building from the Late Roman period was discovered just below surface. Its walls were extremely thick, but in a poor state of preservation. Part of the structure was paved with thick cement floors. A row of small cubicles (1 × 1 m) was in the western part of the building and to their east––three elongated spaces oriented east–west (Fig. 5). The southern space was paved with large white tesserae, into which a large round stone that could be a column base, or a yam from an oil press in secondary use, was incorporated. The middle space was paved with cement and in the pavement’s foundation was a plastered conduit that connected it to the southern space. The northern space was corridor-like, had a drainage channel and led to another room in the east that was not excavated. To its north was another row of cubicles that was hardly excavated.
Several places in the building showed evidence of two construction phases (1, 2). The earlier phase seems to have been built of ashlars, whereas the walls added or repaired in the second phase consisted of poured cement. Remains of ceramic pipes and square hypocaust tiles were discovered in the southern area. One tile was found in situ and tile impressions were discerned on the floor.
Two possibilities with regard to the purpose of the building were proposed. The first was a bathhouse, based on the tiles (the small rooms were perhaps furnaces and water tanks). Yet, the coarse mosaic (devoid of any ornamentation) was incongruent with this kind of building. The second was an industrial building that processed a liquid, which required a drainage system (wine?). However, the ceramic pipes beneath the floors were not in keeping with such an interpretation.
Attempts were made in the southern part of the area to locate other parts of the spectacular ‘mask and garland’ mosaic that was revealed in the 2000 season. This is one of the most magnificent Hellenistic mosaics ever discovered in the region and it was important to narrow down its stratigraphic provenance. Mosaic parts hitherto found were thrown into a refuse dump. The northern part of the dump was excavated this season and yielded numerous architectural elements, including fragments of frescos and pieces of a white mosaic that probably belonged to the undecorated sections of the same floor; no other colored sections were retrieved. The large Roman building postdated the refuse pit with the mosaic, as small mosaic fragments were found sealed beneath one of the building’s floors.
Conservation work was mainly carried out in the areas that had previously been excavated by the Stern expedition.
Area D2. Mechanical equipment was used to dismantle the two massive walls of the Late Roman period that remained ‘hanging’ over the early Iron Age remains.
Area G. This deep centered area of the tell was prepared and covered at the end of the season, except for a small section that was excavated.
Area H. Work on the Roman villa east of the Roman temenos had begun, including cleaning, leveling the ground, reinforcing and stabilizing walls, as well as documentation.
Simultaneously, infrastructure work that mainly involved the preparation of paths and steps on several of the approach routes to the tell, was undertaken.