An account of the 23rd season of excavations
- The Expedition: Composition, Strategy, Organization
- Iron Age I - IIA
- Iron Age II - Area D2
- The Persian Period - Lower D1 and D2
- West Sector of Upper D1 - Hellenistic and Roman Remains
- D4 - Roman Period
- Upper D1 - Eastern Sector - Roman Period
- A note about the Roman Stratigraphy at Dor
- A Note about Pyrotechnic Installations
- On-Site Analyses
- Conservation and Site - Presentation
- Automated Registration
The Expedition: Composition, Strategy, Organization
The staff assembled at the Nahsholim seaside resort on June 26th for preparation and orientation, and volunteers arrived on the 28th. The 29th was the volunteer orientation day. Thursday and Friday - June 30 and July 1 - were the first days of full work, devoted mainly to weeding and excavating the 'post plastic age' - down to last years' plastic covers. Excavation began in earnest the following week, though in one case - the low part of D1, which had not been excavated since the mid-90's - cleaning operations took a full week. Excavation continued, with a short pause for a three-day-weekend and a changeover of some of the volunteers in mid-season, and started winding down on August 1. End-of-season photos were taken on August 3-4 (including balloon-photographs, taken by "Sky Balloon inc."), and the fields covered and closed on the 5th. Staff finished their office work on the 8th and left on the 9th of August.
The excavation was sponsored by the Israel Exploration Society and five academic institutions, including three Israeli universities - The Hebrew University of Jerusalem (henceforward: HUJ), Haifa University, and the Weizmann Institute of Science (WIS) and two American ones - The University of California, Berkeley (UCB) and The University of Washington, Seattle (UW).
The season was supported by endowments, grants and donations from the Berman Institute of Biblical Archaeology, the Kimmel Center for Archaeological Science, the Zinman Institute of Archaeology, the University of California (in the form of travel grants and student scholarships), the Stella and Charles Guttman Foundation, the Roswell family of Baltimore and an anonymous donor. Additional activities that took place during the season (though not excavation itself) were funded by the Israel Science Foundation (collection of new samples for 14C dating, grant no. 778/00), the Bikura program of the Israel Science Foundation (computerized typology, grant no. 4014/03-50.0); the White-Levy program for Archaeological Publication (publication of the Iron Age strata of Area B) and the Getty Foundation's collaborative research program (J. Paul Getty Trust; no. 016844).
The expedition was lodged at the Nahsholim Seaside Resort and working facilities were extended by the CONRAD (Ha-Mizgaga) museum at Kibbuz Nahsholim and its director, Israel Hirshberg.
The main bulk of the volunteer force was the united UC-UW team (50 on the first half - 45 on the second). To this were added 15 Haifa University study-excavation students (first two weeks); a group of 15 Italian students from the University of Florence for the second two-weeks, and a number of 'individual' volunteers - including returnees from previous seasons.
Three field-schools were held on-site: an English-speaking field school for UC-UW students (accredited by the UC summer school program), a Haifa University study excavation course and a graduate conservation course (see below).
Excavation was concentrated in four sub-areas of area D, on the south-west corner of the mound (map). This choice was prompted by several strategic considerations:
- We wish to achieve a large exposure of all strata, rather than spread our effort on various small areas around the tell.
- Previous research indicates that the initial settlement of the site in the Bronze Age - hardly tapped hitherto at all - may have concentrated on the kurkar ridge at the western third of the mound.
- Long-term conservation policy: Our biggest conservation problems are caused by the high vertical balks that are the inevitable results of deep probes into the Iron Age (and perforce Bronze Age) strata; as well as the lack of proper drainage for rainwater in the bottom of such. An 'amphitheatre-like' digging design - where deep early areas are gradually stepped by rings of later exposures and the whole is opened towards the sea is a solution.
- Site presentation - and especially the explication of stratigraphy to the uninitiated - would also be enhanced by a gradual ascent from the earliest exposures by the beach through successive later strata onto the top of the tell.
Thus the basic excavation design over the last few seasons has been to open new areas that will connect area H and the three D's of the previous excavations into a single large field, leaving the deepest parts of area D2 as they are, excavating a ring of units around them down to Iron Age II or late Iron Age I (henceforward 'upper D2'), to open further units north of D2 and D1 (D4 and 'upper D1') which will be excavated down to Persian / Hellenistic levels, while a strip north of that will be left at the Roman levels - as will area H. A good candidate, under this design, for a Bronze-Age area may be between Garstang's southern trench, west of D2, and the Crusader moat west of D1. To check out this option we re-activated 'lower D1' - unexcavated since Stern's expedition left it in 19971.
The Dor consortium is organized in three intersecting structures: An executive structure - consisting of the directors of the participating groups and institutions, a hierarchic operational structure of unit, area, and field supervision plus technical support staff, and forums consisting of staff members pursuing a joint research topic or professional project.
Ilan Sharon and Ayelet Gilboa (HUJ and Haifa U. respectively) directed the project. Gilboa was in charge of the two 'early' areas - D2 and 'lower D1'. Area D2 was supervised by Elisabeth Bloch-Smith (St. Joseph's University) assisted by Talia Goldman (HUJ/WIS) and Amir Haeim (Haifa U.), with Barak Givon (HUJ) as recorder. Yiftah Shalev (HUJ) supervised lower D1, with Dana DiPietro (UCB) and Brendan Haug (UCB) as recorders. A special project of high-resolution excavation of (mainly) industrial waste-pits in D2 (see below) was supervised by Adi Behar (WIS). Andrew Stewart (UCB group director), Sarah C. Stroup (UW group director) and Allen Estes (UCB co-director) were in charge of the 'late' areas - D4 and upper D1. Upper D1 was supervised by S. Rebecca Martin (UCB), assisted by Rebecca Karberg (UCB), Lauren Cannon (UW) and Mont Allen (UCB) and Emily J. Haug (UCB) as recorder. Area D4 was supervised by Martin Wells (University of Minnesota) with Erin Dintino, Nathan Arrington and Jeff Pearson (all UCB) and David DeVore (UCB) as recorder.
The site-formation and archaeo-materials studies of the Weizmann team were
again a main focus of investigation, the forum was chaired by Steve Weiner
and Elisabetta Boaretto with Ruth Shachak-Gross (micro-morphology),
Francesco Berna and Adi Behar (pyro-technology and burnt sediments), Sana
Silstein (XRF and metallurgy), Rosa-Maria Albert (ICREA / University of
Barcelona) - phytolith analysis.
Professional forums active during this season included:
- the Iron Age Ceramics & small finds 'readings' - of D2 and lower D1 staff with Ayelet Gilboa; 'late period s' (Classical - Roman) ceramics and small finds 'readings' - Andrew Stewart and S. Rebecca Martin with D4 and upper D1 staff.
- The site-formation and archaeo-materials studies of the Weizmann team were again a main focus of investigation, the forum was chaired by Steve Weiner, with Elisabetta Boaretto (analysis of charcoal and radiocarbon dating) Ruth Shachak-Gross (micro-morphology), Francesco Berna and Adi Behar (pyro-technology and burnt sediments), Sana Silstein (XRF and metallurgy), Rosa-Maria Albert (ICREA / University of Barcelona) - phytolith analysis.
- The archaeo-zoology forum was particularly active this season, and new sifting, picking, and recording were experimented-with. It consisted of Guy Bar-Oz (chair), Noa Raban (macro-fauna) and Irit Zohar (fish) - all from Haifa U.
- The newly-installed computerized registration and graphics forum was run by Svetlana Matskevich (HU / Heidelberg University), Talia Goldman and Yiftah Shalev aided by Bob Roswell.
- Three publication forums were working concurrently with the excavation - for areas B1 (Elisabeth Bloch-Smith, Ilan Sharon, Ayelet Gilboa, Svetlana Matskevich), D2 (Ilan Sharon, Yiftah Shalev, Talia Goldman, S. Rebecca Martin) and D1 (Andrew Stewart, Allen Estes, Sarah C. Stroup, S. Rebecca Martin).
General staff included Sveta Matskevich (architectural drafting), Howard Karesh (photography), Ravit Linn (architectural conservation), Ruth Gross (artifact registration and conservation), Vera Damov of Haifa U. (artifact drafting), Israel Hirshberg (museum director) and Itamar Mitler of Haifa U. (administration).
Remains of these periods were excavated this season in 'lower D1'. The excavated area consisted of a large (at least 10 x 5 m. ) courtyard, surrounded by 90 cm. thick walls made of small fieldstones. North and west of these walls were several other rooms. However, the narrow space between the courtyard walls and the baulks prevented the excavation of these rooms (the areas east and south of the courtyard are as yet unexcavated). The whole arrangement - rooms around a [presumably] square central court - is reminiscent of the area G house excavated in 1988 - 2004. Another feature shared with the latter structure is the longevity of the structure. The lower D1 house was used at least from the later part of the Iron Age I (our 'Ir1b horizon') to the beginning of the Iron age IIA (our 'Ir1|2' or 'Ir2a'), with a possible localized destruction during this range (see below). The walls around the courtyard, as well as some of the floors and features in it, were already recognized by the University of Saskatchewan team which excavated in area D1 in the '90's ( Excavations and Surveys in Israel 14: 61-71). However, the extremely uneven surfaces (see below), as well as several Persian pits which were not recognized (below, under 'Persian and Hellenistic Periods') hindered the isolation of clean assemblages.
The earliest phase thus far encountered is in the southern part of the area, adjacent to the southern wall of the courtyard. In the southwest corner there was a large patch of bright red debris (it is not clear at this point if this is burnt mud-brick-material or hamra soil) and within it a dark, ash-like area. In these matrices were found the crushed remains of a 'wavy band' Cypriot-style pithos with a stylized palm design on its neck.
Above that was an accumulation of many thin phytolith surfaces, containing typical Ir1b ceramics.
The later phase of the courtyard (the latest phase excavated this season) consists also of several surfaces - typically crushed kurkar makeup with phytolith layers on top - which were very uneven - separated by as much as 15 cm. from each other at some points and compacted to a single surface at others (notably at the southern edge of the area). The topography of all of these surfaces, however, is similar: they all slope down from the walls towards the center - making a bathtub-like hollow in the center of the courtyard. Bisecting the northern half of the courtyard is a trough-like installation with a sharp vertical V section, sloping down like a playground-slide into the central hollow. This may have been some sort of installation for gathering and perhaps processing dry-goods of some kind. A clue as to its function may be several dozen olive pits which were gathered from one of the resurfacings of the slide-like installation. Alternatively, the steep slopes of these surfaces may be a result of some (yet unexplained) post-depositional and post-use process. In the corner of the courtyard and associated with some of the same floors was a small tabun. The ceramics from these surfaces seem to reflect an 'Ir1|2' going on 'Ir2a' chronological horizon.
One unusual find was a burnt wooden object. Found under the lowest of the floors relating to the 'slide installation' it is unclear at this point whether it relates to one of the resurfacings of the latter, or to the earlier, Ir1b horizon. A pile of in situ pottery directly under the object, not yet excavated, will provide the answer next season. When first found, it was thought to be a thick wooden beam. On excavation, however, it appeared to be constructed of at least two layers - in which the planks were laid cross-grained. It is thus either a piece of furniture, or a section of roofing material. Like the crushed pithos it may represent a partial destruction episode within the life of the structure, after which the courtyard was restored with nothing more than a raising of floor level.
The key to understanding the stratigraphy of the area is a set of two perpendicular header walls, built of huge kurkar blocks measuring 100 x 55 x 35 cm. One (W10606) was found already in 1989 in the northwest corner of 'old D2' and the other (W04D2-065) was first excavated last year. These two presumably form two of the walls of a large Iron Age II public structure.
When first excavated, it was observed that W10606 covered the walls of 'Benny's building' of phase D2/8. It was thus dubbed 'phase 7' together with a set of crushed-kurkar floors which also covered the same building (but its relation to the wall was missing). Since the latest floors relating to 'Benny's building' (phase D2/8a-b) were dated Ir2a, the header structure has to be later than that.
A set of several superimposed thick crushed kurkar floors (some with sections of pavements on them) was found both south of W04D2-065 - where at least two superimposed surfaces interspersed by fills were excavated - and north of it - where at least three surfaces are recorded thus far (the first having already been exposed last season). The relationship between these floors and W04D2-065 is problematic. Later pitting activities (see below) severed most of the connections. The rough construction of W04D2-065 (and the fact that its extant preservation comprises only two to three courses) puts in mind a foundation rather than a superstructure - and the persistence of small cobbles along the wall may indicate a foundation trench. However, there were several points of actual contact between the crushed kurkar and the wall. Also, assuming that the wall is later than the floors would allow the building it belonged to a uncomfortably narrow chronological interval (see below). To the best of our judgment at this point, then, the walls and floors co-exist.
The floors were found almost denuded of primary artifacts. The fills between them contain mainly pottery of Ir2a, with a few pieces that may be more recent - Ir2a|b or Ir2b. Several Cypriot sherds in these fills date to CGIII. As we know that the ashlar building overlies structures whose end date to Ir2a, its construction must be dated to the latter part of that period at the earliest. That earliest material in the pits overlying the header structure dates to the early - mid 8th century (see below). This means that after centuries in which this area of the town was used for ostentatious public construction (since early within Iron Age I), it was abandoned and the use of the area changed to open-air industrial activities, before the Assyrian conquest.
The exact time-frame for the ashlar building depends on the absolute chronology one employs for the Iron Age IIA. Employing the low chronology, as we do, means that the upper (Ir2a) floor in the underlying "Benny's building" dates about the first half of the 9th century. The ashlar building was constructed soon thereafter (around the mid-ninth century), and went out of use c. 770 BCE (see below). Employing the high chronology (Iron Age II starting in the early 10th century) "extends" the life of the building. It means that it could have been constructed much earlier - around the mid-10th century, perhaps slightly later and leaves quite a lengthy time unaccounted for by floor levels etc., till its abandonment c. 770 BCE
Phase D2/6 consists of pits cutting the buildings and floors of phase 7. Three of these were dug this season. The inner stratigraphy and the different pottery horizons between the pits indicate that this phase must have lasted some time. Of the three pits, pit L05D2-517 clearly cuts pit L05D2-544. Both of them cut W04D2-065 and the crushed kurkar floors. The pottery in these pits - though not as abundant as that of pit L05D2-802 (see below) and containing but few complete vessels, nevertheless exemplifies a distinctly earlier chronological horizon. Where the bowls of L05D2-802 are mainly of Assyrian or Assyrianizing types - with folded-over and ridged rims, those of L05D2-517 and L05D2-544 are either flat-rimmed with red wheel-burnish or thin Phoenician red slip types. Where the cylindrical jars of L05D2-802 invariably lack a neck and are square-rimmed, those of L05D2-517 and L05D2-544 have 'Hazor-like' short necks with rounded rims and a ridge. The almost complete lack of Assyrian influence in the ceramics of pits L05D2-517 and L05D2-544, and even some of the latest types which appear in Assyrian destruction assemblages in Israel, indicate that pits L05D2-517 and L05D2-544, which post-date the destruction of the 'header structure' were filled-in quite some time before the Assyrian occupation of Dor.
Another discovery, which may have both chronological and cross-cultural significance, is a sherd of Aegean or Aegeanizing skyphos, bearing Late Geometric motifs, found in pit L05D2-517. Such pottery, possibly manufactured in Cyprus, is attested at Al Mina and elsewhere from the mid-eighth century B.C.E. In addition, a fragment of a commercial jar in pit L05D2-517 was inscribed with three Hebrew or Phoenician letters in red.
As to the function of pits L05D2-517 and L05D2-544, nothing much can be said at this point. Pit L05D2-517 has several distinct layers within it - some apparently burnt and some not. One of the highest tip-lines, excavated already last season had charred grain in it (according to a preliminary field-assessment by Ehud Weiss - six-rowed barley). The basal layer of pit L05D2-517 (which covered pit L05D2-544) was a thick phytolith surface. One fill layer in pit L05D2-544 contained some not-yet-identified [non-metallic] slags.
The picture in pit L05D2-802 is very different. This is actually a large complex of pits with several sub-features in it - at least 10m. in diameter and over 1.5 m. deep - parts of which were already dug in L05D2-802 and off-and-on since. It contains huge amounts of pottery and industrial wastes. The pottery repertoire is extremely restricted - almost all of it cylindrical Phoenician commercial jars, of well-known 7th century types. The second abundant type are small coarse-ware jugs virtually unique to this pit. The other sherds which are found seem to be typical household vessels of the Assyrian period at Dor (including one fragment that may belong to a "genuine" Palace-Ware bowl). Occasionally other waste, like a broken (and unintelligible) scarab and a small fragment of a bone or ivory carving were found.
Portions of the pit were excavated in detail in order to record the exact locations of the macroscopic artifacts (slags, vitrified materials, ceramics) and the associated sediments. Metal concentrations were mapped and mineral compositions were determined on-site. The vertical sections of the pit were also analysed in this way in order to obtain a 3D picture of the distributions of the different assemblages of materials disposed in the pit. About a third of the studied area contained assemblages rich in metals and glasses, whereas at the same stratigraphic level, the remained was comprised mainly of phytolith-rich layers and associated ceramics, with no evidence of metallurgical waste. The samples are now being analysed in detail at the Weizmann Institute in order to learn as much as possible about the types of metal industries responsible for this waste, and to gain more insight into the formation and diagenesis of phytolith layers.
The information obtained will hopefully contribute to a better understanding of the exploitation of mineral resources in the late Iron Age that can be compared to the well-known one relating to the Bronze Age. Late Bronze Age metallurgy in the Eastern Mediterranean, as documented in both historical and archaeological sources, is based on copper extraction at or near the mines (principally on Cyprus), and the shipping of ingots to consumers who then melt it and mix it with tin brought from elsewhere. The evidence from Dor (and elsewhere in the Mediterranean) indicates that the same pattern persists in the early Iron Age.
The economics of smelting metals at the point-of-entry to the mainland is not easy to explain. On the face of it, all the resources necessary for metal-production (ore, fuel, and probably expertise) were available right in the Troodos foothills. Why, then, go to the trouble of transporting weighty and largely unproductive materials from the mining districts in Cyprus (presumably) to the seashore and thence overseas? The answer might lie in one of two directions (or both): a) that it has something to do with the switch from Bronze production to Iron (or both Bronze and Iron from the same ores); b) that it has something to do with the centralization of production under Assyrian imperial control.
What, if anything, the metallurgical industry has to do with Phoenician commercial jars or with the idiosyncratic jugs found in quantities in the same waste-piles is a complete mystery at this point. Clearly, the elucidation of these problems will take much more work.
This period did not form one of
the goals of the season. In both lower D1 and D2 we were, by and large, below it at the outset - while upper D1 and D4 were concerned with later strata. The few Persian remains encountered were intrusive features. Wall foundations and robber trenches of phase D2/5 which went into lower strata and three large pits in lower D1. From one of the latter came a small votive terra-cotta head - a grotesque lopsided, grimacing male face - that finds many parallels in the western Mediterranean and in Cyprus.
Mention should also be made again to a feature once thought to be Persian, and finally proven not to be. We have suspected for a long time that the large architectural complex dubbed 'The Persian Palace' in D1 and connected to the so-called 'Big Mama Building' in D2 is actually Hellenistic in date. This was based on the lack of clear connections between the walls and the Persian-period floors thought to be associated with them, and on a few Hellenistic sherds found in and below the 'Big Mama Wall' when it was being dismantled.
The question always was, though, how much trust can one put in small sherds, retrieved from a wall that was exposed for many years? The point of post-depositional disturbances was driven home quite forcefully this season, when a family of gophers settled in area D2 and was wreaking havoc with the stratigraphy even as we were excavating.
After the wall was completely removed, however, a small pit was found under it. The finds from that pit were purely Hellenistic, and included three complete vessels (two 'potholders' and a lamp as well as a coin. These finally settle the question of the date of the wall which overlies them, and will provide a secure terminus post quem for the construction of the 'Persian Palace'.
In the 2000 season our attention was drawn to an ashlar-built E-W wall found in 1995 just beneath the surface on what was then the north baulk of area D1 (W16020). What was special about it was that it was molded and stuccoed at its base. When first found and for several seasons afterwards it was dismissed as late (Stern, E., Gilboa, A., Berg, J., Sharon I. and J. Zorn 1997) despite the fact that Roman walls seemed to cut it. Excavations in 1998 had shown that a later wall, abutting this 'Monument' was part of a Hellenistic complex at the entrance to the acropolis. Several architectural fragments and a statue of Nike found in 2000 (Stewart, A. and S.R. Martin 2003) set us searching for a Hellenistic public structure at this region of the town, which would have been dismantled and had its architecture reused in later Roman construction. The 'Monument' was one candidate for such. Towards the end of the 2000 season we opened several new units north of D1 to trace the walls of this building and attempt to date it. We did find a N-S wall, W16850, which corners W16020 and is similarly molded, but the end of the season and the nearness of the remains to the surface rendered this search inconclusive.
This year we returned to the same problem, opening another series of surface squares north of D1. In the westernmost of these we traced the continuation of the western N-S foundation wall of the 'Monument' to its northwestern corner. The 'Monument' building is 12 m. wide. Owing to proximity to present-day surface and later construction (see below) neither the molded superstructure nor the floors survive on the north part of the building - only the lower foundation courses. The deposits presumably cut by these foundations were Persian. The construction-methods of the 'monument' foundations are also unique for Hellenistic Dor - large ashlar blocks, lying flat side down. It was possibly constructed by a builder conversant with Greek construction techniques. As normal in Greek architecture, it was undoubtedly planned as a free-standing structure. Obviously the molding on the outside walls was made to be seen. It was quickly absorbed, however, into the agglomerative local tradition when walls - in the Phoenician 'ashlar pier' style - were abutted onto it with complete disregard to the molded decoration; converting the open space between it and the complex west of it into a series of rooms.
North of the line of the north wall of the 'Monument' a Roman street-pavement was encountered - possibly the south edge of the N-S street running through areas H and F or an E-W street with which the latter made a T junction. The beginning of a N-S drain - possibly the same one which runs under the aforementioned street - was found under it. The edge of the street made a neat curb with the line on the north wall of the 'Monument' - so either the 'Monument' itself still existed in the Roman period, or - more likely (see below) another building was built on the same foundation.
More information is also available on the structure that cut the 'Monument'. It is a larger building - its walls extending beyond those of the 'Monument' at least on the south and on the east. Its two main (N-S) longitudinal walls are constructed of flat ashlars on a rubble foundation - perhaps stylobates for colonnades. The rubble foundation of the western wall clearly cuts the molded 'Monument' wall. In the eastern wall the flat ashlars of the stylobate protrude eastwards from the foundation, and the protrusion is beveled underneath. A possible explanation is that the floor east of this wall was meant to be a step lower than the floor to the west. A scenario based on these assumptions would be to reconstruct the building as an east-facing stoa. An elaboration of that scenario would be that this stoa replaced a similar - though smaller - Hellenistic building.
West of the western of the two walls referred-to above was found an extent of floor - constructed of cement on a bed of cobbles. Two floors of the same construction were superimposed on the east side - the original one-step-down floor was replaced by one of the same elevation as that on the west. The deposits under these floors are still Roman. Several subterranean installations were discovered here.
On the east side a well-constructed kiln was discovered. It is stone-built, dug into the ground on the east, south and west, and open to the north. The firing chamber was divided into two clay-lined basins - where the clay was baked by the heat of the fire. West of the wall, and also sealed by the floors relating to it is another stone-lined installation - but this one had no evidence of firing in it.
In the area called 'D1 sector 1' in 2004 was found a large Roman building - the outer walls of which extend into D2. Two phases were recognized in the structure (the upper phase being further divisible to two sub-phases). This building forms a discrete stratigraphic / architectural unit whose correlation to adjacent areas is not straightforward. Therefore, we decided to dig it as a separate area henceforward and renamed it 'D4'.
Our problem in 2004 was to determine the function of the building. The size of the structure, thickness of walls and depth of foundations signify an edifice of some importance. The thick cement and /or rough white mosaic floors, as well as several channels, basins and drains indicate the processing of some liquid. Two very different hypotheses were put forward: a) a large civic building - possibly a public bathhouse; or b) a large industrial facility - but the nature of the industry remained unclear. While this season's work added a lot of data - the same problems persist.
In order to further elucidate the plan and function of the building we extended its exposure this season by opening new units to the north and to the east. These revealed four new rooms.
The most important data for the understanding of the structure come from the NE room. Along the robbed western wall of the room and under the cement floor (but not sealed by it) runs a terra-cotta water pipe of 15 cm. diameter. Near the middle of the pipe there are marks on the cement floor for some installation - a basin? - attached to the floor and nearby is an up-spout in the pipe. The pipe is robbed as it approached the south end of the room and it is not clear if it originally turned under the floor and went elsewhere or just ended.
In the latest use-phase of the building (or perhaps even after its destruction) the cement floor was cut by two installations. One is a circular depression - 250 cm. in diameter - cut through the plaster face of the floor and down to the cobbles which make the sub-flooring. At is center is a hole with raised ridge around it. Yeshu Dray suggested - when he visited the excavation - that this installation is a base for a donkey-powered mill. The other installation is a large (110 diameter) tabun preserved for most of its height (about 120 cm.). It was dug some 90 cm. deep under the floor and obtained its oxygen from surface level by means of a terra-cotta pipe (of the same type and diameter) as the water pipe in [the early phase of] the same room.
Does any of this help in elucidating the function of the building? The water pipe under the floor strengthens the supposition that the structure has something to do with large amounts of water. The pipe is of the same type and diameter as the main water-supply pipe found in areas A and C extending northward from the terminus of the aqueduct in area B. In fact, in a previous article about the water supply and distribution system at Roman Dor (Berg, Zilberstein and Sharon 2002) we hypothesized that the water main circled the high tell from the north (the [presumed] water-level at the terminus of the open gravitational conduit was c. 12 m. a.s.l. - not nearly high enough to scale the 14.50 m. high tell directly). We presumed that the pipe skirted the NW point of the tell, where a bath-house was excavated in area E, and made its way south to area F - where a section of piping was found in front of the main entryway into the temple complex. If it then continued southwards, area D would be a logical end to its circuit. The elevation at which the D4 water-pipe was found is consistent with such an explanation. Our structure may then be a public bath-house, as previously hypothesized, or some sort of a fountain-house distributing water to various pools and installations. The vague term 'fountain-house' is employed here, rather than nymphaeum, to stress that the building is not a decorative, stylized monument but a much more utilitarian structure. In its reuse phase this room might have been a commercial bakery - with an industrial-sized mill and a large oven. Such commercial bakeries have been found in Pompei, Ostia, etc. (Junkelmann 1997: 110-127).
The next room westwards also had several pyrotechnical installations in it. One is a clay-lined tabun-like dome - smaller than the one just described and apparently built on the floor surface rather than dug into it. The other does not seem to have any lining, but is rather built of an haphazard stack of small stones and hypocaust tiles, stuck together with clay which was baked by the heat of the installation.
The third, westernmost room excavated this season is again of a different nature. It had no installations in it, and its floor is not made of cement. Two distinct floor levels - the upper of them made of glycimeris shells - were found here. Whereas the other rooms were found virtually empty, there was a rich assemblage of pottery and several coins both above and between the floors of this room - providing a late 1st century/2nd century CE date for the building.
This sector encompasses the [robbed] western wall of the D4 building and the area between it and the 'monument' structure to the north. This area is most problematic for several reasons. First - it suffered heavily from stone-robbing. Virtually no walls were left in it and the deposits are crisscrossed by robber trenches. Though the lines of the robber trenches are often visible it is next to impossible to determine which phase[s] the ghost walls belong with, and hence any reconstruction of the plans of the phases is doubtful. Also, this sector forms a stratigraphic 'no man's land' between the western sector of upper D1 - where the Roman strata are almost completely eroded (except for wall foundations and subterranean features) and Hellenistic remains are encountered at almost surface level; and D4, in which the depth of Roman strata is considerable.
Nevertheless two phases - both dating to the Roman period - can be distinguished. There are possible sub-phases in both. The upper phase contains some bits of cement floors and cement-and-rubble foundations while the lower one consists of accumulations (or intentional fills) of usually unconsolidated horizontally-deposited sediments, and ashlar walls reusing cut-down column drums and capitals from an as yet undiscovered Hellenistic colonnaded building, some of which were published in 2003 (Stewart and Martin 2003). It is not quite clear which of these Phase 2 'floors' formed proper floor surfacse and which are floor makeup or simply constructional fills.
In almost all areas excavated at Dor at least two separate Roman construction-phases were noted. It is tempting to simply think in terms of two Roman Strata e.g. STR I = H/1 = D1/1 = D2/1 = D4/1 = B2/1... STR II = H/2... but this may be misleading. In some cases we have clear differences in architectural conception and layout between the phases - e.g. in area H where phase 2 is a residential insula replaced in phase 1 by monumental temple construction; or [parts of] area D2 where phase 1 is characterized by thick cement/conglomerate foundations which disregard all previous construction whereas phase 2 consists of thin walls which are essentially a rebuild of the Hellenistic ones. However, such criteria are often nonexistent elsewhere and there is a high degree of continuity. For instance - both phases of D4 seem to relate to the same basic architectural layout - and it is one of massive walls and deep foundations. I.e. it is possible that D4/1a; D4/1b and D4/2 all relate to D2/1. The only way to reliably work out the interrelations between different architectural complexes is to 'close the stratigraphy from the bottom' i.e. find a common structural element under the Roman strata and work out how many phases and sub-phases there are above it. For this reason, for the time being, we chose the tedious way of dividing the field to different areas and sub-areas according to [presumed] architectural complexes and working out a different stratigraphic scheme for each of them.
The southwestern part of the city in the Late Roman era was occupied, then, by an extraordinary concentration of heating installations. These include six installations found in area D3 as long ago as 1993; at least seven in the eastern sectors of area D2; and at least five in H. To these were added at least five excavated this season in areas D4 and upper D1 this year.
We tend to use the terminus technicum 'tabun' for any clay-built installation that was fired in situ to produce a beehive-like terra-cotta cone. This is of course a misnomer not only because the original meaning of the Arabic word is 'oven' - and none of these installations were household ovens - but because the term groups together installations of different types - most likely designed for diverse functions at a wide range of temperatures - ovens, kilns, furnaces - and some perhaps were never heated (other than for making the installation itself) but were used as vats or containers.
Indeed, even the installations uncovered this year reveal a wide range of types (and probably of uses). We have installations primarily constructed of clay , and ones made of stone ; ones with separate firing chambers and ones in which the load (whatever it was) and the fuel apparently shared the same space; side-loaders and top-loaders; ones with a wide air flue (for oxygen intake or smoke output or both) and ones with a small aperture - probably for a tuyère - i.e. blast furnaces.
Although the significant concentration of heating installations dates to the Roman period - they are not all contemporary. Stratigraphically some belong to the very last occupation (in their respective areas) while others were found sealed beneath the final floors (but see the cautionary note about the stratigraphy above).
The one characteristic shared by all the installations thus far uncovered is the almost complete lack of [macroscopic] slags or wasters. It is quite frustrating that even after we have been excavating these installations we still do not know what industry (or industries) they represent. Whatever it was - it must have been economically significant, as a large part in what must have been a prominent sector of town was devoted to it.
One last observation concerns the urban setting of these activities. In Politics II.8 Aristotle says:
"Hippodamus, the son of Euryphon, a native of Miletus, the same who invented the art of planning cities... was the first ... who made inquiries about the best form of government. The ideal city of Hippodamus was composed of 10,000 citizens divided into three parts-one of artisans, one of husbandmen, and a third of armed defenders of the state. He also divided the land [of the city] into three parts, one sacred, one public, the third private."
The plan of Roman Dor seems an unseemly (from this point of view) mixture of these domains. In area H, a prosperous residential district is peremptorily taken over by a sacred precinct, across the street from which is an industrial park. Meanwhile, just a few meters to the south, a civic structure is built over a kiln, while another is being converted to an industrial installation. Whether or not Dor had a 'Hippodamian' street-grid in the Roman period, it appears that in spirit it is anti-Hippodamian indeed.
The Weizmann team operated on site a Fourier Transform Infrared spectrometer for mineral analyses, as well as an X-ray Fluorescence spectrometer for metal ion anlysis. They also used petrographic microscopes for analyzing sediment components. About 500 samples were analysed on-site and the research now continues in the laboratory at the Weizmann Institute. The sediments exposed both in D2 and D1 were analysed in real time during the excavation in order to adapt the mode of excavation and the recording to reflect the findings. In D2 several more prominent phytolith-rich layers were identified, as well as one that appears to have been burned in-situ. In D1 a systematic analysis of the various cements and mortars was begun. High concentrations of lead were discovered at several locations, and the various types of installations were sampled in detail in order to try to identify their original functions. In addition, detailed mapping and analysis was carried out of the large pit in D2 (see above for more details).
We continued the efforts begun last season to begin architectural conservation on some of the old excavated areas. The strategy followed was the same as last year - a combination of a one-week on-site conservation course and work by the volunteer force - both supervised by one professional conservator (Ravit Lin). We worked on the same area as last year - area H, concentrating this time on two projects - the area of the shops facing the Roman street and rebuilding a doorway and staircase which led from the lower level of H to the upper level.
Fifteen graduate students participated in the conservation course - mainly archaeology students from Haifa University and architecture students from the Tel Aviv University's historical building conservation program. The course was concentrated into five 8-hour days (7:00 - 4:00 with an hour's break) and included five hours of fieldwork and three hours of lectures daily.
Following last year's project of building a public path from the beach to the site - in order to channel the visitors and prevent them from endangering themselves and the remains - we both repaired the stairway built last year from the southern beach to the top of the mound, and constructed another one - leading from the western beach to the top of the tell between area H and area F.
Advances in the program to fully computerize the workflow from field registration to final publication, begun in 2003, included the testing of migrating the excavation's databases from MS Access to MySQL. Using MySQL's OBDC handlers meant that - for the individual staff member the two systems are completely transparent. Most did not even know whether they were working in Access or in MySQL. The new database system will, however, be able to support many users, simultaneously accessing the same database on a network. Such multi-tasking is necessary for several users (e.g. a recorder, several unit supervisors, a photographer, an architect and an artifact registrar at the museum all update the same area's data at the same time.
A second step towards the same goal was the testing of a wireless network connecting the tell and the museum. This would, in the future, enable recorders and supervisors to enter data directly onto a server in the museum from their workstations on the tell.
The last step towards doing away completely with manual registration on the tell was the installment of a barcode tag printer on the tell. Instead of manually filling out basket lists, maintaining a graphic log and write out tags - and then re-enter the same information on computer in the afternoon, the automated recorder's station can print out a tag as soon as the entry in the automated basket list is complete, thereby significantly cutting down on the workloads of recorders. Moreover, the barcode on the tag can be read by a barcode reader which automatically pulls out the relevant record[s] from the database for viewing or editing once the tagged artifact comes up during artifact reading.
This work has been made possible by a donation of equipment and expertise by Mr. Bob Roswell of System Source Inc.
Junkelmann, M., Panis Militaris. Die Ernährung des römischen Soldaten oder der Grundstoff der Macht, 1997, Mainz am Rhein.
1 A short note on area-naming is in place here. Area names at Dor are completely arbitrary - denoting merely a contiguous set of excavation-units supervised by one area supervisor and sharing a 'bank' for locus, basket and image IDs. Sub-area designations occasionally change from season to season. Thus the area designated 'D1 sector 1' in 2004 was renamed 'D4' this season - the reason being that the stratigraphic sequence found there is different than that of the original D1 (if anything, it correlates with the D2 sequence). The 'D1' designation for both 'upper D1' and 'lower D1' was kept for historical reasons - as both of these were excavated in the past as parts of the same area. In retrospect - this should have been changed. These areas henceforward will be excavated by different teams, pursuing different goals and different periods. Similarly, this season has shown that within 'upper D2' there are two different zones (nicknamed here 'the Monument area' and "outside the Monument area") with differing occupational sequences. We will need to think about renaming the sub-areas of D come next season.