• Figure 6 Trench D Excavation
  Figure 6
  
  Excavating in Trench D, the 30-meter section trench, looking east
  
  from the Tell Deir ‘Alla archive).
  
  Halbertsma (2019)
  from Halbertsma (2019)
• Figure 7 Trench Layout
  Figure 7
  
  Layout of the trenches of the 1960s excavations at Tell Deir 'Alla (Franken 1969, 13).
  
  Halbertsma (2019)
  from Halbertsma (2019)

• Figure 13 Map of Iron Age sites
  Figure 13
  
  Map of Iron Age sites discussed in this chapter
  
  (after Van der Steen 1999, 179).
  
  Halbertsma (2019)
  in the Jordan Valley from Halbertsma (2019)

• Location Map from
  Fig. 1
  
  Map of Deir 'Alla region, based upon topographic map 1947
  
  Van der Kooij (2006)
  Van der Kooij (2006)
• Jordan Valley Sites from
  Figure 6
  
  Archaeology of the Jordan Valley.
  
  

  • White squares, main populated areas cited in historical documents
  • white dots, archaeological sites visited and reappraised in this study
  • gray dots are archaeological sites not studied here (lack of evidence and/or available literature) but of potential interest for future studies
  • gray squares, paleoseismic sites
  • black squares, geomorphological sites studied by Ferry et al. (2007)

  
  
  Ferry et al. (2011))
  Ferry et al. (2011)
• Fig. 2 Map of Iron Age I
  Figure 2
  
  Map of the Jordan Valley with Iron Age I sites mentioned in the thesis. The open circles are modern cities.
  
  Halbertsma (2019)
  sites in the Jordan Valley from Halbertsma (2019)
• Fig. 13 Map of Iron Age sites
  Figure 13
  
  Map of Iron Age sites discussed in this chapter
  
  (after Van der Steen 1999, 179).
  
  Halbertsma (2019)
  in the Jordan Valley from Halbertsma (2019)

• Location Map from
  Fig. 1
  
  Map of Deir 'Alla region, based upon topographic map 1947
  
  Van der Kooij (2006)
  Van der Kooij (2006)
• Jordan Valley Sites from
  Figure 6
  
  Archaeology of the Jordan Valley.
  
  

  • White squares, main populated areas cited in historical documents
  • white dots, archaeological sites visited and reappraised in this study
  • gray dots are archaeological sites not studied here (lack of evidence and/or available literature) but of potential interest for future studies
  • gray squares, paleoseismic sites
  • black squares, geomorphological sites studied by Ferry et al. (2007)

  
  
  Ferry et al. (2011))
  Ferry et al. (2011)
• Fig. 2 Map of Iron Age I
  Figure 2
  
  Map of the Jordan Valley with Iron Age I sites mentioned in the thesis. The open circles are modern cities.
  
  Halbertsma (2019)
  sites in the Jordan Valley from Halbertsma (2019)
• Fig. 13 Map of Iron Age sites
  Figure 13
  
  Map of Iron Age sites discussed in this chapter
  
  (after Van der Steen 1999, 179).
  
  Halbertsma (2019)
  in the Jordan Valley from Halbertsma (2019)

• Fig. 2 Aerial view of
  Fig. 2
  
  Aerial view of Tall Dayr ‘Allå
  
  (after G. Van der Kooij 2006)
  
  Kafafi (2009)
  Tall Dayr ‘Allå from Kafafi (2009)
• Deir 'Alla in Google Earth
  Deir 'Alla
  
  click on image to explore this site on a new tab in Google Earth

• Fig. 2 Aerial view of
  Fig. 2
  
  Aerial view of Tall Dayr ‘Allå
  
  (after G. Van der Kooij 2006)
  
  Kafafi (2009)
  Tall Dayr ‘Allå from Kafafi (2009)

• Plate IVa Phase C Earthquake
  Plate IVa
  
  Photograph of section 5.50/15.50-20, with earthquake crack, dated to phase C
  
  Franken (1969)
  Crack from Franken (1969)
• Plate IVb Phase C Earthquake
  Plate IVb
  
  Section Drawing of section 5.50/15.50-20, with earthquake crack, dated to phase C
  
  Franken (1969)
  Crack drawing from Franken (1969)
• Plate Va Seismic displacements
  Plate Va
  
  A typical section through phases A-D courtyard deposits. Section to the left cf. fig. 3. The black burnt deposit clearly shows the disruptions by the shifting of the earth. In the centre: collapsed sides of a pit. Looking S.
  
  Franken (1969)
  (Phases A-D) from Franken (1969)
• Plate Vb Phase B Flower
  Plate Vb
  
  Trench D, looking s. Foreground phase A courtyards. Burnt floor of B 1 furnance. Top left the round tower, (wall K 13), cut through wall F (G) 8. Below the round tower remains of wall E 7 and phase D courtyards.
  
  Franken (1969)
  Structure and burnt floors in Trench D from Franken (1969)
• Figure 21 The same Phase B
  Figure 21
  
  Trench D100, looking south from D500. Visible is the heavily burnt surface of one of the installations (also visible in figure 18), and the various earthquake cracks complicating the stratigraphy in this part of the tell
  
  (from the Tell Deir ‘Alla archive).
  
  Halbertsma (2019)
  Flower Structure and burnt floors in Trench D from Halbertsma (2019)
• Figure 6 Trench D Excavation
  Figure 6
  
  Excavating in Trench D, the 30-meter section trench, looking east
  
  from the Tell Deir ‘Alla archive).
  
  Halbertsma (2019)
  from Halbertsma (2019)
• Plate VIb Phase A Flower
  Plate VIb
  
  Phase A accumulation over the red burnt L.B. ruins. Thick ash deposit (D 908), interrupted by pits and cracks.
  
  Franken (1969)
  Structure from Franken (1969)
• Plate IXa Reed foundation and
  Plate IXa
  
  Wall F 8 in section, showing the reed 'foundation' and cracks running through it
  
  Franken (1969)
  cracks from Franken (1969)
• Plate IXb Bowed Walls
  Plate IXb
  
  Wall J 3 and wooden beam underneath. On the left wall J 1
  
  Franken (1969)
  from Franken (1969)
• Plate Xa Phase J road levels
  Plate Xa
  
  Phase J road levels, looking s. and stone pavement running along wall J 1
  
  Franken (1969)
  from Franken (1969)
• Plate Xb Phase K wall
  Plate Xb
  
  Phase K, wall K 13, the round tower dug into the ruins of wall F 8
  
  Franken (1969)
  dug into ruins of wall F8 from Franken (1969)
• Plate XIa Phase M wall
  Plate XIa
  
  Phase M, wall M5
  
  Franken (1969)
  from Franken (1969)
• Plate XIb Phase M cistern
  Plate XIb
  
  Phase M cistern, showing the circular pattern of the compressed fill in the horizontal section.
  
  Franken (1969)
  from Franken (1969)
• Plate XII Fill of Phase M
  Plate XII
  
  Fill of Phase M cistern near the surface. In the centre, a wall of a house built on a clay deposit over the cistern with stone floor. In the burnt debris of this house, a mediaeval grave
  
  Franken (1969)
  cistern from Franken (1969)
• Fig. 13 burned layer separating
  Fig. 13
  
  The burned layer, which separates the two phases
  
  (photo by Yousef Zu‘bi)
  
  Kafafi (2009)
  the two Middle and Late Bronze Age phases from Kafafi (2009)
• Fig. 17 burned phase at the
  Fig. 17
  
  General view of the burned phase at the south foot of the tall
  
  (photo by Yousef Zu‘bi)
  
  Kafafi (2009)
  south foot of the tall from Kafafi (2009)
• Fig. 18 four burned wooden
  Fig. 18
  
  General view of the four burned wooden pillars and the stone pavement south of them
  
  (photo by Yousef Zu‘bi)
  
  Kafafi (2009)
  pillars and the stone pavement south of them from Kafafi (2009)

• Plate IVa Phase C Earthquake
  Plate IVa
  
  Photograph of section 5.50/15.50-20, with earthquake crack, dated to phase C
  
  Franken (1969)
  Crack from Franken (1969)
• Plate IVb Phase C Earthquake
  Plate IVb
  
  Section Drawing of section 5.50/15.50-20, with earthquake crack, dated to phase C
  
  Franken (1969)
  Crack drawing from Franken (1969)
• Plate Va Seismic displacements
  Plate Va
  
  A typical section through phases A-D courtyard deposits. Section to the left cf. fig. 3. The black burnt deposit clearly shows the disruptions by the shifting of the earth. In the centre: collapsed sides of a pit. Looking S.
  
  Franken (1969)
  (Phases A-D) from Franken (1969)
• Plate Vb Phase B Flower
  Plate Vb
  
  Trench D, looking s. Foreground phase A courtyards. Burnt floor of B 1 furnance. Top left the round tower, (wall K 13), cut through wall F (G) 8. Below the round tower remains of wall E 7 and phase D courtyards.
  
  Franken (1969)
  Structure and burnt floors in Trench D from Franken (1969)
• Figure 21 The same Phase B
  Figure 21
  
  Trench D100, looking south from D500. Visible is the heavily burnt surface of one of the installations (also visible in figure 18), and the various earthquake cracks complicating the stratigraphy in this part of the tell
  
  (from the Tell Deir ‘Alla archive).
  
  Halbertsma (2019)
  Flower Structure and burnt floors in Trench D from Halbertsma (2019)
• Figure 6 Trench D Excavation
  Figure 6
  
  Excavating in Trench D, the 30-meter section trench, looking east
  
  from the Tell Deir ‘Alla archive).
  
  Halbertsma (2019)
  from Halbertsma (2019)
• Plate VIb Phase A Flower
  Plate VIb
  
  Phase A accumulation over the red burnt L.B. ruins. Thick ash deposit (D 908), interrupted by pits and cracks.
  
  Franken (1969)
  Structure from Franken (1969)
• Plate IXa Reed foundation and
  Plate IXa
  
  Wall F 8 in section, showing the reed 'foundation' and cracks running through it
  
  Franken (1969)
  cracks from Franken (1969)
• Plate IXb Bowed Walls
  Plate IXb
  
  Wall J 3 and wooden beam underneath. On the left wall J 1
  
  Franken (1969)
  from Franken (1969)
• Plate Xa Phase J road levels
  Plate Xa
  
  Phase J road levels, looking s. and stone pavement running along wall J 1
  
  Franken (1969)
  from Franken (1969)
• Plate Xb Phase K wall
  Plate Xb
  
  Phase K, wall K 13, the round tower dug into the ruins of wall F 8
  
  Franken (1969)
  dug into ruins of wall F8 from Franken (1969)
• Plate XIa Phase M wall
  Plate XIa
  
  Phase M, wall M5
  
  Franken (1969)
  from Franken (1969)
• Plate XIb Phase M cistern
  Plate XIb
  
  Phase M cistern, showing the circular pattern of the compressed fill in the horizontal section.
  
  Franken (1969)
  from Franken (1969)
• Plate XII Fill of Phase M
  Plate XII
  
  Fill of Phase M cistern near the surface. In the centre, a wall of a house built on a clay deposit over the cistern with stone floor. In the burnt debris of this house, a mediaeval grave
  
  Franken (1969)
  cistern from Franken (1969)
• Fig. 13 burned layer separating
  Fig. 13
  
  The burned layer, which separates the two phases
  
  (photo by Yousef Zu‘bi)
  
  Kafafi (2009)
  the two Middle and Late Bronze Age phases from Kafafi (2009)
• Fig. 17 burned phase at the
  Fig. 17
  
  General view of the burned phase at the south foot of the tall
  
  (photo by Yousef Zu‘bi)
  
  Kafafi (2009)
  south foot of the tall from Kafafi (2009)
• Fig. 18 four burned wooden
  Fig. 18
  
  General view of the four burned wooden pillars and the stone pavement south of them
  
  (photo by Yousef Zu‘bi)
  
  Kafafi (2009)
  pillars and the stone pavement south of them from Kafafi (2009)

• from Fischer (2006)

• MB II (?)/III-LB IC at the eastern base of the tell
• MB II (?)/III-LB IB at the southeastern base of the tell
• LB IA-end of LB on the northern slope
• LB II at the north-western base and the south-western slope of the tell
• MB III (?)-LB II at the southern base of the tell (see p. 224 Table 10)

• the initial phase, viz. MB II/III
• the follow-up settlement which includes major parts of the LB I and II
• the final stage which ended with a general conflagration and only some limited rebuilding, viz. the end of the LB II

• from Fischer (2006)

• from Franken (1969:40-42)

• from Franken (1969:44-46)

1.1.2 — The Iron Age I deposits at Tell Deir 'Alla

... the Iron Age I remains at Tell Deir ‘Alla consist of no less than 4 meters of well-documented accumulated archaeological deposits. These deposits contain significant architectural features, as well as a wealth of finds such as figurines, ornamental pottery, bronze objects, and personal ornaments. As mentioned above, these Iron Age I deposits have been thoroughly documented, but were published focussing more on the pottery chronology (Franken 1969).

... Abandonment, destruction, and resettlement: change

Several sites in the Jordan Valley appear to be abandoned, or have been abandoned, and subsequently resettled in the period of the 12th century BCE. Clear evidence for this pattern is attested at Tell Deir ‘Alla. During the Late Bronze Age Tell Deir ‘Alla housed an extraordinarily substantial religious structure, or sanctuary. This sanctuary had been built already during the Middle Bronze Age, on an artificially levelled platform. It consisted of a central cella, which functioned as the ‘holy of holies’ of the sanctuary, and was flanked by several store-rooms which contained ceremonial and functional pottery, of both local and imported Aegean origin, imported objects such as cylinder seals, and inscribed tablets. The sanctuary’s architecture shows features of Egyptian building characteristics, comparable to the Fosse Temple at Lachish (Franken 1961, 365). Based on these characteristics, it was postulated that the sanctuary possibly functioned as a hub for a regional market-economy (Franken 1992, 178), trading between the Jordan Valley, Egypt, and possibly Syria and Lebanon. Somewhere after 1180 BCE the temple was destroyed most likely by an earthquake. The specific date was attested by the terminus post quem provided by a cartouche of Queen Taousert from the latest layers of the temple. This destruction event is witnessed in a clearly recognisable destruction layer, with mud-bricks burnt to a degree of vitrification (Franken 1961, 367). This event did not cause the inhabitants to abandon the site immediately, however, as it is apparent from the archaeological record that attempts were made to rebuild parts of the sanctuary, and possibly salvage some of the temple’s inventory (Franken 1969, 20). Another fire resulted in the end of this phase, and the site was briefly abandoned. However, this abandonment phase did not last long, as on top of the debris from the Late Bronze Age sanctuary a series of industrial installations were built. Franken suggests that this was done by a group of newcomers, as mentioned above, largely on the basis of a new pottery repertoire unlike that of the Late Bronze Age inhabitants (Franken 1969, 20-21). These phases, Franken’s Phases A-D, fall somewhere in the second half of the 12th century BCE. These phases have been published, but largely limited to the analysis of the pottery chronology (Franken 1969).

... Tell Deir ‘Alla Phase B belongs to the so-called ‘first period’ (phases A-D as published by Franken (1969, 33)), a period attributed to semi-nomadic habitation of the mound following the Late Bronze Age destruction (e.g. Franken 1969, 38; Van der Steen 1996, 68). In general terms the period is characterised by a limited amount of architecture, a large quantity of dug pits, and accumulations of courtyard layers. This ‘first period’ started with the repurposing of the remaining standing architecture of the LBA [Late Bronze Age] sanctuary, possibly for shelter.

4.1.1 — Tell Deir 'Alla during the 'First Period'

Franken subdivided the stratigraphy of the Iron Age I at Tell Deir 'Alla in two main periods: the 'First Period' (Phases A-D), and the 'Second Period' (Phases E-M). The former, which is the focal point of this thesis, begins when the remaining Late Bronze Age architecture was used for 'squatter' habitation (Phase A). The remains of the Late Bronze Age sanctuary were cut into, as attested by many dug pits and a large depression of around 2 meters deep. These features gradually filled back up with wash layers containing Iron Age I pottery, but were still intermixed with Late Bronze Age pottery. Franken attributes this to the digging activities at the site, which resulted in the inclusion of Late Bronze Age material in the Iron Age I assemblage. The large 2 meter deep depression was, according to Franken, used for storage of fuel such as brushwood. This fuel caught fire and burnt over at least the entire excavated area, resulting in ashy patches of a white and pink colour, sometimes as much as 20 cm thick. This fire caused the last standing architecture from the Late Bronze Age to collapse, and subsequently wash over the accumulated Iron Age I layers (Franken 1969, 33-34).

In Franken’s publication of the Iron Age I stratigraphy, this is where Phase B starts. A sequence of three installations positioned one atop the other were observed, with possible evidence for at least one more directly north of the former three, dating to Phase B (Franken 1969, 36-37). According to Franken, these installations were enclosed by four mud-brick walls, creating a square shape (see figures 26 and 27). The original inside width of the installations was approximately 2 meters. Excavating this phase proved somewhat problematic due to several substantial cracks running through this area (see figure 21), causing the stratigraphy to sink at several places. Franken notes that the cracks would dry out during the excavation, causing the soil to shift up to 20 cm (Franken 1960, 389; Franken 1969, 38).

After the last use-phase of the Phase B installations, they were destroyed. Subsequently, numerous small layers began to cover the installations, marking the beginning of Franken’s Phase C. While two new walls were built on the remains of the installations, it is unclear what how they can be interpreted. This area in Trench D was heavily weathered due to its proximity to the tell’s surface. Numerous layers containing burnt clay are deposited, often with pits dug into them.

4.2 Stratigraphic analysis of the Phase B excavation trenches

A starting point had to be defined to begin working with the Tell Deir 'Alla archive and explore its potential. To this end a well-defined context had to be chosen, small enough to allow the author to investigate it in the span of this MPhil research, and significant enough to contribute to the current scientific discourses. The Phase B installations met these criteria, and therefore were chosen as a 'pilot study' for the current research. As the Phase B installations fall largely within two trenches, Trenches D100 and D500, and the overall chronology of Phase B is most clearly represented here, this provided a well-defined and manageable starting point. As mentioned above, the relevance of these installations for the Iron Age I Jordan Valley made these all the more appealing. While they are discussed in the literature, their archaeological contexts haven't been fully published yet. While it still proved a substantial task to get fully familiarised with the excavation system and the documentation, the selection of the Phase B installations as the focal point of the current research and the limiting of the analysis of the stratigraphy to these two trenches, provided a realistic starting point for working with the Tell Deir 'Alla archive.

In this section a detailed analysis of the stratigraphy of the Phase B installations will be provided. As mentioned above, the preceding Phase A is very clearly visible in the stratigraphy. It appears as a dense sequence of wash layers, as described by Franken. In the subsequent period many changes occur, as suddenly various phases of building activities can be observed. Franken already reflected this in a change of stratigraphic phase, ascribing the subsequent phase as Phase B. This division still holds true, which is why it will be adopted in this thesis.

4.2.1 — Data from the archive

While nothing is guaranteed when working with old archives, a surprising amount of data concerning the Phase B installations could be retrieved. Apparently, Franken had already realised in the field how important the Phase B installations would become, as these features are covered relatively generously in the archive. Numerous photos document the features, as well as various drawings, sections, and sketches in the fieldnotes by Franken himself. Apart from the usual drawings, which were made for every excavated square, smaller and targeted top-plans and cross-sections were also drawn. Below is a summary of what the archive contains regarding the Phase B installations.

Sections

As was usual for every excavated Trench, all the sections in the sub-squares of D100 and D500 were drawn. However, due to the steep slope towards the north of trench D no north-section was made for sub-square D100. This was not unusual, as Franken applied the 'walking baulk' method described in chapter 2. Instead, the south section from adjoining square D500 is referred to, as an 'inverted' north section for square D100, and is documented as drawing Dllb (see figure 22). The regular section drawings for sub-square D100 are DI (the large eastern section covering the entirety of Trench D, see figure 9 and figure 24), DII a (the south section), and DIII (the west section, see figure 25). Apart from the regular section drawings, which were made for every excavated square, smaller cross-sections of soundings made to clarify the stratigraphy were also drawn.

Top-plans

As with the section drawings, Trench D100 had several major top-plans, being D1, D2, D3, and D4. These top-plans all portray the Phase B installations, in various stages of the excavation process. D4, for example, shows the situation when the installation are encountered. D2 shows remains of the first use of Phase B, as well as some Late Bronze Age architecture. D3 shows the first installation (see figure 26), and D1 the latest (see figure 27). Other than these drawings, which were made in the field, summarizing drawings were made combining all of the top-plans. In these the furnaces are plotted one atop the other. Other than these drawings, the archive also contained several smaller top-plans from Phase B. These are often difficult to trace back to their stratigraphic location, as not all of them contain clear measurements forthe grid system.

Fieldnotes

The fieldnotes relating to Trenches D100 and D500 are in one single notebook. The fieldnotes were written by Prof Dr H. Brunsting, a Dutch archaeologist, who was the sole supervisor of the excavations of Phase B in Trench D (see figure 19). The fieldnotes pertaining to Phase B were written between 8 February and 27 March 1960, indicating it was excavated over the period of roughly a month and a half. The fieldnotes vary in quality and content, with certain deposits being described in much greater detail than others. References are made to the section drawings and top-plans, and certain sketches and excavation strategies explain decisions made in the field. The overwhelming majority of the fieldnotes, however, describe very briefly what work was done in what area of the trench. This leaves out descriptions of the deposits that were excavated, as well as the precise location of many finds, complicating stratigraphic analyses.

... Photographs

Thus far a total of 55 separate photo’s portraying information from Phase B in Trench D have been recovered. They vary in quality, and it is apparent that several times multiple photographs were shot of the same feature at more or less the same angle. These were likely made with the ‘Miranda’ camera (see figure 18), which often required taking several of the same shots to ensure a positive outcome when processing the negatives. The majority of these photographs capture one or more aspects of the Phase B installations, indicating these were a clear priority for the excavators.

... 4.2.2 – Franken’s publication of Phase B

In the 1969 publication, Franken published his stratigraphic interpretation of Phase B (Franken 1969, 36-40). While he published several top-plans, as well as a cross section of the installations, these were summaries of the actual drawings. Franken’s main goal for the 1969 publication was to publish the pottery chronology, which should be kept in mind when interpreting the stratigraphy on that basis. He published the walls with different numbers than the ones used in the field, but did keep the original deposit numbers. This makes retracing his steps somewhat complicated, as the variety of numbers in the 1969 publication do not reflect the documentation in the archive, which is less detailed.

Franken identified 5 sub-phases in Phase B, based on the stratigraphy. These cover the entire period that the furnaces were in use, the first of which he postulates was likely built in the very last part of Phase A. He notes that the east, west, and south walls of the installations form the installations' shape, which stays more or less the samethroughout Phase B. Franken recognised three separate installations, but the first one was dug through a possible earlier one constructed in Phase A. As such, a sequence of 4 installations was recognised, all constructed one atop the other. The exact outline and location of the phase A installation could not be retraced, but the walls from the first Phase B installation were cut down into the older walls.

The installations do vary somewhat in shape. The first one, D153, appears to have two small and narrow openings protruding towards the north. The opening was wider for the second installation, but for the third installation appeared blocked. These openings appeared to shift slightly northward with each new building phase. Stones were incorporated into the walls of the chambers of the first installation. These stones had split from heat exposure, and were coloured dark red. The consistency of the floors was hard, and they'd burnt to a depth of ca 5 cm. Burnt clay made up the fill beneath the installations, and the soil around them contained burnt clay particles and "many tiny drops of metal" (Franken 1969, 36-38). Franken was quite brief in discussing the actual stratigraphy of the phase, and doesn't explain on what basis the sub-phases were divided.

Franken makes an interesting side-note in his publication, which is relevant for later interpretation of the installations. As it is part of his initial interpretations and described the archaeological deposits, it should be mentioned here:

It is however a curious fact, that slag was only found in large quantities from phase E onward. And this is not slag from melting metals but from the burnt walls of kilns or furnaces. It can be found in masses near the kilns of the modern potters village of Kerami. From the present evidence the accumulation of deposits on the tell in this period h as to be explained mainly by the destruction of furnaces and moulds, and this is possible as these deposits contain a large amount of burnt clay particles, and ash in patches or large stretches in abundant. Slag and stones were probably systematically removed from the surface by the smiths and thrown down the slope so as not to get mixed in the clay used for moulds. The villagers from phase E on brought it back while quarrying fortheir houses (Franken 1969, 38).

Table 3

Subdivision of phases with corresponding field-numbers from Trench D100 and D500

(after Franken 1969, 39-40).

Halbertsma (2019)

4.2.3 — Re-evaluating Phase B's stratigraphy

With the data gathered from the archive the chronology of Phase B could be re-evaluated. The cross-sections proved a reliable starting point for analysing the stratigraphy, as it was clear that they were complete and relatively detailed. All the sections were joined together to see if Franken's sub-phases could be recognised, and if reinterpreting the stratigraphy could elaborate his analyses. Fortunately, all the sections could be connected neatly to one another. This made interpreting phases in the stratigraphy possible, and thus also re-evaluating Franken's subdivisions. However, as mentioned above, large cracks in the sections do very much complicate this stratigraphic picture, as deposits appeared to indeed have shifted over 20 cm at certain places.

While several top-plans were made for the different phases of the installations, these did not appear to connect well to the sections. On the top-plans several height-measurements were indicated. However, in various cases the areas connected to the sections were left blank, or the planned deposits and features simply did not correspond to those indicated on the sections. As such, it often remained unclear where exactly a wall was supposed to enter- the section, or which deposit in the section connected to which deposit in the top-plan. In short, while containing valuable information, horizontally the documentation, especially with regards to the top-plans, is less than ideal.

Another complicating factor was the observation that several layers had shifted and subsided throughout time. Apart from the obvious consequences of large earthquake cracks, it appears that the pressure of the top of the tell caused layers to shift downward in the centre, and outward towards the edges, as attested by the light curve of the various interfaces. By consistently measuring the lowest point of the sub-phases, the relative differences in elevation will provide an idea of the thickness of the layers, and how quick building events follow one another. For that purpose, in the analysis of the stratigraphy a rough estimation of the elevation relative to the fixed point on the mound will be provided, consistently measured from the lowest point of the deposit. These heights will be used to connect the top-plans to the sections where possible.

The reinterpreted chronology

The stratigraphy of the Phase B is clearest in section DI lb (see figure 22), of which Franken also published aversion in his 1969 publication (Franken 1969, 38). As mentioned above, this is actually the south-section of Trench D500, and as such should be taken as an inverted north section for Trench D100. Here the sequence of the successive Phase A wash layers is very clear, as are the changes that occur at the start of Phase B. The south-section of the installations, an unnamed section, does not show this transition, as south of the installations the situation does not appear to change much in comparison to Phase A. Unfortunately, this section does not cut the southern part of the installations, resulting in a lack of coverage for the installations' building sequence on the southern part. Instead it covers the are a just south of the southern walls, where it shows a sequence of accumulated layers covering Phase B. The west-section, section DIII (see figure 25), also shows the chronology very well, with Phase A represented by an accumulation of wash layers, and construction events marking the beginning of Phase B. However, the area of the installations is represented here by a single mudbrick wall, marked as D7, with numerous cracks through the middle, complicating distinguishing separate building phases. The east-section, Section DI (see figure 24), is more helpful in this regard. It also shows the transition from Phase A to Phase B, and clearly contains several construction events regarding a mudbrick wall. Numerous layers run up against this east-west oriented wall, which is marked on the section as walls D7 and D9.

From these sections, most clearly exemplified through section Dllb of which a Harris Matrix and section drawing is provided (figures 22 and 23), the sequence of events outlined below was reconstructed. The finds from Phase B were traced back to their original sub-phase, and listed below. As the pottery is already published in the 1969 publication, only their registry numbers and category are added. For the other find categories a small description of the object is added, and where known, their approximate find location.

Sub-phase 1

After numerous wash layers, a mudbrick wall was built in an east-west orientation (unit 2), roughly at 13.70m. This wall is clearly visible in section Dllb (see figure 22), and might be visible in section DI (wall D7 cuts a smaller wall with wide flat bricks, which might be the same feature, see figure 24). Layers associated with this sub-phase are described in the fieldnotes as burnt, containing a lot of charcoal.

Unfortunately no top-plan was made for this sub-phase. Also, no detailed information exists regarding to the find-locations of the associating objects. ...

Sub-phase 2

The area was completely levelled (unit 4 and 30), as is visible in section DIIb, at roughly 13.50. Afterwards a substantial wall was constructed on the western side (wall D7 on DIII) which likely connected to the bricks seen in Section DIIb (unit 6). These walls belonged to the first installation, of D153. The opening which is visible from D154 in section DIIb appears cut into the wall, indicating the wall was likely opened at a certain moment in time. Subsequently the opening in the oven filled up (unit 7 and 8), as did an area just east of the installation (unit 9). Most of this sub-phase was then covered by another fill, possibly to even out the surface once more (unit 11). Though visible in the sections, the layers from this sub-phase are not described in the fieldnotes.

This sub-phase includes the earliest of the recognised installation floors, which is depicted in top-plan D3 (see figure 26). The centre of the floor was measured at 12.90 height in the top-plan drawing, but at the opening in the wall of the installation it is noted that the surface dropped 35 cm downward, placing it at approximately 13.25. The top-plan for this sub-phase of the installations shows a square feature, with a demarcated surface area (D153). This surface is described as heavily burnt in the fieldnotes. It has an inner dimension of ca. 2.20 by 2.20 meters, but the outer walls are not clearly drawn. Several stones can be observed on the outerwalls of the installation, as well as specifically drawn mudbricks. These stones were described in the fieldnotes as "wall-reinforcements". Furthermore, there appear to be two narrow gullies toward the northern edge, which are described in the drawings as 'air-ducts', which were apparently heavily burnt. These gullies are not visible in the Dllb section, which instead shows an uninterrupted fill of ca. 27 cm thick. The top-plan shows that in the inner south-eastern corner of the installation a 'test trench' was excavated, as D155. The fieldnote entry for D155 states that a cross-section was made for this test-trench, but unfortunately this drawing was not encountered in the archive. Again, no clear descriptions exist as to the original locations of the finds from this sub-phase. ...

Sub-phase 3

At the start of Sub-phase 3, the entire area was levelled again (units 12 and 13; possibly D149 on DI), at roughly 13.20m. This event also took down part of the large western wall (unit 5; wall D7 in DIII), and possibly also part of the eastern wall (wall D7 in DI). Another western wall was then built, again north-south in orientation (unit 16; possibly D148 on DIII). Mudbricks can be observed in east-west orientation along section DIIb, suggesting that another wall was built enclosing the northern part (unit 14 and 15). The eastern section here is quite unclear, due to the levelling activities of the next sub-phase (possibly DI 149?). According to the fieldnotes, this sub-phase contained the charred remains of a large 'tree-trunk' or 'roof-beam', which yielded a lot of charcoal.

No top-plan was made for this sub-phase, and the find location of only one item belonging to this sub-phase was described. ...

Sub-phase 4

Most of the area was again levelled (interface 17 and 18), at roughly 12.90m. Wall 16 was left in place, and likely reused in this sub-phase.A new substantial east-west oriented wall was built in the east, visible as wall D9 in section-drawing DI. In the north section, DIIb, two newly built 'air-ducts' were constructed (unit 19). It appears that after these were constructed, another east-west wall was built over or around these, possibly incorporating them into a larger structure (unit 28 and 29). This wall was eventually broken down in the same place as the openings of the other installations, and the surface filled with debris (unit 21 and 22; D148).

The 'airducts' are visible in top-plan D3 (see figure 27), but no separate top-plan was drawn for this sub-phase. The 'air-ducts' can be seen as protruding slightly in comparison to the other installations. According to the fieldnotes, they contained a small layer of ash on the inside, and the edges of the bricks had been baked to a very hard consistency. Furthermore, the fieldnotes mention a lot of charcoal was encountered in this sub-phase, as well as several "chunks of lime". ...

Sub-phase 5

While the situation on the western side is quite unclear for this phase, as it reaches top-soil, the eastern side shows that the area was levelled for a final time (unit 25), at roughly 12.70m. Wall D9 appears not to have been affected by this event, and remained in place as attested by various layers running up against it (DI D147, see figure 24). What these layers are is not explained in the fieldnotes. From section DIIb (see figure 22) it would appear that yet another east-west oriented wall was then constructed (unit 25 and 26), which ran into wall D7 in the east section. This gives an estimation of its thickness, averaging around 1 meter. This east-west wall was eventually broken down, although the part in the section appears to have remained intact. The area between the ruins of the wall was filled with debris (DIIb D502), after which a sequence of layers covered the entire area (unit 31). These layers contain several stones, and the fieldnotes mention that the fill contained charcoal and ochre. Layer D502 runs up against this wall, as seen in section-drawing DI, indicating that this is most likely a debris layer washing down from the decaying mudbrick architecture. An interesting observation done by Brunsting in his fieldnotes, is that he thinks he might have removed part of another structure, similar to the latest installation. This installation would have been located just north-east of the other one, sharing the southern wall. This is questioned by Franken in the fieldnotes, however. No mention was made of a possible second installation in the 1969 publication.

Sub-phase 5 is reflected in top-plan DI. This top-plan shows the latest of the installations. The elevation of the centre of the floor is 12.60. On the plan it was noted that the surface towards the northern edge sloped down 12 cm, fitting neatly with the 12.70m read from section DIIb. This sub-phase of the installations appears to have included a wide opening in the north, which is not visible in the section. This installation contained two upright mudbricks against the western wall (see figure 31), which create three separate niches. Its surface showed dark patches of charcoal and ash (see figure 18). The surface was apparently heavily burnt, resulting in a hard-baked surface. None of the objects belonging to this sub-phase have been provided with any precise spatial information. ....

The end of Phase B

After Sub-phase 5 section DIIb portrays only top-soil (see figure 22), and section DI shows that while the large wall remains standing for a while, numerous small layers are slowly deposited up against it (see figure 24). No new building phases are recognised in Trench D following the installations, indicating that either Phase B came to an end after the final installation was covered, or continued in an area not yet excavated.

4.2.4 Discussion of stratigraphy

Based on this re-evaluation of Phase B's stratigraphy in Trench D, a slightly more complicated picture emerges than was published by Franken. It is important to state here, however, that only Trench D was researched for this thesis. As Phase B was also encountered in a number of other excavation squares (to be specific Trenches E, F, G, L and M), several of his phases relate to stratigraphic events occurring elsewhere. That said, the stratigraphy of Phase B is most clearly represented in Trench D, and could already be refined just on the basis of analysing this area. This research has recognised five sub-phases in Trench D, which are associated with Franken's sub-phases 1 to 3. While Franken mentions the presence of his sub-phase 4 in trench D100, during the current re-evaluation of the original documentation no stratigraphic units could be linked to this phase. This suggests that two of Franken's sub-phases were not visible in the stratigraphy of this area, and the preceding three sub-phases actually consisted of five.

The sub-phases proposed above likely begin already in the end of Franken's Phase A, when the first mudbrick feature was constructed. While the exact dimensions of this feature are difficult to establish, as it remains only visible in the north section and possibly in the east section, it is proposed by Franken that this was the earliest installation. This is something that is also mentioned in the fieldnotes. Brunsting wrote an entry in the fieldnotes (10/03/1960) discussing the stratigraphy, which mentions that the ovens had 3 or 4 phases, each with its own floor. Only three floors were recognised in the field (D153, D150, and D148), and only two provided with a top-plan (D153 and D148). It is not unlikely that, due to the close superimposition of these layers and the complicated stratigraphic situation caused by the large cracks in the soil, one or two separate floor levels were missed while excavating. Another possibility is that two floor levels were largely erased already in antiquity, possibly during the extensive levelling activities demonstrated above.

Summarising the sequence of events, it appears that there were five building events (all clearly recognisable in section DIIb) which were all initiated after a levelling event. While these building events appear to follow one after the other relatively quickly, the associated wall in the west section remains largely untouched throughout all five sub-phases, with the exception of one levelling event in Sub-phase 3. This indicates that the general outline of the installations remained the same throughout all five building events, which is reflected in the top-plans. While the installations' east and south walls can only be recognised on the top-plans, the substantial walls visible in the east section, walls D7 and D9, also show continuation of construction in the same location. The first installation was likely rebuilt 4 times after its initial construction, in more or less the same configuration.

The emerging picture shows a rather large mudbrick installation, measuring around 5.5 by 4 meters, oriented at a slight north-east angle, which was enclosed by four walls. The northern wall did not only cover the northern face of the installations, but continued in an eastward orientation into the section. Sometime after this wall was constructed it was opened in the western half of the trench, in front of the installation, as attested by clear cuts in the walls. The remaining debris then was covered with what is described in the fieldnotes and publication as a "clay fill" (Franken 1969, 36). This process was recognised in the stratigraphy 5 times. While the north wall appears to have been often de- and reconstructed, the east, west, and south walls likely remained more or less in the same position.

This leaves two 'anomalies' to investigate, however. The first of these two are the 'air-ducts' from sub-phase 3. These protrude slightly from the northern wall, as attested by top-plan D3. From the top-plans and section DI it would seem that the area to the south of these 'air-ducts' would not contain any architecture, instead consisting of numerous sequential layers. This might indicate they formed part of a different structure, perhaps located directly north of them. As mentioned above, the trench supervisor noted that he might have removed a similar structure to the Phase B installations to their north, and sketched the outline of this structure in his fieldnotes (see figure 33). This situation, however, would require a north-south wall in the section, which was not recognised.

The other anomaly is the large wall continuing in the east section, walls D7 and D9. These two walls, built one atop the other, appear not to have a clearly recognisable function from the top-plans and sections. There are layers running up against the walls on the north and south sides, but no architecture can be connected to these walls. As they do not appear to have perpendicular adjoining walls in the section, it is likely they continued further into the section, into the unexcavated area east of trench D100. The area east of D500, which is sub-square M200, was excavated by Franken in the 1964 season. Remains from this area mostly belonged to the Late Bronze Age sanctuary, which were published in the 1992 publication (Franken 1992). However, also Iron Age I material was excavated in this sub-square, which was later published by Van der Steen (2008c). This study shows that walls D7 and D9 did not have an adjoining wall to their north, as no such wall was recognised in the stratigraphy (see Van der Steen 2008c, 72). The west-section for this area was not published, however. The results from Van der Steen's publication will be further discussed below.

• from Franken (1976)

• Fig. 8 - Plan of Phase C
  Fig. 8
  
  Plan of Phase C
  
  Franken (1969)
  from Franken (1969)
• List of Deposits of Phase C
  List of Deposits of Phase C
  
  Franken (1969)
  from Franken (1969)

• Fig. 1 - Plan, mainly Phase IX (M)
  Fig. 1
  
  Plan, mainly Phase IX (M)
  
  Ibrahim and van der Kooij (1986)
  from Ibrahim and van der Kooij (1986)
• Fig. 2 - Section through Phases I-IX
  Fig. 2
  
  Section through Phases I-IX
  
  Ibrahim and van der Kooij (1986)
  from Ibrahim and van der Kooij (1986)

Fig. 1 is a preliminary plan of the architectural remains found this season, including those remains partly unearthed in previous seasons. The 1967 excavations lay further to the W/NW; only the remains in square B/C5 excavated at that time are included in this drawing. Pl. XII also shows about half of the digging area. Some of the wall fragments drawn in the E part of the area cannot yet be definitely attributed to phase IX — they may belong to X.

The mudbricks of the walls had been placed directly on the surface; no stone foundation was made, but a layer of reed was used instead (a reed foundation was also used for walls of phases VIII and V and partly VI). The mudbricks used mostly measured 47 x 11 x 35cms, but several walls were made of a slightly larger type; other slight differences do occur as well. Almost all the walls were only as thick as the length of a brick, some even as thin as a brick's width. The material used for most of the bricks came from the banded clay beds of the low natural hills nearby, but often different clay/ mud was used for bricks employed in one wall (e. g. the EW wall B/B6.127 had three layers of grey and three of ochre coloured bricks; cf. also ADAJ XXII p. 65). Stones and bricks were only used for square platform-like structures made at some walls, facing each other (e.g. in squares B/A & B8).

The function of these structures is not known but in one case a basin-like construction was made on top of the structure (see below). Most of the rooms/spaces enclosed by walls measure ca. 3 x 3.5m, but some are up to 10m in length. Half of them had a solid roof (roof debris found on the floor), but at least two of the other spaces had been covered by reed matting (in squares B/C6 and B/C7). The connections between the walls dividing spaces are not always clear, because doorways have not always been recognisably preserved, perhaps because of the low level of the wall stumps left. In any case courtyards or alleys between the "house complexes" form a small part of the plan (in squares B/D6 and possibly B/C7) in the area excavated. Several rooms, roofed and unroofed, have floors paved with unworked small boulders (ca. 25 cm.s diameter), with whitish clay plaster filling the holes in between. The small space, originally covered with reed matting, which gave access to the plastered wall with the Bala'am inscriptions ( squares B/C5 & 6, cf. ADAJ XXII pp. 65f and ADAJ XXIII pp. 48f), had a 'bench' on the S and E sides, and maybe on the N side as well, with a shallow depression sloping towards the plastered wall in between; (see Pl. XVI, 2 which shows the remains of the mat as fallen on this ' floor') . At five other spots the same type of ca. 7mm thick lime plaster has been found, embedded in wall debris, nowhere still attached to a wall fragment (the rooms are in squares B/A5, B/A8 - each side of the NS wall -, B/C6 and B/C7, also covered with reed). In each case the amount of plaster exposed covers a surface of about 1/2 m², but in several rooms much of the original plaster may have been burnt. None of the plaster fragments had any writing or drawing visible on the surface. If this was the original case it also has to be concluded that the smooth plaster, including that with the inscriptions, had not been applied with the specific purpose of being inscribed. Several of the rooms have very thin (ca. 15cms) clay plastered brick walls, dividing the room into different compartments (very clear in B/C8), apparently for storage purposes.

The walls of the rooms were not very solidly constructed. Right angles between connected walls were not employed very often. On the other hand, the finishing off of several of the walls and floors has been done with special care. Several of the walls show repair or even rebuilding, and roof collapse during the use of the rooms, giving a new, higher floor, has been noticed at places as well.

The final destruction of the complete building complex occurred suddenly, and was, according to data found in 1967 (long EW cracks stopping at the lower edge of the debris of phase IX), caused by earthquakes. In most places fire had broken out as well, causing severe burning of the roofs and wall faces, as well as of the contents of the rooms concerned.

The 'tawabeen' in B/ D6 are to be connected with phase IX, but most of their stratigraphical context had already been excavated in 1967. The size of two of these ovens is exceptionally small (the diameters at their base are ca. 40, 50, as opposed to 75cms for the third one).

Most conspicuous are two installations uncovered in the W room in square B/A8. One is a kind of basin constructed against the W wall, partly on the floor and partly on one of the low platforms mentioned above. It had been made in situ from the broken body section of a large jar, the rather low sides being constructed of straw-mixed `clay' (now fired, size about 90 x 130cms; see Pl. XVII, 2). A small juglet (reg. no. 2867) was found at its base, but this does not explain its function; some store jars were found standing by the basin, and many loomweights and a lot of carbonised plant material (including wheat) were found in a slight depression in the floor to the S of it. At the opposite wall the other installation was set partly in the floor: a large stone, with a hole for grinding in its rather flat top surface (see Pl. XVII, 1). To the E some more stones had been laid to extend the flat surface, which was surrounded by a row of small flat vertically standing stones, so as to separate it from the surroundings (this complex measures ca. 95 x 70cms). A jug, a store jar and two craters were standing very close to it. The burnt wheat found nearby suggests a mortar for pounding wheat, but any other kind of material could have been ground in this installation.

A large amount of charred wood was found on the plastered stone floor of E room in square B/B8 (for part of it see Pl. XVI,2). The amount represents more wood than could come from a roof, and suggests some kind of furniture or installation there. Although some joints of beams now visible are certainly original, it is not yet certain what the original constructions looked like. The wood remains are being analysed, and two of the groups of beams have been preserved by glueing the soil with the charcoal on cloth.

A large number of moveable artefacts have been found, especially pottery and loomweights, as well as stone objects, and small artefacts of metal and bone. Some of these objects are mentioned below, without discussing their relations inside rooms for this report.

Complete pottery, though often broken by the destruction of the buildings, was found in almost all of the rooms. High concentrations are found in the S room of square B/B5, the SE room of B/C6, the NW room in B/C8, the NE room in B/A6, the S part of the room in B/B8 and the W and S parts of the rooms in B/A8; smaller groups were found elsewhere. A small selection of pottery shapes is shown in Figs. 3 & 4 . None of the different collections of pottery mentioned had one type of pot only (e. g. store jars). All of the larger groups had (storage) jars, as well as jugs; several had one or two cooking pots (B/C6, B/C7, B/B5, B/A6). Others had numbers of craters (B/A6 and B/B7 — see Pl. XVIII,1 in particular). Bowls occurred in the B/A8 and B/C8 groups (in B/A6 and B/B7 the " strainer" type on three legs). Plates were also rare (B/C6, B/B8, with little other pottery, and B/A8) and only one lamp was found (B/C8). Many of the type shapes of the pottery can be found W of the Jordan as well, dating from the Iron Age II (especially B) period.

Loomweights have been found in five large groups (about 30 weights each, sometimes more) and several smaller ones. The large groups were found in rooms in B/C6 (SE), B/C8 (NW), B/A6 (NE), B/A8 (SW corner and SE part), often among pottery and carbonised plant material. Most of these weights had been fired during the destruction, but some are still unfired. The groups show quite a variation in shape. For example, the group from B/C8 (already illustrated in ADAJ XXVII, Pl. CXXXII, 2) has seven types, six of which are circular with a hole perpendicular to the circle, and one is conical in shape, with a somewhat rounded base and a horizontal hole (this type occurs once or twice in the larger groups). It is interesting to note that more or less the same set of shape types has been recorded from Tell Qasile (see B. Maisler, IEJ I, 1950, Pl. 39B). This season again several of the loomweights (mainly those from B/C6 and B/C8) were found with pieces of string/thread still in the holes. This material has to be further analysed, but their existence makes it most probable that the weights will indeed have to be connected with weaving or making mats.

Different sorts of objects made of stone have been found on the floors and in the debris of this phase. Examples are grinding stones, found in a position that indicates their use during phase IX. The few lower grinding stones found were made of sandstone or basalt, and almost all the upper ones were made of sandstone. They were found in different rooms: several in B/C6 (different spots), B/C7 (N part), B/B5 (SE part), and B/A8 in different places. Also quite a number of pestles (see Pl. XIX, 2) were uncovered, mainly made of basalt, but some are hardly worked pebbles. Some of the pestles were found on floors in B/ A8 (NE room) and B/B5 (SE room). Several pebbles have been found with a high polish on several sides, and sometimes with other types of wearing. Their precise use is uncertain, but they too were found on the floors of the rooms in e.g. B/A8 and B/B5. Several other types of objects made of stone have been found, e.g. alabaster spindle-whorls in the E room in B/B5. There were some rather small fragments of basalt bowls (mortars) as well, which were most probably no longer in use at the end of phase IX. However, one basalt bowl on three interconnected high feet was incomplete, but still in use: the object was found in a working area of the W room in B/A8, and extra traces of wearing are visible at the centre of what was left of the bowl's bottom. The shape of the bowl (Pl. XIX, 1) is close to that of the type illustrated by Buchholz in Fig. 20d in his "Steinerne Dreifussschalen des agaischen Kulturkreises and ihre Beziehungen zum Osten" in Jahrbuch des Deutschen Archdologische Instituts 78, 1963, pp. 1-77.

Almost all the metal objects found are made of iron. They include seven leaf-shaped arrowheads, four of which were found in the W and NE rooms in B/A8. Several objects of bone are preserved rather well. This season again some examples of the flat oblong pointed and polished bone tool were found (see Pl. XIX, 3) , some are fragmentary, but the others were apparently ready for use. Of the objects rarely found earlier are small tubes made of bone; three single tubes were found (Pl. XX, 1 shows one of them), plus one double tube (perhaps made of ivory), with incised decoration and filled with black material. This double tube, stored in a jug found in the S end of the room in B/B7 (see Pl. XX, 2), is clearly a kohl tube like those found elsewhere in Palestine, (though rarely, see examples in E. Stern, Material Culture of the Land of the Bible in the Persian Period 538-332 BC, 1982, p. 269 note 18). Kohl tubes are quite common in Egypt throughout a long period (communication G. van den Boorn, Leiden). Also the discovery of small bone or ivory panels Pl. XX, 3) found on the stone-paved floor in B/B7 is quite exceptional. The surface of the panels was decorated densely with small drilled holes, each surrounded by two small concentric circles.

Two other artefacts, of Egyptian origin, are: a small game piece, perhaps made of faience, but painted white (reg. no. 2794), and a faience wadjet eye (reg. no. 2890). Both were found in mixed debris, so perhaps have to be taken as isolated objects not specifically connected with the culture of phase IX.

Apart from the artefacts a large amount of biological remains have been found. This includes small animal bones, e. g. from the floor of the room in B/C7, and a burnt but complete large antler of the Mesopotamian Fallow Deer, placed against the W wall of the NE room in B/A8 (Pl. XXI, 1). In 1979 and 1982 two other antlers of the same type were found in the SE room in B/C6 and the E room in B/B6. The use of the antlers in this context is still obscure, and the study of the animal bones has not yet proceeded far enough to know if more remains of this animal have been found.

Large quantities of carbonised plant material have been found on room floors, inside jars and jugs and in debris and wash layers. The botanists Mr. R. Neef and Professor W. van Zeist were able to give some preliminary results of their studies. A large variety of seeds was found, for example those of flax (apparently for the production of linseed oil), barley, wheat, lentils, bitter vetch and olives. Grapes were found together with cummin in a jug with a sieved mouth in the S B/B7 room. (For previous results of the study of botanical material, see W. van Zeist and J.A. H. Heeres, Paleobotanical studies of Deir `An, Jordan, in Paleorient 1, 1973, pp. 21-37).

Quite a lot of information about the culture of phase IX (M) has been found. A careful study of the remains has to follow. At this early stage of investigation it is difficult to get a clear picture, yet the following may be tentatively said.

The group of rooms/houses do not show a purely domestic function. The deducible number of weaving installations makes this improbable, as does the number and variety of pots in so many of the rooms. A combination of trade and living quarters is conceivable. No clear indication of a sanctuary or religious connotation is found, although the religious plaster texts found in 1967 may suggest this (no further inscriptions have been found this season), and although the " bench" room with this inscription at one end, could be interpreted as such. On the other hand, it is still possible that the quarters excavated so far belonged to a central sanctuary complex, with the deduced trading activities being part of it. However, it is also possible that the quarters belonged to a village or small town community (city walls have not been found, but it is sure that the edges of the settlement have been eroded away).

The people living and working in the rooms had rather poorly built, though well finished, brick houses. On the other hand, their pottery and other objects show some very special items. As far as cultural relations are concerned, it is clear that many of the artefacts are rather widely spread, though sometimes rare, in the S Levant during parts of Iron Age II.

A date for the destruction of phase IX is indicated by ¹⁴C analyses of carbonised seed and leaf material from the burnt debris: sample GrN 5633 gave 2690 ± 60 BP and GrN 8119 gave 2590 ± 70 BP, both when calibrated indicate a date of 770-880 B.C. This information has been published before, but it should be added here that this rough date has been statistically corroborated by ¹⁴C datings of older (phase L) and later (phase VI) deposits. Additional ¹⁴C tests will be made.

... The destruction of the buildings of Phase M, the phase to which the Aramaic text belongs, was also caused by an earthquake. Deir Alla has suffered more earthquakes, not only during the time of habitation but also afterwards. These earthquakes and tremors caused vertical cracks which in the excavated area run mostly in east-west direction. When tracing the frequency of these cracks along a north-south line we find at least one every twenty cms. The tell is thus cut into vertical slices and these slices may have sunken for instance from a few cms. to several cms. and sometimes shifted sideways, whereas most of them apparently under the pressure from higher parts of the tell are inclined to lean out to the north. Expressed in geological terms we have found, be it in miniature size normal faults, reverse faults and even pivot faults. Cracks reaching the present surface must have been caused after the tell had reached that height and cannot therefore be dated to the stratigraphy.

... unless all the cracks that run through the deposits overlying Phase M and their exact position had been recorded, we would not be able to say whether any cracks seen in Phase L stop at the floor levels of Phase M. However we had several other indications. We have recorded cracks and shifts of material that run through the ruined buildings but stop at the point where erosion began to level off the debris. These were caused by a second shock which followed the first one after the buildings collapsed and the fire caused by the earthquake had burned itself out. I shall have to refer to the second shock in relation to the position of the text.

However, the first shock was also recorded. There is a long crack about 10 cms. wide running through the deposits of a little lane which formed during Phase M and is almost 60 cms. high. This crack is closed further to the east but here a horizontal shift could be seen because the crack runs lengthwise through the low stump of a mud brick wall. The clay mortar between the bricks on both sides of the crack does not fit together any longer. The horizontal shift was about 10 ems. Such cracks have not only been recorded on paper but also on "pull offs", a method used in agriculture to take a thin slice of earth to the laboratory (Franken 1965b), in order to keep an authentic record of the accumulation of deposits. The slice is thick enough to make samples from it for microscopic analysis. Incidentally horizontal shifts of more than 30 cms. were recorded. It was the second shock which brought the preserved fragments of the Aramaic text down from their support.

... It should be added that at Deir 'Alla no stone foundations were found. Mud brick walls were built on top of existing surfaces, and occasionally on reed layers or wooden beams. However, as far as they were not buried in debris that accumulated against them (like the lane mentioned already had accumulated 60 cms. against two Phase M walls) they sometimes broke away at floor level during the earthquake. This had happened in the area where the text was found. A large portion of the east wall and the north wall of a room were cut off at floor level. Others broke well above floor level but after the whole wall had shifted. As a result standing remains of walls do not link properly any more.

The small room [where the Balaam inscription was found] measured roughly 2.75 m. x 3.25m. but the corner that was formed by the two walls that broke away at floor level was not rectangular but measured ca 115° as found by us. It did not have a normal floor. In the middle was a pit and all round the floor sloped down toward this pit or basin. The pit is slightly oval, ca 1.50 m. in diameter and ca 40 cms. deep. The floor had been dug into an existing clay deposit and was not lined with plaster. The southern half from the south wall to the middle of the basin was filled with burned clay from mud bricks, pieces of charcoal, sherds and some clay loom weights, all of which did not come from floor level hut had fallen down several meters while still burning. It scorched the walls against which it piled up and burned black the surface underneath. In this fill no fragments of plaster were found and if there had been plaster it had been burned in the fire. From which direction this burning material fell down could not be established. Nor could the walls that broke off at floor level or what broke away from the remaining wall stumps be traced. There were no traces of a burned roof. The fill of the room on the north side and beyond to the north consisted of broken mud brick wall fragments. Here no traces of fire were found, the fall was not so heavy as was the Cast with the burned debris. There were clearly large wall fragments which after their fall left open spaces between the fragments. These holes were filled up afterwards with clay wash. In this clay wash the plaster fragments with the Aramaic text were found. They had fallen between the wall fragments. Three main groups were located by G. VAN DER KOOIJ, one of the co-authors of this volume.

It would seem that the fragments came from a short distance and from the north-east. It is possible that the fall was checked by the layer of wall rubble. No fragments were found still attached to mud brick. In this case the bricks to which the plaster had been attached formed the upper level. The plaster fragments were found on both sides of the original position of the north and cast walls and they came down when a shock hit the site after the fire had subsided. They did not suffer from this fire. We assume therefore that the support of the plaster split during the first shock and that the missing parts fell down and probably burned away, whereas the surviving fragments remained in position until the second shock came. In the passage way northeast of the room there may have been a mud brick built stele with the text or there may have been a small gate inside of which the text was written on the south wall. On top of these wall fragments deposits were found to a height of 1 m. none of which had regularly been walked over after the disaster. On top of that thick destruction deposit there are erosion deposits. We have to assume that some of these erosion levels reflect the return of human activities on the site. But the ruins were not levelled off to form a new horizontal and flat surface. In the excavated area there is an undulating surface over the whole area, consisting of light coloured dusty material which is used as the basis for new building activities. This we have termed Phase N. Between these activities and Phase M the large pit was dug and used as a refuse pit, whatever its original purpose may have been. If it was meant to collect water it was a complete failure. The many earthquake cracks had damaged the subsoil so much that as soon as the pit had been dug to the required depth these cracks opened up because the clay started to dry out as soon as it was exposed. A crack may widen in a few days from a few centimeters to two decimeters. But it was used as a refuse pit and it did not fill up or break down from erosion. This points to a short period of existence. This pit does not yet contain wheel-thrown pottery, but the following period, Phase N, does have such pottery with new shapes.

It has been argued already that there is an intermediate M—N Phase. It is also clear that pottery from that intermediate period must be found on the surfaces that stratigraphically belong to Phase N. Phase N has wheel-thrown pottery and for the introduction of that pottery in Palestine we have at the moment only one site where it has been recognized: tell es-Sultan, which is traditionally identified with Old Testament Jericho (FRANKEN 1974). If we assume that the Jericho Iron Age II buildings belong to the 7th century B.C. we may also assume that the lower possible time limit of Phase M is also to be fixed in the 7th century B.C. This is possible because the erosion went on unchecked. Nevertheless there may have been a period of more than one generation before building activities were resumed on the site.

Pottery attributed to the phase of the Aramaic text is not wheel-thrown. From this great mass of broken pottery we have isolated the pieces that could be restored to the extent that it seems safe to assume that they were unbroken and in daily use at the time of the earthquake. When surveying this group of mended pottery we find few thin-walled vessels with shapes that may be called rare but all seem to come from one tradition stream of potmaking, which may be called the Iron I tradition found in Trans Jordan and Palestine. This tradition did not originate in Palestine. It is not a continuation of the Late Bronze Age methods of pot-making as found in Palestine, although it may have been related to that tradition originally, that is, sometime during the Bronze Age. ...

We have seen that Phase M consists of traces of a situation that must have existed one day in the past when an earthquake hit the site and traces of the impact of the first and the second earthquake shock. We must now consider the value of the interpretation of Phase M as a sanctuary. The earthquake has nothing to do with this interpretation, but had it not been for the earthquake the text might not have been preserved. Also thanks to the earthquake many objects were found which otherwise might have disappeared for ever. The plan of the buildings has to be partly hypothetically reconstructed since some walls were dislocated at floor level leaving barely any traces of where they stood before the destruction. Also a number of objects like the text were knocked about when the shocks hit the site. In contrast to two earlier earthquake phases we did not find human victims in the ruins. This may indicate that the disaster took place during daylight but it seems more likely that the destruction happened at night when there was nobody in these rooms. Somewhere there was a fire burning in a breadoven or otherwise, because the first shock was followed by a conflagration, wooden objects burned away like the looms, of which we found the clay weights in several rooms, and charred beams which may also partly have belonged to other wooden furniture. But what was left, the less perishable objects, was found and reconstructed as far as possible. The combination of the plan and the objects in their reconstructed form give rise to the interpretation of ca twenty workrooms, courtyards and lanes as having been immediately connected with a sanctuary. ...

It may be clear by now that there is no simple and straightforward archaeological answer to a number of relevant questions. Since the claim is made that the stratigraphical position of the text and the association with an archaeological level and the objects found in it is fairly exactly known, the question is why things can be so uncertain. The answer can be constructed empirically in the following way. Fora theoretic answer, see CLARKE 1968 and BINFORD 1972.

According to current interpretations of the literary historical sources the area of the Zerqa delta in the Jordan Valley was occupied by the Israelite tribes by the end of the 13th century B.C., when the Late Bronze sanctuary was still in full use. For centuries after there is no mention of a change of population. In other words, tell Deir 'Alla is situated in an area which, according to written historical sources, was inhabited from the beginning of the Iron Age by one or more tribes that were traditionally related to the Israelite tribes, which, according to the same sources, were then living in Palestine. There are some things we would like to know but do not know in relation with the excavations at tell Deir 'Alla. Exactly when did the Israel related tribes settle in the Zerqa delta? I refute the "solution" that a change in cultural assemblage which took place at Deir 'Alla sometime in the 13th century indicates the coming of these tribes (BINFORD 1972). Unless I have a cultural identity of such tribes I cannot compare, let alone identify, them with the assemblages before or after the change. And since after the change at Deir ' Alla in the 13th century the sanctuary was rebuilt and became again the largest one ever found in Palestine and Jordan in that period, the only question that can be raised is, do the written sources which we know refer to such a situation? I can state, however, that neither assemblage which can be dated to the 13th century at Deir 'Alla is derived from cultures known to have existed in Palestine in the 13th century B.C. or before. Moreover, we have two types of inscriptions from the second cultural assemblage: the Taousert cartouche (YOYOTTE 1962) and the Deir 'Alla clay tablets. They cannot be explained as yet when we follow the type of reasoning that is found in archaeological reconstructions of the Israelite conquest. That is to say, if one thinks fit to follow those arguments, then these inscriptions are prohibitive for the conclusion that we are dealing with such tribes. But if we take the earlier assemblage which seems to go back to the 16th century B.C., then there is equally no sense in attempting to establish a tribal identity. When the fallacies in the archaeological reasoning are clear then there remain two possibilities. Both cultures or one of them may have belonged to Israelite tribes or to non-Israelite tribes, but as long as we stick to existing cultural definitions of early Israelite tribes, they were not Israelite tribes.

It seems now that the following cultural assemblages, the one in the first half of the 12th century B.C. (Phases A-D) and the later ones including Phase M are equally not derived from those found in Palestine. In order not to confuse the argument (as is often done) I am not referring to the contents of the Aramaic text. I remind the readers, however, that we found in Phase L, preceding Phase M, an inscription consisting of a hitherto unknown script, scratched on a potter's tool (FRANKEN 1968). The question is then, how to establish archaeologically tribal identity. This question is even more important than that of the identity of tell Deir 'Alla. But neither can be solved on the basis of our present evidence. Discarding the interpretation of the Aramaic text for the moment, the archaeological interpretation of the situation in its totality does not in the least remind us of anything related to the West Bank in those times, apart from those pieces of pottery that we can now isolate from the main group and that show characteristics of Palestinian pottery. ...

The stratigraphic position of the inscribed stone (R. no. 2000) was simple, and the association with other finds and its location is remarkable. Pl. 16a shows the plan of four rooms. On this plan the entrance to room GG205 from the south is located between walls BB 320 and BB 427. When found this entrance was blocked by a wall fragment. The blockage may have been caused by the earthquake. In the corner formed by wall BB 320 and the blocked doorway eighteen burned clay loom weights were found lying on the floor. Eighty cms. north of the corner and lying against wall BB 320 inside room GG 205 three remarkable objects were found in the debris on the floor:

• the inscribed stone
• a goblet on a high foot with a spout (R. no. 1990)
• an outsize loom weight (R. no. 2006).

Only one relevant radiocarbon date of Phase M exists. The sample consisted of charred grain from deposit BB 303, which was part of the burnt layer of Phase M. The date is 2600 ± 50 years Before Present (Present = 1950). The "conventional" half-life of C¹⁴, i.e. 5568 years, was used for this date. By applying a correction based on dendrochronology, the "true age" of this sample "can be estimated at ca 800 B.C.". See the official publication: J. C. VOGEL and H. T. WATERBOLK, Groningen Radiocarbon Dates X, Radiocarbon 14, 1972, pp. 6-110; this sample GrN-5633 discussed on p. 53.

To understand this date some explanatory notes arc necessary (literature: generally J. W. MICHELS, Dating Methods in Archaeology, New York and London. 1973, pp. 148.167; for the problems involved see especially J.O.D. JOHNSTON, The problems of radiocarbon dating. Palestine Exploration Quarterly, 105, 1973, pp. 13-26).

1. 650 B.C. ± 50 means, statistically, that there is a chance of 66% that the "true age" of the sample falls somewhere between 700 and 600 B.C. If this standard deviation is doubled. the chance would be 93% that the "true age" falls somewhere between 750 and 550 B.C. More C¹⁴ dates would give a clearer indication.


2. Radiocarbon chronology differs from historical chronology:
   
   
   1. The half-life of C¹⁴ is not yet precisely known. Nowadays preference is given to 5730 ± 40 years over the conventional 5568 ± 30 years. Consequently the C¹⁴ date based on the old value has to be multiplied by a factor of 1.03 to get the date based on the new value (i.e. ca 2678 years B.P. — ca 728 B.C.)
   2. Other uncertainties and difficulties concerning C¹⁴ exist.

• from Franken (1976)

• from Franken (1976)
• quoted from Ambraseys (2009)

`I have already mentioned that the site was shaken by earthquake round 1200 B.C. The destruction of the buildings of Phase M, the phase to which the Aramaic text belongs, was also caused by an earthquake. Deir Alla has suffered more earthquakes, not only during the time of habitation but also afterwards. These earthquakes and tremors caused vertical cracks which in the excavated area run mostly in east-west direction. When tracing the frequency of these cracks along a north-south line we find at least one every twenty cms. The tell is thus cut into vertical slices and these slices may have sunken for instance from a few cms. to several cms. and sometimes shifted sideways, whereas most of them apparently under the pressure from higher parts of the tell are inclined to lean out to the north. Expressed in geological terms we have found, be it in miniature size normal faults and even pivot faults. Cracks reaching the present surface must have been caused after the tell had reached that height and cannot therefore be dated to the stratigraphy ... Moreover, unless all the cracks that run through the deposits overlying Phase M and their exact position had been recorded, we would not be able to say whether any cracks seen in Phase L stop at the floor levels of Phase M. However we had several other indications. We have recorded cracks and shifts of material that run through the ruined buildings but stop at the point where erosion began to level off the debris. These were caused by a second shock which followed the first one after the buildings collapsed and the fire caused by the earthquake had burned itself out. I shall have to refer to the second shock in relation to the position of the text.

However, the first shock was also recorded. There is a long crack about 10 cms. wide running through the deposits of a little lane which formed during Phase M and is almost 60 cms. high. This crack is closed further to the east but here a horizontal shift could be seen because the crack runs lengthwise through the low stump of a mud brick wall. The clay mortar between the bricks on both sides of the crack does not fit together any longer. The horizontal shift was about 10 ems. Such cracks have not only been recorded on paper but also on "pull offs", a method used in agriculture to take a thin slice of earth to the laboratory (Franken 1965b), in order to keep an authentic record of the accumulation of deposits. The slice is thick enough to make samples from it for microscopic analysis. Incidentally horizontal shifts of more than 30 ems. were recorded. It was the second shock which brought the preserved fragments of the Aramaic text down from their support' (Franken 1976, 7-8).

`We have seen that Phase M consists of traces of a situation that must have existed one day in the past when an earthquake hit the site and traces of the impact of the first and the second earthquake shock. We must now consider the value of the interpretation of Phase M as a sanctuary. The earthquake has nothing to do with this interpretation, but had it not been for the earthquake the text might not have been preserved. Also thanks to the earthquake many objects were found which otherwise might have disappeared for ever. The plan of the buildings has to be partly hypothetically reconstructed since some walls were dislocated at floor level leaving barely any traces of where they stood before the destruction. Also a number of objects like the text were knocked about when the shocks hit the site. In contrast to two earlier earthquake phases we did not find human victims in the ruins. This may indicate that the disaster took place during daylight but it seems more likely that the destruction happened at night when there was nobody in these rooms. Somewhere there was a fire burning in a breadoven or otherwise, because the first shock was followed by a conflagration, wooden objects burned away like the looms, of which we found the clay weights in several rooms, and charred beams which may also partly have belonged to other wooden furniture. But what was left, the less perishable objects, was found and reconstructed as far as possible' (Franken 1976, 12).

'Pl. 16a shows the plan of four rooms. On this plan the entrance to room GG205 from the south is located between walls BB 320 and BB 427. When found this entrance was blocked by a wall fragment. The blockage may have been caused by the earthquake. In the corner formed by wall BB 320 and the blocked doorway eighteen burned clay loom weights were found lying on the floor. Eighty cms. north of the corner and lying against wall BB 320 inside room GG 205 three remarkable objects were found [in the debris on the floor: the inscribed stone, a goblet on a 18]

high foot with a spout (R. no. 1990) and an outsize loom weight (R. no. 2006). A large piece of charred wood lay beside these objects in front of the passage between rooms GG 205 and GG 102. The goblet (pl. 16b) was only slightly damaged near the rim; the stone and loom weight were complete . (Franken 1976, 15).

• Fig. 1 - Top plan of the
  Fig. 1
  
  Top plan of the recovered architectural remains of the last stage of Phase IX (Area B). Dotted lines refer to later disturbances (pits, erosion, egalisation)
  
  Ibrahim and van der Kooij (1991)
  recovered architectural remains of the last stage of Phase IX (Area B) from Ibrahim and van der Kooij (1991)

• Fig. 1 - Site Plan
  Fig. 1
  
  Site plan showing the squares excavated in 1987 and 1994, as well as the army trenches made in 1991.
  
  Ibrahim and van der Kooij (1997)
  from Ibrahim and van der Kooij (1997)
• Fig. 7 - from Ibrahim and van der Kooij (1997)
  Fig. 7
  
  Phase IX; B/C7.50 with pottery including:
  
  

  • Phoenician jug with herbs
  • Platter (originally with barley gruel)
  • to the right: crushed jar with stored wheat

  
  
  (Photo: Hubert de Haas)
  
  Ibrahim and van der Kooij (1997)

• Fig. 5 - from Ibrahim and van der Kooij (1979)
  Fig. 5
  
  Ibrahim and van der Kooij (1979)

• Fig. 1 -                    
  Fig. 1
  
  Tell Deir 'Alla Excavation Trenches
  
  Franken and Ibrahim (1978)
  Plan of Excavation Trenches of the 1977-1978 seasons from Franken and Ibrahim (1978)
• Fig. 2 -                    
  Fig. 2
  
  Square A/C1,S section
  
  JW: note presence of "earthquake cracks" looking like what one would find in a paleoseismic trench
  
  Franken and Ibrahim (1978)
  Square A/C1,S section with earthquake cracks from Franken and Ibrahim (1978)
• Fig. 3 -                    
  Fig. 3
  
  Square A/C4,S section
  
  Franken and Ibrahim (1978)
  Square A/C4,S section from Franken and Ibrahim (1978)
• Fig. 4 -                    
  Fig. 4
  
  Square B/C5, plan phase "L" (broken truction). line: traces left; dotted line: reconstruction
  
  Franken and Ibrahim (1978)
  Square B/C5, plan phase "L" from Franken and Ibrahim (1978)
• Fig. 5 -                    
  Fig. 5
  
  Square B/C5, plan phase M (broken line: traces left; dotted line: reconstruction).
  
  Franken and Ibrahim (1978)
  Square B/C5, plan phase M from Franken and Ibrahim (1978)
• Fig. 6 -                    
  Fig. 6
  
  Squares B/C and D5, main part of E section
  
  Franken and Ibrahim (1978)
  Squares B/C and D5, main part of E section from Franken and Ibrahim (1978)
• Fig. 7 -                    
  Fig. 7
  
  Square B/C5: reconstruction of the position of the plaster texts.
  
  Franken and Ibrahim (1978)
  Square B/C5: reconstruction of the position of the plaster texts from Franken and Ibrahim (1978)

Parts of the newly made section drawing are reproduced here (fig.2), namely the section in A/C1, parallel to the old section at 15-20m, The NS walls, H18 and H17 (see Franken, op. cit., 1969, fig.81), as well as H15 (mistaken called H19 on section fig. 81) to the E of them still exist, but a connecting EW wall appeared. Also E of wall K4 there seems to be an EW wall. The water hole between walls K3/L2 and K4/L3 still exists, with a continuous sedimentation at its bottom. An EW wall also appeared in the section to the W of the given drawing, W of the NS wall H14. It goes through the new section at a slightly oblique angle; it has a plastered face to the S, connected with a plastered floor. A nicely rounded small oven (ca 60 cm) was built against this EW wall close to NS. wall H14.

Close to it a horse head figurine was found (see pl. XXVI, 2)

The excavated area lies between the old lines O.37 m and 10m. The NS walls H13/J5/K1 and II1/J1 are still clear and a wall on top of J1 in phase K also appeared. It is interesting that these walls have a thin layer of reed at their base visible at several places. The thick deposits of street layers between these accumulations of walls still extended as far as the new S section. Part of an EW wall appeared in section W of the NS wall H1 (see fig. 3). Wall 4 (see fig.3) probably belongs to phases H and J, and wall 3 to phase K. A thick rusty coloured floor (level c.-209.60m) was found; it stopped just before the new S section against mud-brick debris. It probably belonged to phase J. The thick deposits of rubble and ashy layers between the walls belong to phase K or L. Wall 1 (phase L?) seems to have been dug into the contemporary surface of the tell and was based first on a row of flat stones, which were also lying more to the SW. It was connected with an EW wall running about lm N of the new S section. Stones had also been put obliquely against the base of these walls (cf. Pl.XXVII.1). On top of the stone pavement, clay debris had fallen and the fill was continuously used for pits (e.g. graves). (P1-XXVII.1) shows the pavement consisting of flat stones, with stones standing against and partly inside the base of two walls that are forming a corner. The complex is-in the process of being unearthed-looking towards the WSW.)

B/C5 equals the 1967 square EE300, in which the Aramaic plaster texts were found (cf. especially Hoftijzer, Van der Kooij, op. cit., 1976). In the following paragraphs frequent reference will be made to this publication, using the abbreviation Aramaic Texts). Many obscurities in the stratigraphy immediately connected with this find made it necessary to do some additional digging without really extending the excavation area. The 1976 dig gave some suggestions which seem to be important for the interpretation of the texts. This is reason enough to report rather extensively about the excavations in this area.

Firstly, a general presentation of the stratigraphy of the area.

In 1967 excavations were carried out in the area reaching the earthquake and fire destroyed phase M - the top of it being c. 2.25m below the tell surface at that spot: The stratigraphy has not yet been published, but some of it has been described in Aramaic Texts, pp.8-10, and 26-27. (A full report of it can be expected after more of this phase has been unearthed). This season work concentrated on the stratigraphy in the immediate surroundings of the plaster texts find spot. In '67 the plans of the walls in the area had many lacunae - no time was left for a thorough study then. What was known was published - provisionally - in Aramaic Texts, cf. Pl. 17b (N.B.: The NS grid-lines of the '67 excavations 40.50 and 45.50 m about equal the new grid-lines 20 and 25 m from c.d.p. The EW lines-4.50 and O.50 m equal 11.50 and 16.50m from the base line). This 1976 season excavations penetrated somewhat deeper at a few places than what is labelled phase M (in '67 too digging had gone deeper than this phase at a few small spots). it seems useful to give a short description of the chronological order of the excavated deposits and the phases. (Date from '67 combined with those from '76)

1. The earliest deposits touched are charred wash layers, which presumably have to be attributed to phase L.


2. Some structures were built; provisionally they are also taken as belonging to phase L ("L"). Their walls have been excavated, but their floors generally not (see figs.4 and 6-plan and section). The S extension of wall B/D5.51 is uncertain. In '67 a few bricks had been seen at the place indicated by two short parallel lines. There was, however, another indication of this extension: the mat of reed leaves (see below) was seen and described in '67 as going upwards at that spot. Wall B/C5.55 was, apparently, built against the stump of a still older wall (B/C5.70); the same may be true for wall B/C5.56 (still to be excavated). Wall B/C5.63/64 consists of a row of stones with mud-bricks on top of them.
   
   The building was burnt and destroyed.


3. After the destruction much weathering had taken place, which may indicate several years' exposure. Pit 13/D5.65 was constructed (a hole was dug, and the pit was constructed of clay bricks and lined with three layers of clay plaster, coloured grey, yellow and red-B/D5.64; the room outside this structure was filled in again).
   
   Then this well made pit was filled with a wall of mud-bricks, some brick rubble, and vegetable material, which shrunk considerably after some time (no traces of another function for the pit were found). Then the heavily charred earth of B/D5.52 was deposited together with some stones.


4. After an apparently short time new buildings were erected (taken as belonging to phase M). The surface, however, still bore traces of the old walls. A glance at figs. 4 and 5 shows that at least parts of the old "L" plans were used. It became clear in '76 that the joint between walls B/C5.39 and 38 was made with differently coloured mud-bricks than had been used for the walls themselves. (The bricks of the walls contained much of the locally found banded clay; the bricks of the joint did not have this clay at all, but a lot of straw temper; they were grey, the others yellow) A crack in the W extension of this wall 38 had been penetrated by fire, which proves that the fire and crack existed at the same time - probably both caused by an earthquake. In the plan of fig. 5 the same difference in brick type is presumed to have been used at the joint between walls B/C5.39 and 42. The N extension of B/C5.36 is unclear. Its existence is probably testified by bricks seen and drawn (in '67) at the point where it crosses the 16.50 m line, and is strongly suggested by the sloping position of the fallen mat of reed leaves seen and photographed in '67. Very little was seen of B/C5.42 in '67. B/D5.53 probably (not certainly) was a row of fallen bricks. Wall B/C5.52 was built on a layer of reed stems (not leaves).
   
   B/C5.69 is a structure made of mud-brick clay on top of the stump of "L" wall B/C5.55. Two "floors" should be mentioned: the one of room B/C5.34 with a rounded pit with a graduated sloping edge (see Aramaic Texts, P1.17a), and floor B/C5.58 which is clearly visible. Judging from the rubble on top of floor 58 there certainly was no permanent roof above it. There was, however, a large mat woven from reed leaves, at least five layers thick. The size of the mat was roughly 3.5 m (NS) to 5.5m (EW) (a small EW trench had been made, supervised by A. van As, in B/D6, NE of B/C5, which produced this evidence).
   
   Then some destruction occurred, with fire. In B/C5.34 burning rubble (with brick clay and wood) tumbled down, possibly the roof or something coming from a higher place (so Franken). The heat penetrated into the unplastered floor and the wall bases. Burning rubble also fell into the room SW of this room 34, but NE of it very little burning rubble fell on floor B/C5.58. Brick rubble (unburnt) fell on B/ C5.69, and probably by now the B/D5.53 bricks had been deposited. Very little weathering took place, and then a complete destruction followed. At first the mat of reed leaves came down, but it was immediately followed by the chalk plaster with (and without) writing and whatever object it adhered to, and parts of the walls. No burning occurred, but some of the bricks were still charred, apparently by the previous small scale destruction. The main rubble deposits are B/C5.30 (on top of 34) with combination 1 of the plaster texts and B/C5.57 with combination 2.
   
   Cracks inside the deposits indicate earthquake shocks. The destruction was complete, though some of the walls were still standing 1 m high; rubble engulfed them. Erosion and wash affected the surface and this made phase M clearly distinguishable from layers deposited later.

A special paragraph has to be devoted to a reconstruction of the original position of the plaster texts, deduced from the data from '67 and'76 (see fig.7).

The direction in which the pieces had fallen has been reconstructed already (see Aramaic Texts, pp.26f.) by indications given by the positions of the different fragments. The result was that the original position must have been around the NE "corner" of room B/C5.34 (=EE334 of the old grid) The plaster of Comb. 1, the upper part of a column, and some other pieces, had fallen inside room 34, together with unburnt brick rubble on top of the previously fallen burnt rubble. The pieces of Comb.2, the lower part of a column, had fallen together with rubble B/C5.57 on top of the NW corner Of floor B/C5.58 (slightly more to the E than indicated on P1.17b in Aramaic Texts). Most probably there was a continuous wall between 34 and 57, namely wall B/C5.36 (see above), although a doorway cannot be excluded for certain.

Combination 2 must have been fallen from a very short distance (this can be deduced from the arrangement of the fragments in situ). Comb.l. came from a greater distance (the fragments were lying together quite unrelated). Comb.2 was apparently lying against the slope of the stump of old wall B/D5.51 and the new wall B/C5.36 (some rubble had fallen in between).

Consequently it most probably had been attached to something flat standing against the E face of this wall (The flat object may have been solidly connected with the wall, e.g. just projecting a little bit out of it; there are reasons enough to maintain the idea that the plaster was attached to something else than just a wall, see Aramaic Texts, pp.25ff.). It is reasonable to assume that all the writing and reading was done at one side only as the texts with related contents are close together. This would mean, that Comb.l. had been originally rather high up on the object against the E side of wall 36, and had fallen, together with parts of the wall, towards the W or SW inside room 34. The conclusion from this would be that the texts were written on plaster applied to something flat standing high up against the E face of wall 36. The wall (this part being c.3m long) formed the W end of an area which was bordered to the S by wall B/ C5.52 and structure B/C5.69 along the base of this wall; to the N it was bordered in one way or another by the bricks of B/D5.53 and /or a wall standing .there or slightly more to the N. The area stretched about 5 m towards the E. Its floor B/C5.58 had clearly often been used (indicated by numerous thin floor layers; in the NW part some large sherds and a complete lamp were found). The woven matting cover has been described above. Only the W end and the N edge of this area has been excavated. Doorways have not yet been found.

The surroundings of the area are partly clear: rooms border it to the W (34, with a strange pit-floor) and S (probably a room, with some useful finds - see below) and at a greater distance, at 5m from bricks B/D5.53, also to the N (the complex of four rooms excavated in '67; see Aramaic Texts, P1.16a); the area in between is unclear.

It is not clear how long the plaster with text had been in existence. However, it can be taken for certain that it was there already before the first small scale destruction with fire took place, by which room 34 got the burnt rubble on its floor, and the few pieces of burnt and some unburnt rubble fell on floor 58 and structure 69. The weathering of the rubble on or close to floor 58 was negligible and less than that of rubble 34. Probably no roof covered 34 after the first destruction (this is indicated by the fact that there was very little or no rubble between the lowest pieces of plaster and the surface of the burnt rubble). So rain could do its work in 34. Above 58, however, a new woven matting cover was quickly placed (it is possible also that the old matting cover was not destroyed at all). The writing on the plaster of Comb.2 had a lot of damage, probably caused by water. Most or all of this damage was done during the time before the collapse (the places on the different fragments where this damage occurs indicate this). Comb 1 also had water damage, but very little, and this may have happened after the collapse, as the fragments were lying in the apparently wet rubble, or during the destruction (the matting cover had fallen just before the plaster came down). This would mean that it was raining during the final destruction.

• Fig. 1 - Plan, mainly Phase IX (M)
  Fig. 1
  
  Plan, mainly Phase IX (M)
  
  Ibrahim and van der Kooij (1986)
  from Ibrahim and van der Kooij (1986)

• Fig. 3 -                     
  Fig. 3
  
  Example of section drawing with description of the deposits.
  
  Wall 1, phase F wall 8. Wall 2, phase C wall 5. Wall 3, phase B wall 10. D 267: Clay wash from wall 1. D 268: Gritty deposit (from building period of wall 1). D 269; D 276: pink coloured layers of organic material and fine clay deposits. D 279: clay wash. D 281: burnt surface, down to a depth of c. 10 cms. Organic material and ash, overlaid by thin levels of organic material. D 282, 329, 330: pink coloured organic material and clay. D 289: clay rubble from wall. D 292: clay wash from wall. D 299: partly red, burnt rubble from wall. D 331: gritty rubble from wall. D 332: rubble and wash from wall. D 333: Organic material and clay. D 334: Organic material and clay wash, overlaying a burnt surface. This surface runs under wall 2 and heat discoloured the earth to a depth of c. 5 cms. This burnt surface runs over levels of clay debris and wash from wall 3. Note the shift of slices of deposits, showing that usually the material sinks down on the north side. Cracks running through wall 1 did not cause a shift of material and have not been drawn. The crack just north of wall 1 caused a difference of height of the deposits of c. 25 cms., as can be seen on the burnt surfaces, drawn with a thick black line.
  
  Franken (1969)
  from Franken (1969)

• Figure 3 -                     
  FIG. 3
  
  Time-stratigraphic correlation chart of Iron IIb excavations throughout an extensive region of Israel and Jordan. Datum for correlation is earthquake debris or rebuilding horizons assigned to a single, 750 B.C. seismic event. Locations of cities are shown in Figure 1. Strata names at each site are from primary archaeological reports.
  
  Symbols
  
  

  • P = pottery date
  • C-14 = radiocarbon date
  • T = dated text material
  • D = military destruction layer with historical date assigned
  • A = historically dated architectural style
  • E = earthquake destruction layer
  • E ? = probable earthquake destruction layer (significant rebuilding horizon without evidence of military conquest)

  
  
  Austin et. al. (2000)
  Time-stratigraphic correlation chart of Iron IIb excavations throughout an extensive region of Israel and Jordan from Austin et. al. (2000)

• Fig. 1 - Site Plan
  Fig. 1
  
  Site plan showing the squares excavated in 1987 and 1994, as well as the army trenches made in 1991.
  
  Ibrahim and van der Kooij (1997)
  from Ibrahim and van der Kooij (1997)

H. J. Franken, Excavations at Tell Deir 'Alia 1: A Stratigraphical and Analytical Study of the Early Iron Age Pottery (Documenta et Monumenta Orientis Antiqui 16), Leiden 1969

H. J. Franken (and J. Kalsbeek), Potters of a Medieval Village in the Jordan Valley: Excavations at Tell Deir 'Alia, a Medieval Tell, Tell Abu Gourdan, Jordan, Amsterdam 1975

H. J. Franken, Excavations at Tell Deir ‘Alla: The Late Bronze Age Sanctuary, Leuven 1992; ibid. (Reviews) LA 42 (1992), 389–392. — PEQ 129 (1997), 79–81

Aramaic Texts from Deir 'Alia (Documenta et Munumenta Orientis Antiqui 19, eds. J. Hoftijzer and G. van der Kooij), Lei den 1976

J.-A. Hackett, The Balaam Text from Deir 'Alia (Harvard Semitic Monographs 31), Chico, Calif. 1984

H. Rouillard, La Pericope de Balaam (Nombres 22-24): La Prose et /es "oracles" (Etudes Bibliques, N.S. 4), Paris 1985

Picking up the Threads ... A Continuing Review of Excavations at Deir 'Alia, Jordan (eds. G. van der Kooij and M. M. Ibrahim), Leiden, 1989

The Balaam Text from Deir 'Alia Re-Evaluated-Proceedings of the International Symposium Held at Leiden, 21-24 August 1989 (eds. J. Hoftijzer and G. Vander Kooij), Leiden 1991.

The Chronology of the Jordan Valley during the Middle and Late Bronze Ages: Pella, Tell Abu al-Kharaz and Tell Deir ‘Alla (Contributions to the Chronology of the Eastern Mediterranean; ed. P. Fischer), Wien (2006).