Four phases, not evenly preserved, were identified in Area H, corresponding to four chronological phases:
Phase IV (Islamic). The entire area is covered with a horizontal surface layer (average elevation –205.60 m), moderately sloping from east to west, which is ascribed to a modern phase (L0, L4, L7; Figs. 3–5). Materials from the Islamic period, ascribed to this layer and collected in several areas of the excavation, bear witness to a partial reoccupation of the site, with the reusing of building materials and the adaptation of many existing waterworks.
Phase III (Middle/Late Roman). Beneath the Phase IV layer, a level of collapse, best preserved in H3, was identified. It was partially disturbed in H1 and H2, probably after 1940, according to the findings.
The collapse was preserved in situ and studied from a geological standpoint. As a terminus ante quem for this phase, the total absence of any Byzantine potsherds from H1, H2 and H3 should be considered.
Phase II (Early Roman). Some loci that were partially unearthed under the collapses can be assigned to the Early Roman period.
Phase I (Late Hellenistic). Several Late Hellenistic structures, clearly dated by findings, were identified only in the layer under the collapse of Early Roman walls in the southeastern part of Field H1.
The structures unearthed in the collapse layer are organized following the known plan of the city and were aligned with the paved road (L5), running for 16 m in an east–west direction in Field H2. The walls belong to different chronological phases and the assemblages associated with them provide a temporal range extending from the Late Hellenistic, with a level of use dating to the reign of Herod, to the end of third–beginning of fourth centuries CE.
Walls 16, 17, 57 and 93 enclose a square room (5×5 m), in which the remains of collapse are visible (L91, L94). The collapse was overlain with a solid layer of chalk, lime, and pottery fragments (L86). This whitish pressed floor layer occupied the entire eastern portion of the field. The potsherds contained in it included the so-called Kefar Hananya ware, dating to the first–third centuries CE, mixed with a few oil-lamp fragments of Types Luc1, 2 (50 BCE–50 CE; Loffreda 2008a:41–42) and Luc3 (Middle Roman; Loffreda 2008a:45–46).
On top of Layer 86, walls were rebuilt maintaining the same orientation of the earlier structures. This later construction dates back to the Middle Roman period, while the areas were still in use, up to the Late Roman period, at which time there is evidence of collapsed walls and consequent abandonment. Walls 46, 87, 89, 92, and the upper courses of W16 whose construction technique used natural polygonal blocks, differed from that of the lower parts, which belonged to the same chronological stage as the hard-packed Floor 86.
A similar stratigraphical sequence was verified in the western portion of the field. The southwestern oriented Wall 14, which is built of roughly hewn polygonal blocks laid out in two parallel rows and forming a double face, had two building phases; its foundations were connected with those of the three walls orthogonal to it: W16, W2 and W104. The hard-pressed soil layer (L106) must be considered a floor that was associated with the latest building phase of W14.
Layer 106 covered a stone cluster (L103), which contained Late Roman pottery, mainly amphorae, a fragment of a basalt tripod mortar, and a triangular one-hole basalt anchor of the type commonly used in antiquity on the shore of Lake Kinneret.
The upper two courses (height 0.5 m) of W2, built of two stone rows, are assigned to the same Late Roman phase. Wall 15 (length 6 m), oriented southeast-northwest, was exposed in the northwestern section of the field. The wall is built of roughly dressed blocks, sometimes with one face smoothed, placed in regular courses and bound with light mortar mixed with small-size debris; it was rebuilt with a slight rotation of the axis, as can be seen next to its conjunction with the earlier W15.
Next to W45, a hard layer of pressed soil (L110), similar to Layer 86, was found; upon it was an almost entirely restorable flat pan of Type Teg17 (Loffreda 2008a:209), dating from mid-third–late fourth centuries CE. Wall 45 has the same alignment as W55 in Field H2, where it is reasonable to hypothesize the entrance to the complex, which overlooked the paved Road 5.
The collapse (L91, L94) is composed of well-hewn rectangular blocks; the manner in which they are placed suggests that they originated from the elevation of the earliest phase of Walls 93 and 16.
Under the collapse, a basalt rounded fishing-net weight was found, as well as a cluster of 54 prutot, mainly datable to Hasmonean times, although the latest of which pertains to the prefects Marcus Coponius or Ambibulus (6–9 CE or 9–12 CE). This data is consistent with the chronology of the pottery from the same context, which includes oil lamps of the Karm er-Ras type (Loffreda 2008a:41) and thick-walled amphorae, leading us to date the collapse of this early building after the first decade of the first century CE.
Beneath Floor Layer 106, the collapse (L103, L105) occupied the space that was separated by W104. This wall, which is contemporary with the earliest phase of W14, has a basalt grinding stone (Olynthus mill) reused as a threshold. Under the collapse and on top of a compact earth floor (L102), Early Roman pottery, including Herodian oil lamps and a portion of a limestone hydria, as well as a remarkable group of nails, parts of a latch and a bronze artifact (specillum), was found.
To this chronological stage pertains the small rectangular room enclosed within Walls 2, 14, 42 and 104. It must have been originally double in size, reaching the older phase of W15 as its northern boundary. In this case, Walls 41 and 42 must be considered a later superimposition.
A foundation row, protruding 0.3 m out of the upper courses, was identified along the southern face of W15’s western portion; it represents the remains of an earlier construction phase. A similar alignment, projecting only 0.2 m from the wall, was exposed in a small trench (1.3×1.5×1.5 m) along the southern face of W15’s eastern part. This sounding made it possible to verify the presence of a two-course wall (W44), below a filling of earth (thickness 0.3 m) under the northern portion of W2; both walls have the same orientation. Moreover, when a whitish dump layer (L39; thickness 0.8 m) was removed, a pressed-earth floor layer (L98) was reached, overlain with a globular Kefar Hananya cooking-pot of Type Pent10 (Loffreda 2008a:184–185), dating to the first century BCE–first century CE, and associated with three Hasmonean coins. On the contrary, the material from Dump Layer 39 included traces of burning, animal bones, half shells, pottery from the Early to Late Roman periods, glass fragments, including a ribbed bowl (coppa costolata) of Isings Type 3 (Isings 1957:17–21), and coins from the first–third centuries CE.
A usage phase attributed to this period is identified by the material found on top of a pressed-earth floor section (L96), bound by W97 to its southwest. The pottery assemblage included a complete globular cooking pot with outward-curving long neck and a partially preserved amphora of Type Anf3 (Loffreda 2008a:119–120), as well as three coins of Alexander Jannaeus.
The building remains in this field emerged immediately below the surface; generally, they belong to two different periods that are not always easily distinguished, due to their continuity. Moreover, some walls show signs of heavy subsequent reworking. The chronological span of the finds in this field ranges from the Hasmonean to the Late Roman periods, with a predominant third-century occupation phase, as well as several Herodian contexts.
It should be noted that the northeastern area of the field presents an altered stratigraphical scenario, as inferred from the presence of an artificial pit that was cut probably due to agricultural needs in modern times.
The Road (L5)
A basalt-paved road (length 16 m, width c. 3 m), aligned east–west, was exposed in the center of Field H2. The road presents a cambered surface, with the sides sloping to the east and scaling the sidewalks. The upper surface of the road (summum dorsum) consists of paving stones in a tight arrangement, bedded into a layer of rubble and mortar. At an unspecified time, the road was heavily restored with a further layer of hard and solid cement material mixed with potsherds. This intervention was implemented to integrate the areas plundered for building material, probably to rehabilitate the water supply network, and the original convexity of the road surface. The axis of Road 5 appears to have determined the buildings’ orientation, both in Field H3 and in Field H2. The road’s orientation is thus consistent with the urban topographic arrangement noticeable in the east quarter of the city. By extending the course toward this direction, the new paved road joins the north extension of the cardo maximus (Magdala 2008, Fig. 2) and qualifies as one of the decumani of the city. On the opposite side, the road leads toward the pass of Wadi Hamam, which was the main road linking the lake to the western Galilee.
The street and the structures overlooking it were covered with whitish pressed soil layer (L6), not assignable to any building. This layer, which is dated by an Ayyubid fals (twelfth–thirteenth centuries CE), was in turn covered with another surface (L4).
In the gaps between the road’s polygonal basalt flagstones, two coins of Alexander Jannaeus minted in Jerusalem, were found. The ceramic finds indicating the usage phase of the paved road are, on the whole, Late Roman.
Concerning the buildings adjacent to the road, there is only a collapse (L12) connected to a wall (W55), which should have reached the northwestern structures of Field H1.
The Conduit (W60, W61, W116)
The conduit (width 0.6 m, depth in excess of 1 m) was revealed east of the road. It consists of solid lateral walls and interior parapets for the covering stones, mainly of basalt lithic elements.
This portion of the conduit seems to have served as an inspection chamber where different branches of the water network joined. The fill in the conduit contained several animal bones, a few Early Roman potsherds and two Herodian coins. Both the conduit and the line of squared blocks (W59), which is probably a smaller channel connected to the conduit, were overlain with a debris layer (L54), full of Late Roman potsherds.
Two rooms, formed by Walls 25–27 and 29, were identified in the west area of the field. The two rooms, which are linked through the door in W26, were affected by a collapse that also involved the passageway between them. The collapse is predominantly composed of well-cut and refined basalt blocks, typical of fair-faced masonry. The disruption layer (L24) containing the collapse (L13) can be distinguished by the whitish color, the extreme compactness, the limy composition and the presence of numerous ceramic inclusions from the Middle to Late Roman periods.
The excavation did not yet reach the floor in this room; yet, a small trench close to W27, where a monolithic threshold is still in situ, made it possible to collect Late Hellenistic/Early Roman finds, including limestone vessels, Herodian oil lamps and cooking pots, which are clearly indicative of the levels of use prior to the collapse.
To the west of W25, a large room (at least 3×4 m) with remains of a collapse was exposed. The finds from the collapse go back to the Middle/Late Roman period.
At the moment, it still needs to be confirmed whether these areas are connected to the ones uncovered in the southwestern sector of Field H3; it can be assumed that they are by calculating their volumetric development.
The collapse (L10) to the west of W31 is related to a rectangular room, of which only the foundations of the façade wall were fully recovered. Within Collapse 10 was an alignment of blocks that belonged to an arch or a pillar, which collapsed from west to east. The collapse layer contained Late Roman potsherds, coins dating from the Hasmonean period to the third century CE, and part of rectangular-shapedmillstone of the so-called “hopper-lever” Olynthous type. The floor has not yet been reached, but it should correspond to the “footprint” of the millstone.
Below the surface in this area, the whole eastern and central sectors appeared to be sealed by a solid deposit of crushed stone and gravel. Once removed, a hard-packed collapse level was revealed. About ten rooms can be discerned in a fairly uniform orientation and characterized by alternating stones and ashlar masonry.
The walls in this field clearly show that they were restored, reinforced or doubled in different phases.
This is the main wall (length 10.7 m, width 0.6 m) in Field H3, occupying the central sector. The wall, preserved at least four courses high, is built of two regular rows of well-hewn basalt blocks, set as headers and stretchers, with a core of lime-based mortar. At a later stage, the wall was reinforced by placing a row of hewn and un-worked stones against its western side. On W8’s eastern side, it is joined by W66, W48 and perhaps also W9, all built in the same manner. It is reasonable to assume that the rooms on the east and west sides of W8 were parts of the same building. A hoard of 21 coins was discovered in W8. The coins, minted in the neighboring autonomous cities, including Gadara, Abila, Hippos and Nisa Schitopolis, can be dated from the first half of the third century CE; however, countermarks on some of the coins indicate that they were reissued. At present, the discovery context of the coins represents a terminus post quem for the destruction date of the complex. The collapse dates back to after the Diocletian era (284–305 CE), as other findings seem to confirm.
Area east of W8
Three rooms, linked to each other, were exposed. The northern room is enclosed within Walls 8, 47 and 48 and contains a collapse layer (L50); the manner in which the stones are set indicates that it should be ascribed to the collapse of a double arch, which happened when the doorway in W48 was already obstructed.
Two construction phases were identified, at least in Walls 48 and 79, which were originally two pillars, with the space filled up between them.
The central room was occupied by a collapse layer (L11) and the pottery ascribed to this layer, as well as to the level of use, dates back to the Late Roman period. A niche in W47, whose substructure is only preserved, was covered by Layer 11 and contained an amphora of Type Anf14 (Loffreda 2008a:127) from the Middle Roman period. A similar niche was exposed in the eastern side of the same wall (Fig. 8, W47) and in the southern W9. The building was partially exposed in the southeastern corner of the field; at a later stage, in the Early Islamic period, it was founded on the southern room’s structure, as the pottery from this context confirms. This late construction probably used plundered masonry stones from the upper courses of W47.
Beneath the collapse layer (L67), the remains of an earlier building were exposed.
Area west of W8
A T-shaped pillar (W18) built of basalt blocks arranged in headers and stretchers and bound together with lime based tenacious mortar, was exposed in the center of this area. The T-shaped pillar suggests the presence of corresponding structures in line with its southern, eastern and western sides. In fact, a structure (W85) was found in the same axis, 1 m toward the west; to the east, at a distance of 1.1 m, the “head” of a platform (L72) is laid and to the south, a pillar (W22) protruding 0.4 m out of the wall, was revealed. The platform may have been a base of a stairway that accessed the higher floors.
Between W18 and Platform 72 was a collapse layer (L73) that yielded a remarkable pottery assemblage, which can be dated from the late first century CE to the mid-third century CE.
Exactly in the middle of the space between W18 and W22, a column (W19) has been located. Pillars and columns must have held up a system of vaults, as confirmed by the number of dressed and, in some cases, trapezoidal stones and by the orderly distribution of the collapse, whose axis is the line of W18, W19 and W22.
Wall 23 was exposed on the same line as W22. The wall was originally a couple of pillars, which in a later phase were connected with a rough patch of un-worked stones in an inferior technique. Between W22 and W8, the foundation row of an earlier wall (W70) was exposed.
The collapse layer (L21) contained a large quantity of Middle Roman pottery, as well as coins from Herod’s reign to the third century CE.
Next to the western pillar in W23 was a lead weight with the relief image of the Phoenician goddess Tanit (Callegher 2008). A few examples of this kind of weight are known from well-defined archaeological contexts. At least two other weights come from Magdala; one from the area of the quay dating to a later phase, and the other is from the antiquarian market.