The excavation mainly focused on the eastern area of the Franciscan compound. The unusually low level of the lake (c. 5 m under the upper red line) made it possible to reach the pavements of the pools (D1, D3, C3, E11, E22, M31) and to investigate the extended water supply and sewage network.
Soundings were carried out in the Water Tower (A1), to clarify how it functioned and its relationship to the aqueduct and to the water basins (A5, B5); in the aquarium(E27) and in the adjacent room (E28); in another room (E19) and conjoining spaces (E16–E18); due to the large amount of squared bricks (20×20 cm) of the bessales type and several fragments of tubuli found on the surface, as well as vertical spaces for tubuli inside the walls, Rooms E18 and E19 can now be interpreted as a caldarium of a bath; in the adjacent room (E31) where traces of burning on the basalt flagstones suggest the presence of a furnace opening (praefurnium); and in Areas G and I, where a residential quarter was identified.
The pottery classification of the site follows the typology set out byS. Loffreda (2008a–c) and a detailed report of the Magdala excavations appeared recently (De Luca 2010–2013).
The excavation of the pool (Fig. 2) confirmed the two building phases of Complex C that had already been identified in the 1970s. The bottom of the pool consists of well-worked chiseled basalt slabs; the southern part of the pavement was originally covered with a mosaic, of which a few traces survived. Near the spot where the change in the pavement is visible, there is also a seam in the masonry of the eastern and western walls, indicating a later restoration work. The basin of the pool was enlarged in the second building phase and a drainage channel, connecting the conduit (C14; Fig. 3), was opened in the southeastern corner. The iconography of the mosaic discovered in C6 (Fig. 3), the traces of white tesserae mosaic in C7 and C12, the running-water pool, the remains of earlier masonry reused as a bench in C10–C12 and the remains of clay bricks for a floating floor (suspensurae) in C7, suggest that Complex C was part of a thermal bath that was in use, with some adaptations, at least from the first half of the first century CE, before the First Revolt, until the fourth century CE.
This stepped pool was restored (Fig. 4). The earliest phase corresponds to a north–south wall (W345), which was partially dismantled at the level of the pavement, although four higher courses were preserved and covered by the steps.
The paving slabs (L346) were laid contemporary with the construction of W345, whereas the northern wall (W348) and the western sidewall (W347) of the pool partially covered it and therefore, belong to the second building phase.
Hence, the pool was enlarged in the second phase toward the north with the construction of W348 and restricted on the east with W347.
Moreover, in a later stage the uppermost three rows of W347 were rebuilt with well-hewn squared blocks, arranged in the headers and stretchers building technique. The findings from the pool, which include a bronze tool (specillum), unguentaria and a bone hairpin, attest to its use, with successive adaptations, from the Late Hellenistic to the Late Roman periods, when the caldarium and tepidarium in E19 underwent restoration.
Columned Building D1
Several soundings were carried out in the controversial columned building D1, misinterpreted as a synagogue, to clarify its function and its connection with other buildings.
It was possible to attest that the original use of the structure was in some way connected to water. In a later phase, it was transformed into a water reservoir, also by raising the pavement, probably for manufacturing activities. The water was supplied via some rectangular openings inclined toward the interior of the building; the best preserved are visible in the southern wall (W1200) that belonged to the first phase and was built of perfectly smoothed blocks in courses of varying height. The western sidewall (W1199) was equipped with similar openings, which at a later stage were roughly enlarged. In the earliest phase, the supplying-drainage system consisted of a sluice gate that regulated the water flow toward the discharge conduit D8.
Stepped Pool E2-E7
Although the absence of certain pavements and the plundering of masonry make it difficult to “read” the layout of the rooms and interpret their function, the soundings in E2 (Fig. 4) offered significant contributions. Remains of stone pavements with traces of mortar were unearthed in the central sector at an elevation of -207.89 m; in the southern sector traces of mosaic bedding were visible along with several mid-sized tesserae in black, white and red hues; in the northern sector, an older lower pavement was exposed; at the same level to the east was a small covered channel, overlooking a six-step basin that broadensup at the fourth step into a rectangular basin (1.0×1.8 m). The pavement was partially covered by a partition wall (W985), aligned north–south.The southern side of this basin (W986) was built of worked stones arranged in regular rows and connected with W985, while the northern side (W988), corresponding to the southern wall of E6, was roughly built of un-worked polygonal stones arranged in irregular courses. Wall 988 is later than the flight of stairs because its foundations were adapted to the steps, reducing their size and precluding their use. Therefore, it is possible to imagine an original basin with steps, oriented west–east, which was in use before the middle of the first century CE, judging by the Late Hellenistic–Early Roman assemblage from the fill of grayish clayey mud. Later, the basin was reduced in size and no longer in use; eventually, it was completely obliterated with the resetting of the walls during the Middle Roman period.
This room (Fig. 4) consists of two levels. The lower one is the proper stepped pool and the upper level could have been the undressing room (apodyterium). To the east, a stone plastered bench runs along a narrow corridor that widens out toward the south into a plastered pavement with a small basin.
The uppermost fill layers in the pool were excavated down to the fourth step of the staircase in 1975 by V. Corbo and S. Loffreda; these consisted of an abandonment level that contained findings from the Byzantine to the ‘Abbasid periods (sixth–ninth centuries CE).
Beneath a debris buildup layer, a level of collapse was identified, overlain with several mixed potsherds and a quantity of square bricks for suspensurae and fragments of clay tubuli, coming from the adjacent E19. The collapse, which extends from southwest to northeast, consisted of a covering basalt slab, a well-hewn quadrangular block, several building blocks and a basalt column. Moreover, the pertaining disruption layer contained fragments of colored plaster, mortar and several shattered glass panes.
The collapse, which is ascribed to an upper floor, covered a levelof abandonment that contained a large quantity of common pottery, e.g., amphorae, cooking pots, and lekythoi, dating up to the fourth century CE, while the associated coins extend no further than the third century CE.
Several articles of toilette, such as bone hairpins, a bronze handle for a glass aryballos, a bronze spoon, a gold granulated earring, complete clay unguentaria and a wooden comb, were found on the pavement, in association with coins that provide a chronological range extending from the first century BCE to Caracalla (198–217 CE).
The earliest building phase corresponds to the remains of a north–south wall, razed to floor level, which was incorporated in the lowest step and in the masonry of the basin’s southern side. The western wall (W995; preserved height 2.44 m) clearly shows two different building techniques: the upper courses are of well-worked squared blocks, alternately arranged lengthwise and widthwise, placed on top of the remains of a pre-existing structure, built of roughly worked calcareous stones.
The flight of steps leading to the pool, which shows a gradual narrowing toward the west, was conditioned by the presence of some earlier structures.
The pavement of the pool (-210.03 m) consisted of seven squared masonry blocks; in the western sector, two large squared basalt blocks were laid down forming a sort of step that led to a deep basin (0.62×1.62 m; -210.35 m), with a non-paved bottom.
The presence of two large basalt corbels, in situ, on top of the western wall, leads us to offer the existence of a small balcony.
Few minor traces of hydraulic and colored plaster are still visible, in situ, on the eastern side of the basin and, together with the fragments collected in the mud that filled the pool, suggest that the sides were decorated.
Water flows into the pool from a natural source, via two underground trickles that enter through two openings in the western wall, where some remains of limestone bedrock are preserved and incorporated in the well-dressed masonry. Before reaching the pool, the water was apparently filtered in a stepped reservoir (E7) excavated in the eastern side of E2-E7. The pool was connected to the water collector (E12) through a stone squared conduit, provided with an outstanding mensula (small table), which flows in the southeastern corner of the southern wall (W997).
Water Collectors E12, E21
Room E12 (Fig. 4) consists of two levels: the upper one has a pavement of flagstones covered with mosaics, now totally destroyed, and the lower level is a large underground water channel. This channel is unfortunately interrupted in the west by a blockage of large blocks and its top is blocked by the collapse of the original covering. The northern edge of this water conveyor was built as a wall and topped with a U-shaped pipe of hydraulic mortar, which runs west–east close to the mouth of the E11 conduit and then enters into the western wall of E19 (caldarium).
Several fragments of clay unguentaria, fragments of colored plaster and a second/third century coin were found atop the upper pavement. The mud fill of the lower channel contained early Roman potsherds, a complete fritillus-type ungentarium, in association with a coin of Pilatus and fragments of marmoridea (plaster with marble appearance).
E12 was connected in a later phase to Water Collector E21, through an additional duct along the western wall (W998) of the caldarium, and though E13 and E12, which seem to have been a kind of water collector, as attested by the different branches coming out from it. E21 is connected to E13 through an opening in the western side, to the caldarium in E19 through a covered channel penetrating below its southern wall (W525), and into the main conduit E20.
This is a large underground basin (Fig. 3) with a fountain (nymphaeum?), divided into two nearly rectangular parts by a segmental arch; its entrance is presently placed south toward the Channel E20, whose covering at this point is stepped. The perimeter wall of E22 is built of dressedstones in alternate courses.
At least two building phases were identified in the complex design of the structures.
The good workmanship of the arch, supporting a double line of elongated basalt covering blocks, consists of worked stones, well finished and arranged in rows in the intrados and in the southern face.
From the southern entrance hall of the pool, two steps under the span of the arch lead to a basin, which has a plastered bench against its western side. The same plaster was also applied to the surface of the pillars of the arch and on the northern wall of the basin, indicating the same building phase. Here, masonry remains of an earlier stage were identified in the northwestern corner. The masonry of the basin’s eastern side (width 1.1 m, height 2.10 m), although less accurate than the western side, carries the thrust of the arch. This wall was partially dismantled in the southern room to build a covered conduit that runs in the southern sector of the adjacent room (E22a). The latter, although altered, was connected to the water network as shown by the presence of three rectangular outlets now visible in the northern and eastern sides.
In the northern wall of E22, water flows out from the spout of a fountain with a Y- shaped clay tube (fistula; -209.30 m). The jet of water was forced out due to the pressure of a cascade vertical system. The water of this fountain came from a covered conduit that was partially encompassed in the southern wall of the caldarium with suspensurae (E19).
This fountain, which turned the pool into a kind of nymphaeum, might have been fed by the U-shaped Channel in E12.
The pavement of the basin (average elevation -208.88 m), lightly sloping to east, consists of irregular basalt blocks and rubble that were originally coated with hydraulic plaster.
Concerning stratigraphy, three strata were identified.
Stratum III (-209.60– -209.80 m): a level of abandonment, consisting of an accumulation of very fine and dusty clayey soil with few ceramic inclusions, dating from the Late Roman to the Islamic periods.
Stratum II: A hard packed conglomerate layer of non-worked stones, including cobblestones, pebbles, limestone and basalt flakes, bonded with light mortar, was discovered only in the entrance hall. This layer, which was bound by a barrier of blocks and stones before the arch, closed up the earlier entrance from south and raised the pavement of the entrance hallup to the level of the most recent entrance. The findings give a terminus post quem of the Late Roman period.
A layer of waste, containing fragments of ochre, reddish, pearl-white, and turquoise green painted plaster that was probably used for decoration with stripes and floral patterns, was identified in the basin. In addition, a few pieces of egg-and-dart stucco were collected. The same context yielded a group of rare wooden findings, exceptionally preserved in the water-saturated mud, including pieces of bars, planks and muntins with nails, joints and wedges. These wooden pieces, which may have been parts of a trellis for false ceilings, were placed in overlapping horizontal strips that apparently crossed and ended a little below in long and very thin panels. A mat of intertwined canes and vegetable materials, such as vine-leaves, thin canes, olive branches and palm, bound by lime mortar has been found in several places between woods and frame. The mortar contained lake sand, small shells, fruit stones and vegetable frustules.
: The collapse of the wooden structure covered a layer of abandonment (thickness 0.5–0.6 m) where several findings were found, including glass aryballoi, soft limestone vessels, a fish-hook, cramp-irons, blades, nails, faunal remains and a remarkable assemblage of pottery dating from the end of the Hellenistic to the end of the Early Roman periods. Among the forms to be noted are Kefar H
ananya type globular cooking pots, Pent10 (Loffreda 2008a
:184–185), Pent11 (Loffreda 2008a
:185–186), and Pent12 (Loffreda 2008a
:186–187), cooking bowls, Teg12 (Loffreda 2008a
:204–205), and Teg14 (Loffreda 2008a:206–207), two intact unguentaria, amphorae, Anf12 (Loffreda 2008a
:125–126), and Anf13 (Loffreda 2008a
:126–127), stone vessels, and Herodian oil lamps (Loffreda 2008a
:42–45). Moreover, the mud fill contained an extraordinary group of wooden vessels, including a plate, a small cup and two rounded cups, which have been transferred, together with the comb from E11, to Pisa to be consolidated and restored at the Centro di Restauro del Legno Bagnato del Cantiere delle Navi Antiche.
Therefore, the homogeneous pottery assemblage from this stratum testifies to a phase of use preceding the sudden destruction that occurred before the end of the first century CE and is ascribed to the violent conquest of Magdala by Titus and Vespasianus.
This is a rectangular basin (1.20×2.25 m), whose northern, eastern and southern walls (height 2.1 m) were built of irregular masonry stones and placed against the last southern pillars of the aqueduct. The western wall is formed by the eastern side of the pillars and their intercolumniation.
A basalt rectangular corbel jutting out 0.3 m from the eastern side toward the interior of the basin seems to indicate that the basin, which may have served as a water reservoir, was roofed. The use of hydraulic plaster to coat the inner wallsconfirms that the function of M31 was connected with water.
The excavation in the basin exposed beneath a hard packed limylayerwith few Abbassid-Ayyubid potsherds, a collapse of long basalt slabs laying along the short side of the basin. The collapse (thickness 0.6–0.8 m) covered a layer of moist grayish and coarse soil, containing a large quantity of yellowish light-clay pottery. The assemblage includes more than fifty cylindrical mugs (height 20 cm) with a pinch body along the upper quarter and with a high neck and a lightly everted lip. This clay type is known from Early Islamic contexts and is ascribed to the Khirbet el-Mefjar production.
A small handle is joined to the bottom of the mug and due to its position it appears not to have any holding function, but could have been used to hang the mugupside down.
The tall mugs could have equipped a kind of a wooden wheel for lifting water—a noria; considering the contiguity between M31 and the Late Roman-Byzantine aqueduct, this is the most reasonable hypothesis.
Wall 317 (Figs 4, 5: View 1-1) has been revealed along the eastern closing wall of the large quadriporticus F (FC). Built of polygonal stones in horizontal rows with rock filling, it was bound with very tenacious mortar; the surface was coated with a lime layer, analogous to that of FC. Excavations to the west of FC have shown that a lower foundation (W332) had existed prior to FC (Fig. 5: View 2-2).
The walls of the structure were thus over 3.6 m thick. Exploration in the east part of Channel E20 allowed the analysis of masonry in the northern corner of the quadriporticus, as built of large lithic elements and the revealing of the FC’s foundations that were built of large stones with a central boss and drafted margins, which protruded c. 0.5 m from the face of the FC. This structure was covered by the construction of W317, which also caused the flattening of the bosses.
The quadriporticus, judging by the material discovered in the deep soundings, should be assigned to the Late Hellenistic period, namely the Hasmonean era.
The area west of W317 was affected by a collapse of blocks, lying from west to east and ascribed to the collapse of the FC’s upper structure.
The purpose of Sounding F18 (see Fig. 5; View 2-2) was to understand the function of W317. The following layers (see Sarti et al. 2013) have been verified in the sounding, below the surface:
I) Whitish beaten earth and rubble (L325), leveled over the upper horizon where, amongst the finds, a Constantius II coin (350–361 CE) was found;
II) A collapse layer (L329) below Layer I is composed of worked stones mostly dressed, rubble and several fragments of colored plaster, especially in proximity to the wall. The coins retrieved from the collapse dated from Alexander Jannaeus to Herod the Great;
III) A layer of lime sand with pebbles and shell remains, which is characterized by the absence of any ceramic or other anthropogenic remains. The sediments’ surfaces follow the west–east direction in the south and north sections, and south–north direction in the east section;
IV) A cobble pavement (L331) of various sized natural pebbles placed in correspondence with W317. Potsherds from the Hellenistic–Early Roman periods were discovered in the ballast, as well as two coins from the first century CE. The ballast’s surface, slanting from west to east, was partially affected by the collapse of roughly-hewn lime blocks before being buried;
V) A dark clayey layer, containing pre-Roman material (L401), before both L331 and apparently W317 were placed.
Wall 317’s surface was coated with two different plaster layers; the lower was rough plaster with grayish hydraulic mortars and the upper consisted of smooth lime.
Many pieces of colored plaster in red, yellow and greenish hues were uncovered in the collapse (L329), and the façade’s elevation of W317 seems to have been decorated with plaster panels imitating marble.
Some parallelepiped-shaped basalt blocks (length 0.7 m), which were incorporated in the stonework of W317, bore holes (diam. 14 cm).The degree of smoothness in the holes shows wear, probably due to the extensive friction with metal rings, ropes or chains. These blocks were used for the mooring of boats; several comparisons can be mentioned, e.g., in Moregine (suburb of Pompei), Caesarea Maritima, Rujm el Bahr, Leptis Magna, and Portus. Two of the blocks have been found broken, but four others are still in situ, in the pier (W317).
This arrangement of the wharf, dating back to the Early Roman period, replaces an earlier anchorage that is evinced by a big mooring stone inserted in the eastern side of the perimeter wall (W282) of the quadriporticus that encompassed the Roman pier. A sixth mooring stone in situ is found in the southeastern corner of the adjacent impressive building (E32, E33, E35). The masonry preserved in this corner consists of ashlars with bosses and dressed margins, set in alternating rows and typical of the Hasmonean period.
A casemate buttresses the substructure of the eastern wall (W355) of this monumental building, to the north of the quadriporticus; more than 10 m of W355 had been exposed to date toward the north (Fig. 6).
We can reasonably presume that this edifice was a Hasmonean lookout tower, similar to the ones around the Dead Sea, e.g., Rujm el-Bahr, Khirbat Mazin and Qasr et-Turabeh.
The quay seems to date to the Early Roman period and, probably, was no longer in use after the Late Roman period because of a natural blockage of silt, limestone and sand, as were identified in Layers I–V (see above).
What needs to be still clarified is the chronological and functional relationship between this dock and the breakwater, discovered in 1960 by the “Link Expedition” (Fritsch and Ben-Dor 1961
), and is now buried under sand. It should be noted that the breakwater is chronologically different and located, at a distance from our dock toward the lake, at a different altitude, and is now buried under accumulation of lake-water sediments.
The Area South of W317
A trench (2×8 m) at the southeastern corner of Quadriporticus F made it possible to identify a whitish layer (thickness in excess of 1 m), severely disturbed by roots and almost entirely consisting of extremely compact limy material (L351). This layer yielded a large number of bricks (imbrices and tegulae) together with Late Byzantine potsherds and several fragments of glass. The potsherds included an African terra sigillata dish decorated with the emperor, dressed in tunic and holding a long scepter between two small palms in his right hand. This type corresponds to Form ARS 104C (variant of Stamp 230; Hayes 1972:104, C23) that, according to Hayes, dates back to 550–625 CE. The 47 coins recovered from this layerare severely corroded, mainly dating to the Byzantine period, with a bronze 3-solidi exagium of the fourth–fifth centuries CE.
Five meters from the southeastern edge of FC, Layer 351 covered a flight of three stairs, built in W317 and contemporary with it; therefore, it should be considered as part of the port structure, which continues southward in the direction of Area M, beyond the boundary of the quadriporticus. The stairs led to the landing area and from there, maybe through arcades, of which elements have been discovered in the collapse, it might have been possible to access the eastern porticus.