In the present excavation a short stretch of a finely-constructed sloping pavement was exposed. The basalt paving is interpreted as the ramp of an ancient bridge that crossed the original Jordan riverbed outlet from the mainland in the north to the site of el-Sinnabra/Tel Bet Yerah. On the basis of the archaeological data exposed in the present limited excavation and the evaluation of the relevant archaeological evidence from the tell, in conjunction with the contemporaneous historical sources and the accounts of early travelers and researchers, it is proposed that the bridge was constructed in the Umayyad period and functioned down to the Mamluk period (seventh to fifteenth century). Three squares (Sqs. 1–3; Fig. 1) were excavated in three adjacent areas.
 
Square 1 (4 × 6 m; Fig. 2). A small patch of a packed-earth and pebble living surface (L101) was exposed running up to a segment (c. 5 m long) of a stone wall (W104), oriented northwest–southeast. The wall was constructed of large- and medium-sized basalt stones bonded with heavy-duty mortar, and was extremely damaged by the pipe-laying activities. In the course of the excavation, the water table rose due to the winter rains and flooded the entire square, preventing further excavation. The stone size and the mortar of W104 resembled those of the long wall exposed in square 3 (W108; see below), about 50 m to the northwest, and the two walls had a compatible although not identical orientation. It seems therefore that these walls belonged to the same structure, although this remains conjectural.
 
Square 2 (1.6 × 2.0 m). A tiny patch of a packed-earth living surface (1.2 × 1.4 m) was uncovered in a very small probe excavated for the insertion of a Bezeq telecommunication box.
 
Square 3 (c. 6 × 11 m; Figs. 3, 4). A long stone wall (W108), delimiting the western side of a sloping surface constructed of a solid stone bedding and overlain by a basalt slab paving (L105, L106, L109), was uncovered. Wall 108 was exposed for a length of 10 m and had continued c. 1 m further northwest, with the latter section being damaged. It probably continued also in a southeasterly direction beyond the excavation limits. Its minimum width was 0.85–0.90 m and there was evidence for additional stone-built elements constructed integrally with the wall on its western side. These elements were damaged by the contractor and were furthermore rendered inaccessible by the raised water table that flooded the area to the west of the wall. Wall 108 had two faces of carefully laid, large, roughly-worked basalt stones, with smaller stones filling the intervening spaces. All the stones were fixed with thick, hard lime mortar containing many shell grits. The inner face of the wall was exposed in two small probes (L110, L111) that were dug into the stone bedding. These probes revealed that the long wall was extant for two courses and that its internal face was coated with a thick layer of hydraulic plaster that continued horizontally as a thick plastered floor surface (L111; Fig. 5) running directly below the stone bedding. The high water level prevented digging below this plastered surface and it is possible that there were additional constructional layers under L111.
The basalt stone bedding (L105/L109; 5.0 × 7.3 m), exposed in the area where the basalt slab paving was no longer extant, was laid down over L111 and consisted of two to three compact layers of medium to smallish basalt stones. Some of these stones were worn whilst others were more jagged (Fig. 6). The earth layer in which the stones were embedded was waterlogged due to the impervious plaster layer below it and contained broken shells and many water-worn small body sherds dated to the Roman and Byzantine periods. The uneven upper surface of L105/L109 sloped up in a southeasterly direction: 208.01 BSL at the northwestern end and 207.56 BSL at the southeastern end, with a rise of 0.45 m over a distance of 7.3 m. The basalt stone paving (L106) directly overlay the stone bedding at its northern lower end. The paving slabs that originally overlay the rest of the bedding must have been removed for secondary use after the bridge fell out of use. Here was uncovered a 3.2 m length of basalt slabs carefully laid in eight rows without mortar bonding (Fig. 7). The northeasternmost end row of slabs consisted of nine fine, large basalt slabs (0.20–0.25 × 0.3 × 0.7–0.8 m), laid lengthwise and forming an even straight edge, whilst all the other rows consisted of slightly smaller slabs of varying dimensions (c. 0.20–0.25 × 0.30–0.50 × 0.25–0.30 m) that were laid width-wise. As with the bedding to the south, the basalt paving sloped up in a southeasterly direction (208.25 BSL at the most northwestern end exposed to 207.82 BSL at the southeastern end, a rise of 0.43 m over a distance of 3.2 m). A very slight step up or ridge (c. 3 cm) was visible between the fifth and the sixth row.
A wall that ran parallel to W108, must have delimited the eastern side of this basalt paved surface and its underlying stone bedding, but it was not uncovered. Initially it was presumed to be beyond the excavation limits, and a trench was dug mechanically east of the extant basalt paving in order to trace it. The trench revealed that a sewage pipeline running alongside the northeastern border of the excavation at a depth of c. 4 m below ground level must have previously destroyed the eastern continuation of the basalt paving, its underlying bedding and the delimiting eastern wall. Consequently the original width of the paved ramp is not known, and it can only be noted that it was at least five meters wide.
The long wall, the bedding and the overlying paving are all part of a single construction interpreted as the northern ramp leading up to the bridge that was built over the ancient Jordan riverbed, connecting the mainland to the site of Tel Bet Yerah.
The small finds from the excavation that could contribute to determining the date of the construction of the bridge and its duration of use were limited to a small quantity of pottery sherds, a few glass fragments and five identifiable coins.
 
The Pottery. About 250 sherds in total were retrieved in the excavation, most of which were body sherds and almost all were small and water worn. No sherds were recovered directly on or between the basalt paving slabs (only a couple of which were slightly moved aside). The several sherds that were found on the stone bedding and in between the stones of L105 and L109, included a few Roman and Byzantine sherds, a few fragments of terracotta pipes and several glazed cream-ware sherds dated to the Umayyad and Abassid periods. The c. 40 small worn sherds that came from the small probe L111, dug down to the plaster layer below the stone bedding, were attributable to the Roman and Byzantine periods. A few Crusader and Mamluk sherds came from a very small area of accumulated debris located about 3 m north of the northern edge of the stone paving (L107; c. 1 × 2 m, designated square 4; the excavation of square 4 was initiated but discontinued due to flooding). The dearth of reasonably-sized diagnostic rim sherds in the excavation led to the decision to present a photograph of a selection of Early Islamic cream ware sherds from the stone bedding L105/109 (Fig. 8) and of a selection of Crusader and Mamluk glazed and hand-painted sherds from the very small area L107 (Fig. 9). On the basis of the admittedly small quantity of pottery it is possible to tentatively suggest that the bridge was constructed after the Roman and Byzantine periods, possibly in the Early Islamic period.
 
The Glass. Six glass fragments retrieved from the stone bedding L105 and L109 were identified as a wine glass, a stemmed lamp and a large vessel, all dating to the Byzantine period.
 
The Coins. Five identifiable coins were uncovered in the excavation. Three were found on and in the stone bedding L105/109: a coin of Constantius II dating to the mid-fourth century CE (IAA No. 143072) and two Umayyad coins (seventh–eighth century; IAA Nos. 143071, 143074). Two coins came from the accumulation in the partially excavated small area with the Crusader and Mamluk sherds L107: one dated to the Zangid Dynasty, to the reign of Mahmūd bin Zangī (1146–1174 CE; IAA No. 143075) and the other to the Ayyubid Dynasty, to the reign of el-Nāsir Yūsuf Salāh el-Dīn (1174 CE; IAA No. 143073).
 
The discovery of the bridge ramp begs the evaluation of additional available data that may illuminate the history of the bridge. What circumstances may have necessitated the construction of a bridge from the northern mainland to the tell? Who was the central authority that may have had the motivation and the means to undertake this building project? How long did the bridge function? The historical written sources and the archaeological data from the tell, from the Roman period and later, that may shed light on these questions are presented briefly below.
 
The Historical Sources. In the Early Roman period, Josephus records that in 67 CE Vespasian and his son Titus camped with three legions at Sennabris, located thirty furlongs (stadia) from Tiberias, in preparation for their attack of the Jewish rebels at Tiberias and Tarichaea (War III, 447). In the Byzantine period the site of Sinnabri is mentioned in the Talmud as located next to Bet Yerah (Jerusalem Talmud. Megilla I, I).
The Syrian geographer Yaqut el-Hamawi (1179–1229 CE) records that the Damascus-based Umayyad Caliph Mu’awiya Ibn Abu Sufyan (661–680 CE) built his winter residence (qasr) at el-Sinnabra, six kilometers south of Tabariyya/Tiberias, the capital of el-Urdunn/Jordan province (the Arabic sources are quoted in Mayer 1951). According to el-Baladouri, the palace was subsequently seasonally occupied by the Caliphs Marwan ibn el-Hakam (683–684 CE) and Abd el-Malik ibn Marwan (685–705 CE). El-Sinnabra was still under Umayyad control in 744 CE when the tribes of Urdunn pledged loyalty here to the Umayyad regime under Caliph Yazid III. Occupation of the palace at el-Sinnabra may have ceased with the fall of the Umayyad dynasty to the Abassids around 750 CE, but el-Sinnabra, by now possibly a village, seems still to have been settled in 979 CE when it was the meeting site of representatives of the Hamdanid and the Fatimid dynasties (Gil 1997:355).
In 1113 CE, el-Sinnabra—called Sane Boria in the Crusader records—was the site of the battle between Baldwin I of the First Crusade and the Ayyubids, the Frankish army camping west of the bridge. In July 1187 Saladin’s camp was located at el-Sinnabra en route to encountering the Crusaders at the Battle of Hattin. The village may not have survived beyond the Crusader and Ayyubid periods, as after the thirteenth century the literary sources refer specifically to the bridge of el-Sinnabra.
Already in 1106 CE the Russian pilgrim Abbot Daniel recorded the existence of two bridges spanning two Jordan river outlets:
“The Jordan river flows from the Sea of Tiberias in two streams which are three bow-shots apart, and which after a separation of about half a verst reunite as one river which is called Jordan. At the source, fish abound and there two stone bridges, very solidly built upon arches, through which the Jordan flows, span the two streams”.
 (Le Strange 1896).
From Daniel’s account it is patent that in the twelfth century CE the Jordan flowed in both the northern and the southern riverbeds, and that two bridges spanned the outlets from the Sea of Galilee. These bridges provided a northern and southern access from the mainland to the island of el-Sinnabra.
In the Mamluk period, Ibn Fasal Allah describes in the fourteenth century CE the junction between the northern and southern segments of the Jordan river at the bridge of el-Sinnabra. In the fifteenth century Qalqashandi designates the border between the provinces of Safed and the Jordan valley at the el-Sinnabra bridge below Tiberias, indicating that the bridge was an important landmark.
 
The Archaeological Evidence: Roman, Byzantine and Umayyad periods. Few archaeological finds from the Roman period have been uncovered on the tell. The site of the first century CE Roman camp at Sennabris has not been discovered yet. The Roman legions may have camped on the mainland immediately north of the northern Jordan river outlet and the tell, on the raised plain known as Ard el-Mallaha, today occupied by Hasar Kinneret (see Fig. 1). Moreover the short-term encampment may have left very limited evidence.
A tri-apsidal sixth century Byzantine church, exhibiting several building phases, was excavated at the northern end of the tell. After it had fallen out of use at the end of sixth or in the early seventh century, a modified building was rebuilt over it in the Early Islamic period (Delougaz and Haines 1960).
A fine fortified public-building complex excavated on the tell in the 1950s was originally identified as a Roman-Byzantine synagogue (Bar-Adon 1956:53–54), but has since correctly been reinterpreted as an Umayyad fortified qasr-palace residence with an adjacent bathhouse (Reich 1993; Whitcomb 2009:241–242). On the basis of this archaeological evidence, the identification of the Umayyad el-Sinnabra with the site of Tel Bet Yerah (Khirbat el-Karak), which was challenged in the past, is now firmly established. Recent excavations carried out beneath this palace floor retrieved coins that consolidated the date for the construction of the central building after 650 CE, and for the bathhouse attached to the outer palace wall at the end of the seventh century (Greenberg and Paz 2010; Area GB-T). In addition, the siphon pipelines exposed in the recent excavations carried out by the author on the west bank of the Jordan riverbed, are evidence of extensive engineering works carried out by the Umayyad rulers in order to supply fresh water to this palace (Alexandre 2013).
No buildings from the Crusader or Mamluk periods have been uncovered on the tell.
 
Discussion. The historical sources and the archaeological data do not provide support for the construction of a bridge to Tel Bet Yerah in the Roman or in the Byzantine periods. The main Roman road from Bet She’an/Scythopolis to Tiberias probably passed to the west of both the Jordan river outlets without accessing the tell itself.
By contrast, both the historical sources and the archaeological data provide strong contextual support for the construction of the bridge in the seventh century Umayyad period, when there was a central government that had both the impetus and the means to construct a bridge that would specifically connect el-Sinnabra/Tel Bet Yerah with the Umayyad district capital at Tabariyya/Tiberias. In the Umayyad period the well-travelled roads from el-Ramla, Beisan/Bet-She’an and Dimashq/Damascus accessed Tabariyya via el-Sinnabra, where the Caliph’s fortified seasonal palace was located (Fig. 10). In this period water may have flowed from the lake in both the northern and the southern Jordan riverbed outlets creating an island. Following the observation of an early nineteenth century British expedition that the mound “appears to have been artificially surrounded by water” (Irby and Mangles 1823:300), it is even possible to tentatively speculate that earthworks may actually have been carried out in the southern riverbed in order to create a moat surrounding and protecting the qasr.
The historical sources specifically cite the el-Sinnabra bridge as a landmark in the Ayyubid/Crusader and Mamlukperiods, from the eleventh to the fifteenth centuries. Some ancient bridges crossing the Jordan River south of the Sea of Galilee were observed by travellers and researchers in the nineteenth and early twentieth century, these bridges being an essential component of the major historical international west–east thoroughfares connecting Syria with the Mediterranean coast. The major bridges recorded were the ten-arched Jisr Umm el-Qanatir (the same bridge also sometimes called Jisr Semakh),and the Jisr es-Sidd, located respectively about a kilometer and two kilometres south of both the Jordan outlets, and Jisr el-Mujamieh, located south of the confluence of the Yarmuk with the Jordan (for the location of these bridges see Conder and Kitchener 1882: map sheet VI). The more local bridge at el-Sinnabra was mostly overlooked by the travellers, and it has been suggested that the major Jisr Umm el-Qanatir located c. 1 kilometer further downstream was designated the “el-Sinnabra bridge” of the medieval sources, rather than the bridges to the island itself (Ben-Arieh 1965:172, note 6).
However, the ruins of two bridges connecting the mainland with el-Sinnabra over the northern and southern Jordan outlets were still visible in the 1920s, and the Finnish pioneer archaeologist Aapeli Saarisalo provided specific details of the ruins of the northern bridge (although he mistakenly considered it to have been of Roman masonry):
“The line of the piers of the bridge runs from the corner of the kvutzah [Hasar Kinneret YA] … to the northern end of Khirbet Kerak … The bridge runs first 50 m towards the southeast, then 94 m to the east, then again 50 m towards the southeast until it disappears under the present highroad. There remains to be seen now only the foundations of the old arches. The hewn stones of the surface have all been carried away. Only here and there some detached stones remain and in the masonry there are three single ones still in place. The plaster is very hard. I have seen blocks, containing small stones cemented with this hard plaster, taken from this bridge and used as pavement stones in the neighbouring colony. There are traces of another bridge of the same type also in the southern end of the island, just beside the modern bridge that crosses the present outlet” (Saarisalo 1927:76–77).
The present excavation uncovered part of the northern ramp of this ancient Umayyad bridge, including a short section of a fine basalt paving. The paving was not visible in Saarisalo’s day as it was probably covered over by the swampy lands.
 
To conclude, the small section of the well-built basalt-paved ramp belongs to a bridge; it slopes upwards in a southeasterly direction. The limited finds from the excavation support a construction date for the bridge in the Umayyad period, and suggest that is continued functioning into the Mamluk period. Alongside these finds, the archaeological evidence from the tell, historical sources and the early researchers’ accounts, all lead to the understanding that the paved ramp was part of a bridge that was constructed in the Umayyad period to connect the mainland with the palace on the island of el-Sinnabra. This bridge continued in use down to the end of the Mamluk period, and with its southern counterpart, it was probably the el-Sinnabra bridge cited in the medieval sources.