The Combined Caesarea Expeditions, 1996 Summer Season

by: Kenneth G. Holum and Avner Raban


The Combined Caesarea Expeditions summer 1996 season extended from May 26 through July 25 on land and underwater. The project directors are Avner Raban and Joseph Patrich of the University of Haifa, and Kenneth G. Holum of the University of Maryland. Participating institutions in 1996 were the University of Haifa, the University of Maryland, Trinity College, and the University of Oklahoma, along with a large number of volunteers and staff from the U.S., Israel, Canada, and other countries.(1)   In addition to the summer season, the Israeli directors have conducted excavations on land throughout the year, within the framework of the Caesarea Tourist Development Project. The year-round excavations continued during 1996, but on a smaller scale than from 1992 through the end of 1995.

As is well known, Caesarea (fig. 1) is alarge urban site, located on the Mediterranean coast of Israel. King Herod the Great founded the city at the end of the 1st century B.C.E. on the site of an earlier Hellenistic town named Straton's Tower, and Herod's foundation flourished through the Roman, Byzantine, Islamic,and Crusader periods, until destroyed in 1265 C.E.  In all periods,Caesarea functioned as a commercial, industrial, administrative, and residential center for the neighboring hinterland, and also as a seaport, especially in the century or two after Herod built the artificial harbor Sebastos adjacent to the city.

Again in 1996, a team of diving archaeologists continued exploration of Caesarea's harbor during the summer season, concentrating as in 1995 on the intermediate harbor basin and the division between it and the inner basin, now mostly landlocked (fig. 2). The divers excavated six probes in Area QN, in the western part of the intermediate basin, to study the stratigraphy of harbor debris on the sea floor and the topography of the underlying bedrock. Scattered building stones of various sizes appeared in the area, as well as large chunks of kurkar (the local calcium-cemented sandstone), but there were no sherds, other artifacts, or the fine-grained sediments that usually characterize the bottom in harbor basins. It was as if the area had been thoroughly dredged, either by later inhabitants of Caesarea or by the rip current. Dredging may have occurred as late as the Crusader period (12th-13th centuries), or this stratigraphic lacuna may represent the suspected dredging of this anchorage during the Early Islamic era, sometime before the late 9th century C.E.   The topography of the bedrock suggests a generally uniform elevation, now 4.6-5.1 m. below MSL, but with an underwater gorge sloping toward the southwest to a maximum depth of almost 7.0 m. against the Herodian quay in Area Q.

A second underwater objective in 1996 was Area T, in the eastern part of the modern anchorage adjacent to the bathing beach.  Here the most prominent man-made feature is a wide sea-wall incorporated into the seaward fortification of the Crusader city at the shore on the north and crossing the harbor almost to the south side, where there are collapsed remains of a rectangular tower, now submerged.  Here the divers extended east-west trench TN1, begun in 1993, to expose both sides of the sea-wall . On the lee side the fill consisted of four layers:

a. Surface material of mixed, circulated, and marine-coated building stones, recent deposits, and eroded sherds, 1.4 to 1.9 m. below MSL
b. A fill of ashlar blocks from terrestrial buildings, some coated with stucco, mixed with sixth-seventh century Byzantine sherds, 1.8 to 2.3 m. below MSL
c. A layer of compact mud mixed with fine sand and Late Byzantine sherds that originated in brackish water or on land, 2.2 to 2.7 m.below MSL
d. A thin layer of decomposed kurkar chunks with some limestone cobbles mixed with friable chunks of lime and a few Early Roman sherds, in a layer integrated with the crumbled bedrock of poorly consolidated, laminated kurkar

The sea wall's first course of ashlars was bedded in the second of these layers, at 2.2 m. below MSL.  Over 6.5 m. wide, the sea wall was founded on reused column shafts laid parallel along its western side. Mostly displaced, these columns came to light on the western side of the wall, and beneath them several wooden planks were exposed.  On the western side the mud layer is much thicker, reaching a depth of 3.6 m. below MSL.

TN2 is another probe on the western side of the sea wall about 20 m. north of TN1.   In this trench, excavated only during the last days of the 1996 season, a layer of wooden planks came to light covering the entire 3 x 6 m. excavated area.  The planks appear to continue eastward under the base of the sea wall.  The top most planks lie 2.3 m. below MSL and are covered by some kind of cloth.  There are at least two layers of planks beneath them, some of which terminate on the west in a scarfed joint.   The excavators observed no tree nails, frames, or other fastening features, so if the planks are remnants of a ship's hull it is probable that the "skeleton first" technique of ship construction was employed, which became standard only during and after the Byzantine era.  The planks were covered by loose sand, building stones, and eroded sherds, and quantities of sixth-seventh century pottery, including intact vessels, appeared in the same stratigraphic context.  Small finds included a few nicely preserved glass vessels and a "lump" of more than seventy Byzantine bronze coins of the forty nummi type, minted about 600 C.E. and bearing countermarks of the Emperor Heraclius (610-41 C.E.). Another find, of uncertain consequence, was a gold solidus of Constans II dated 651-54 C.E.  This coin, minted a decade or more after Caesarea fell to the Muslims, might confirm the hypothesis that extensive debris in Area T represents deliberate dumping to block the anchorage and impede a Byzantine sea-borne raid.  Whether this dumping tookplace before or after the actual raid of 685 C.E. is an open question.  Discovery of later Arab and Crusader sherds in higher levels, however, might indicate later dredging and further harbor use, sometime in the ninth century, when dredged fill was dumped along the shore south of the medieval fortifications.

Area SW was a single probe positioned randomly 50 m. west of TN2, excavated to study the stratigraphic sequence in this location.  At 2.3 m. below MSL the sandy sea floor was littered with building stones and rubble in a loose, wave-circulated mixture.   Below 0.3 m.of this deposit, at 2.6 to 3.3 m. below MSL, was a distinct mud layer with many broken Byzantine jars and much other pottery, including earlier vessels, mostly imported amphoras and local jars from the 4th and 5th centuries C.E. Beneath the mud layer was a sand layer more than one m. in thickness, which flowed back into the trench and made excavation difficult. Nevertheless, near the end of the season the divers reached a mass of black basalt blocks not of local origin, probably the remains of ballast either jettisoned from a vessel or contained within the hull of a sunken ship.  Along with TN2 and other possible ship remains, this discovery will be the objective of further underwater excavation in 1997.

On land the project reopened Area LL (figs. 3, 4, 5, 6),to the north of Area T and the Inner Harbor.   Here a team from the Hebrew University had excavated in the mid 1970's, exposing a domestic quarter of the Islamic period (Hebrew University phases 1-3) and discovering a large, corridor type warehouse (horreum) from the Byzantine period (Hebrew University phase 4).(2)  The Combined Caesarea Expeditions have renewed this excavation as a major emphasis of future research.   The objective is to explore the rich Islamic domestic occupation that the Hebrew University team discovered, as well as the Byzantine warehouse and other occupation from the 4th through 7th centuries, but eventually to expose a large tract of what is expected to be a Hellenistic and Early Roman domestic quarter.

In 1996 the team opened three trenches, LL1, LL2 and LL3, adjacent on the south to the Hebrew University excavation. During 1996 the excavators devoted a high proportion of their time to penetrating and removing modern (19th-20th century, Bosnian and post-Bosnian) debris deposits and fragments of Bosnian domestic structures.  Ancient levels explored ranged from 4th-7th century Byzantine through much of the pre-Crusader Islamic, 7th through 11th centuries.

A high concentration of recirculated Late Hellenistic and Early Roman pottery (2nd century B.C.E.-1st century C.E.) in the earliest Byzantine fills, including many fragments of cooking vessels and tableware, suggests dense occupation in the earlier periods, perhaps domestic in character, but the earliest structure uncovered in 1996 was fragmentary plaster surface 2052 in trench LL2, at an elevation of 4.27 m.,(3) dated to the 4th century.  In the 7th century, at the end of Caesarea's Byzantine period, surface 2048, a compact floor of crushed kurkar and gray mortar at 4.39 m., abutted on the north the first phase of east-west wall 2028.  To the west, the excavators exposed north-south wall 1002, preserved more than 2.5 m. high, a continuation of the eastern wall of the corridor horreum that the Hebrew University team uncovered earlier, and parallel to it 5.75 m. to the west wall 1093, the eastern wall of the building's central corridor.

All three Area LL trenches yielded the same Islamic domestic occupation found in the adjacent Hebrew University trenches (phases1-3), but the excavators did not progress far enough as yet to integrate the new evidence with the old.  In LL1 both north-south horreum walls remained in use, but in association with a new floor, 1031/1039 at 4.84 m., constructed of kurkar slabs.  Associated with this stone floor was an apparent settling basin (1083) and several drains, dating from the 9th or 10th century and perhaps renewed in the Crusader period. In LL2, meanwhile, wall 2028 was rebuilt and functioned between the 8th and 10th centuries as the northern wall of a courtyard and a room, or of two rooms, represented by superimposed pavements 2026 (5.05 m.) and 2030 (5.22 m.) in the central part of the trench and by pavement 2041 on the west (fig. 5, 5.29 m.). North-south wall 2009 and associated pavement 2010, in the northeastern part of LL2 at a higher level (6.28 m.), dated to the Early Crusader period (12th century).  Here the team also excavated a stone-lined cellar or cess-pit belonging to a modern (19th to early 20th century) Bosnian house.  Except for walls 1002 and 1069, all of these structures, along with similar fragments of an Islamic house in LL3, lack the scale and formality that define monumental commercial or public architecture, and they belonged, therefore, to relatively small-scale houses or shops that characterized this quarter of the city at least from the 8th century to the 13th.

A special discovery was a group of eleven gold belt ornaments found in Islamic fill levels in LL1 but as the excavator, Professor Martha Risser, has already established, originating in the 6th or 7th century.  The fill levels in which the ornaments appeared surrounded catchment basin 1083, dating to the 9th or 10th century, but had presumably been recirculated from an earlier context when the basin was built.  The ornaments likely formed a coherent assemblage, deposited for safekeeping in a ceramic jar or other container that is now lost.  Dated by a cruciform monogram of the name Stephanos on a buckle tongue, the ornaments were apparently part of the elaborate costume of a governor or other member of the imperial aristocracy.

On the Temple Platform, Area TP, the project has been excavating during its summer seasons since 1989 (figs. 7, 8, 9). On this site, where Professor Avraham Negev excavated in the early 1960's, CCE has discovered the remains of an Early Christian church, octagonal in plan, built above the robbed foundations of King Herod's celebrated temple to Roma and Augustus (figs. 7-8).  During the first seasons, in 1989 and 1990, the team recovered ceramics and coins that indicated a date for the church ca. 525-550 C.E., an estimate that some in the team would now like to lower to ca. 500 C.E.   The temple, of which massive foundations came to light in 1995, would date to the original period of urban construction at Caesarea under King Herod's aegis, presumably 22-10/9 B.C.E., and indeed the excavators have uncovered construction fills, suitably dated by ceramics and coins to the end of the 1st century B.C.E.(4) (cf. below), that represent leveling beneath the pavements of the temple itself and of the surrounding esplanade.  On the other hand, Area TP has yielded virtually no Roman pottery or coins and no fills or structure that can be dated to the 1st through 5th centuries C.E.  This was to be expected, since during that long period the entire site would likely have been sealed beneath the temple's pavements and superstructure.  Similarly, after the construction episode ca. 500-550 C.E., the church and its surrounding pavements appear to have survived until the 10th century, and in the interim there was little or no building across the entire site.   Then, about 950 C.E., the building succumbed to a mixed domestic and commercial occupation represented by a few wall fragments, subterranean grain bins, and numerous cisterns.  There followed, certainly by the 12th and 13th centuries, but perhaps earlier, a renewed monumental phase (fig. 9) associated, perhaps, with the triple-apsed Crusader church located onthe Temple Platform's southwestern flank, or, less likely, with Caesarea's Great Mosque known from literary sources.(5)

Hoping to learn more about both the church and the temple, the team opened two trenches, TP19 and 21, in the northern part of the area.  In TP19 the excavation exposed a medieval well (19019), highly fragmentary remains of a pavement and wall foundations dated 10th century or later, and a thick soil layer (19055) densely packed with 4th-6th century C.E. ceramics that presumably represented a fill purposely imported when the church was built, in order to elevate and level the terrain beneath its pavements.  Below this fill, mortared onto the bedrock, was 19014, a north-south leveling course of kurkar blocks, elevation 10.42 m.  Partly concealed beneath the inner octagonal foundation of the later church (19003), this leveling course could represent the position of the temple's western cella wall.

In TP21 the team reexcavated a deep pit in the south center of the trench dug by Negev's project in the early 1960s but left undocumented and subsequently backfilled.   In this pit the excavators penetrated to bedrock at 8.78 m., only to reveal that here the temple foundations had been robbed out entirely, presumably by the builders of the church, leaving only evidence of quarrying and leveling the bedrock to receive the temple foundation.  Deposits of undisturbed Byzantine fill at the bottom and circumference of this pit (21041, 21053) suggested that the church builders had removed the temple foundation blocks before themselves backfilling in preparation for laying the church pavements.

Along the western side of TP19 and TP21 (figs. 10, 11) the excavators studied a massive north-south foundation wall, TP19039/21044, 1.9 to 2.5 m. wide, constructed of squared kurkar blocks on its faces laid in consistent courses, but in its core of less regularly coursed blocks set in find sand.  This wall lies parallel to a similar foundation wall 6.3 m. to the east,on the eastern side of TP19 and 21, which flanks a north-south paved street and extends alongside the street to both north and south.   In the absence of clear evidence, we interpret these foundations, which exist elsewhere in Area TP as well, as the substructure of walls massive enough to support vaulting over the streets and adjacent buildings.  The date of the foundation walls is still unclear, but they may have been founded as early as the 10th or 11th century and clearly remained in service through much of the Crusader period.  It appears therefore that for about 250 years after ca. 1000 C.E. a vaulted building, of monumental proportions and oriented north-south, occupied all of TP19 and TP21.  This could have been part of a hospital, monastery, or bishop's palace, for example, associated with the Crusader church, or possibly, at an earlier date, even Caesarea's Great Mosque.   Toward the end of the Crusader period, however, foundation 21044 was partly taken down in TP21, and some of its stones were incorporated into Crusader pavement 21021/21038 that covered the entire northern half of the trench at 12.51 m.  This pavement terminated on the south in wall 21032/21035, and to the south of that wall was another associated pavement, 21006, at 12.44 m.  These remains represent the end of the vaulted building in the later 12th or 13th century, although its successor, of different construction, might have served the same function.

Finally, the project investigated the southern approaches to the temple and the church during the 1996 season, excavating several probes in TP20 and TP22 (fig. 9) between the Herodian through Byzantine period staircase that surmounted the southern Temple Platform retaining wall and the temple and church to the north.  In its southern wall the octagonal church preserves a jamb block and fragmentary threshold of a monumental doorway in line with the staircase that suggested a paved street or walkway between the two structures, yet no clear remnant of such a pavement came to light.  The excavators did penetrate two successive strata of imported fill--the lower one (e.g. 22055 and 22059) at top elevation of 10.90 m. and the upper (e.g. 22050, 22046, 22035, 22025) at 11.80 m.-- that could represent leveling for paved esplanades that surrounded, successively, the temple and the church.(6)

The pavement that the excavators did uncover this year lay at an elevation of 12.83 m.   Designated as loci 20003, 20004, 22001, and 22038, it was 4.0 m. wide and made of kurkar slabs 20 x 20 cm. to 30 x 70 cm. on their exposed upper sides, generally laid across the direction of traffic.  This pavement dated to the Islamic or Crusader period, and was the later successor of the ancient street, continuing northward across the former site of the temple and the church, the latter dismantled, apparently, in the tenth century.   In TP22, as further to the north, the street was flanked by heavy and deep foundations, including 22027, the southward continuation of the unexcavated wall on the eastern side of TP19 and TP21 that is parallel and 6.3 m. to the east of wall 19039/21044.   In TP22, wall 22027 is built in part of spoils, like its analogue to the north, including two ancient Corinthian column capitals. Again, these foundations appear to represent walls massive enough to support vaulting, in this case a vault that covered the adjacent street.

In Area TP, the Temple Platform, the Combined Caesarea Expeditions plan two or three further field seasons, perhaps 1997 through 1999, followed by two years of analysis and publication of Herod's Temple, the octagonal church, and a rich medieval domestic and monumental stratum.  Thereafter, the plan is to concentrate exclusively on Area LL, with the aim of studying the development of the urban plan and domestic and commercial occupation through the various periods of the city's history.  Nor in future seasons will the harbor be neglected, linked as it is to the land site not only in terms of urban space but as a facility that through the centuries accounted in large measure for Caesarea's prosperity.

1. In addition to the directors, staff members included Frederick Winter, executive director; Anna Iamim, chief surveyor; Aaron Levin, photographer; Yael Arnon, registrar and ceramicist; Peter Lampinen, numismatics; Carol Cope, palaeozoology; Lisa Kahn, glass; Howard Williams and Verona Capone, architectural drafting; and Stefanie McCullough, assistant registrar. On the dive staff were Eduard Reinhardt, senior dive coordinator; Heather Reinhardt, dive rotations; and supervisors Derek Threinen, Nina Hodge, Chris Westfahl, Robert Crawford, and Jana Owen. Field staff on land were trench supervisors Audrey Shaffer, Jennifer Stabler, Farland Stanley, Virginia Rotter, Martha Risser, and, part-time, Robert Rausch, Jerry Black, Debbie Finch, and Sara Hecht; and assistant trench supervisors Sheila Van Deventer, Tina Horwitz, Kurt Reinhardt, and Dave Gmiter.

2. Lee I. Levine and Ehud Netzer, Excavations at Caesarea Maritima 1975, 1976, 1979--Final Report,Qedem, Monographs of the Institute of Archaeology, the Hebrew University of Jerusalem, no. 21 (Jerusalem: The Hebrew University,1986).

3. Henceforth all elevations are above MSL.

4. A. Berlin, "Hellenistic and Roman Pottery, Preliminary Report, 1990," Caesarea Papers: Straton's Tower, Herod's Harbour, and Roman and Byzantine Caesarea, ed. by R. Lindley Vann, Journal of Roman Archaeology, supplementary series, no. 5 (Ann Arbor: Journal of Roman Archaeology, 1992), 112-24.

5. For excavations published thus far see K. Holum et al, "Preliminary Report on the 1989-1990 Seasons," in Caesarea Papers, ed. Vann, 104-9; K. Holum, "The Temple Platform (Area TP)," in the Combined Caesarea Expeditions: Field Report of the 1992 Season, ed. by A. Raban, K. Holum, and J. Blakely, The Recanati Center for Maritime Studies, publication no. 4 (Haifa: Center forMaritime Studies, 1993), 57-60; and K. Holum's study, "The Temple Platform: A Progress Report," in Caesarea Papers II (forthcoming).

6. The preserved top of the Herodian retaining wall in Z2 just to the south at 11.87 m., and it is clear that the upper portion of this wall is a later (Byzantine) addition.  Thus the fill levels might correspond reasonably well with successive esplanade surfaces at ca. 11 and 12 m.