Where Is Ezion-Geber? A Reappraisal of the Site Archaeologist Nelson Glueck Identified as King Solomon’s Red Sea Port

024

King Solomon built a fleet of ships at Ezion-Geber which is near Elath on the shore of the Red Sea, in the land of Edom … and they went to Ophir and brought from there gold … and they brought it to King Solomon” (1 Kings 9:26; see 2 Chronicles 8:11–18 for parallel passage).

Ophir is endlessly elusive. But where is Ezion-Geber?

The geographical markers concerning Ezion-Geber in the Bible are so numerous and so locatable that one might suppose it would be easily found. It was the southernmost port city of the Solomonic empire; from there royal ships sailed to Ophir. It was in the land of Edom, on the shore of the Red Sea, near Elath.

According to Numbers 33:35–36, the Israelites camped at Ezion-Geber during their desert wanderings, after they had received the law at Mt. Sinai and before they entered the land of promise. While we have not identified most of the stops named on the route of their desert wanderings, we can locate the area of Kadesh, the next stop after Ezion-Geber. We are also told that the stop after Kadesh was “on the edge of the land of Edom,” a description consistent with the geographical markers in the passage from 1 Kings (and 2 Chronicles). And in Deuteronomy 2:8, we are told that Ezion-Geber is at the end of the Aravah road, another more or less identifiable geographical marker. We surely know where to look for Ezion-Geber.

In 1933 the German explorer Fritz Frank surveyed the Aravah, the depression that extends from the southern end of the Dead Sea to that finger of the Red Sea the Jordanians call the Gulf of Aqaba and the Israelis call the Gulf of Eilat. The Aravah and the Gulf are parts of the Great Rift, which continues south into the Red Sea and, beyond that, into Africa. About 1,500 feet from the northern shore of the Gulf of Aqaba, Frank came upon a low mound or tell, barely distinguishable from the surrounding hillocks except for the unmistakable signs of ancient human occupation. Its Arabic name was Tell el-Kheleifeh. Biblical Edom was located in the southern part of Transjordan, today southern Jordan. Tell el-Kheleifeh is on the Jordanian side of the Aravah, in ancient Edom, almost equidistant between the modern cities of Eilat in Israel and Aqaba in Jordan. Barely 20 feet west of the site a fence marks the beginning of the neutral zone between Israel and Jordan.

Not surprising for the time, Frank identified Tell el-Kheleifeh as Biblical Ezion-Geber.

In 1937, the American rabbi and archaeologist Nelson Glueck, then with the American School of Oriental Research in Jerusalem, led a surface survey in Transjordan, including an examination of the mudbrick remains exposed on the surface of Tell el-Kheleifeh. Based on the potsherds he picked up on the surface, Glueck concluded that the site had been occupied between the tenth and eighth centuries B.C. This fit nicely with the Biblical references in Kings and Chronicles; Solomon ruled between about 970 B.C. and 925 B.C. There are Biblical references to Ezion-Geber 025in the ninth century as well: The Judean king Jehoshaphat (873–849 B.C.) made an unsuccessful attempt to continue the Arabian trade network earlier established by King Solomon. Jehoshaphat’s effort literally shipwrecked. According to the Biblical account, “Jehoshaphat built a fleet of trading ships to go to Ophir for gold, but they never set sail—they were wrecked at Ezion-Geber” (1 Kings 22:48). Later in the ninth century B.C., Judean control of Ezion-Geber was interrupted when Edom rebelled against Judah during the reign of King Jehoram (849–843 B.C.) and set up its own king. Jehoram unsuccessfully attempted to reassert Judean authority (2 Kings 8:20–22; 2 Chronicles 21:8–10). This is the last Biblical reference to Ezion-Geber.¹

Between 1938 and 1940, Glueck directed three seasons of excavations at Tell el-Kheleifeh.² He identified six major periods of occupation and dated them between the eleventh and fifth centuries B.C. This was a significant revision of his earlier chronology, but from the beginning, Glueck accepted Frank’s identification of the site as Ezion-Geber. Though some uncertainty is reflected in Glueck’s field records and in later publications, this identification as Ezion-Geber remained the underlying premise of Glueck’s interpretation of the site’s stratigraphy and architecture.

I have recently completed a reappraisal of Glueck’s excavation of Tell el-Kheleifeh.³ Fortunately, Glueck’s field records are commendably detailed—in contrast to records for so many digs of that period. Artifacts from the excavation are available for study at three principal repositories: the Semitic Museum of Harvard University, the Smithsonian Institution in Washington, D.C., and the Jordan Archaeological Museum in Amman.⁴

In addition, on three visits to the site, I studied the few building remains that have survived the past few decades,^a and I collected some artifacts that had been strewn over the site as a result of recent digging of military trenches and construction of a military observation tower on the mound’s summit. These disturbances in fact produced a wealth of new finds, including a stamped Rhodian jar handle from 200 B.C. and a bronze trefoil arrowhead. These finds, including additional pottery sherds, provide a valuable complement to Glueck’s 1938–1940 assemblage.

My reappraisal has confirmed some of Glueck’s conclusions, but has also suggested the need for some major refinements and revisions. This is hardly surprising. Revision and refinement are inevitable in any scientific 026endeavor. In archaeology, they are an indication of how much we have learned in 50 years of excavation and study. This process should not be confused with criticism.

Nelson Glueck was a pioneer. He led the way into the wilderness and prepared it for those of us who followed. His surveys and excavations remain the paradigm. His studies and articles remain the focal point for the study of the archaeology and historical geography of the Negev and Transjordan.

But refinements in excavation methodology and pottery typology since Glueck’s day have been enormous. Even in his lifetime, Glueck, as we shall see, welcomed revision and correction when based on improved understanding and methods. I have no doubt that, had he lived to complete his work on Tell el-Kheleifeh in the 1980s, he would have come to the same revised conclusions I present here. Unfortunately, he died in 1971 without ever writing a final report on his excavation of Tell el-Kheleifeh.

During the first excavation season at Tell el-Kheleifeh, Glueck exposed approximately 45 rooms on the northern perimeter of the site. The most impressive structure consisted of three long adjacent rooms and three small rooms at the northern end of the long rooms. Today we would easily recognize this as a so-called Four-Room House, a plan typical of Israelite architecture in the Iron Age (1200–586 B.C.). This particular example of a Four-Room House is quite large, about 40 feet on a side. Each of the three long rooms is about 8 feet wide and 24 feet long. Most of the walls are more than three feet thick. The three small rooms (which together comprise the subdivided fourth room in the paradigm of the Four-Room House) are, in the aggregate, about the same size as any one of the three long rooms and, as noted, extend along the northern end of the three long rooms. The building is constructed almost entirely of mudbrick.

Glueck noted that certain walls of this building were perforated with two horizontal rows of holes or apertures. According to Glueck, the upper row of holes led into an air channel that extended along the length of the wall. Glueck interpreted the building itself as a copper-refining plant and the two rows of apertures as flues. The lower row of apertures, Glueck reasoned, permitted the transfer of heat between chambers of the building, while the upper course provided the necessary draft. He described the smelting process in this way:

“Pottery crucibles were placed on loosely packed masses of clay debris which had been burned to the consistency of intensely fired bricks by the heat which permeated through them from the charcoal fires built on them. The already ‘roasted’ ore placed in the pottery crucibles over the hard-baked clay debris was fired by packing charcoal around and over the crucibles in the furnace rooms, which were not roofed over. The strong drafts of air coming through the flue holes would fan the flames to a furious heat. The charcoal was furnished in part by burning palm wood from the vicinity, but most of it was undoubtedly imported from the wooded hills of Edom proper.”⁵

During the 1939 season Glueck determined who was responsible for the construction of the smelter.

“There was, so far as we know, only one man who possessed the strength, wealth, and wisdom capable of initiating and carrying out the construction of a highly complex and specialized site, such as the factory town of Ezion-Geber in its first and greatest period. This was King Solomon. He alone in his day had the ability, the vision, and the power to build an important industrial center and seaport so comparatively far from Jerusalem. With the building of a new Ezion-Geber, Solomon was 027able to have smelted and refined and worked up into finished products the ores extracted from his great copper and iron mines in the ‘Arabah, and was then able to export them directly by sea and by land in exchange for the spices and ivory and gold and precious woods of Arabia and Africa. The wise ruler of Israel was a copper king, a shipping magnate, a merchant prince, and a great builder.”⁶

After 1939, the site became known as the “Pittsburgh of Palestine.”

Glueck himself later revised this romantic interpretation.⁷ There was no large-scale smelting at Tell el-Kheleifeh, he conceded, although he continued to maintain that some industrial and metallurgical activity occurred there, a suggestion confirmed in our reappraisal.

But if the building wasn’t a smelter, what were the two rows of holes that Glueck interpreted as air circulation holes and flues? They may reflect a method of construction now widely attested throughout the ancient Near East but practically unknown when Glueck excavated the site. To strengthen the mudbrick construction, wooden beams were laid within the mudbrick wall, both parallel and perpendicular to its length. The interior and exterior wall surfaces were then plastered with a mud coating that concealed the ends of the perpendicular beams; the beams laid parallel to the wall’s length were completely concealed within it. When this building was destroyed in a fire, the wooden beams burned, leaving holes penetrating the thickness of the wall to mark the former location of the perpendicular beams.

Although the most probable explanation of the holes is the one just described, two other explanations should be mentioned. The upper row of holes may have supported ceiling rafters that spanned the room; but this does not easily explain why the beams were so low—about 5 ½ feet—from the bottom of the wall. A few of the holes in the lower row have a clear half-circle shape, rather than the irregular shape of most of the holes. The irregular holes probably held the wooden beams that were part of the interior wall lattice. But the occasional half-circle shaped holes may have served a ventilation function, providing a flow of fresh air into the mudbrick building, especially to those rooms that contained hearths or ovens. At present, no conclusive explanation of the holes is possible, but they are best explained by the theory of interior wooden wall supports. One thing is clear: the building was not a smelter.

The structure, as we have noted, was built on a typical Four-Room House plan, which was used for large public buildings as well as for private dwellings. We find structures built on this plan in cities, in villages and in wilderness fortresses. This particular Four-Room House is similar in size, design and quality of construction to other monumental Four-Room Houses uncovered since 1940 in such cities and towns as Hazor, Tell Beit Mirsim, Tell el-Fariah (north) and Tell en-Nasbeh. These sites range in location from northern Israel to the Negev. The Tell el-Kheleifeh example is atypical, however, in that none of the dividing walls between the long rooms consists of a row of pillars. Usually in Four-Room Houses, at least one of these dividing walls is a row of pillars, which usually faced a roofless courtyard within the structure.

What was this monumental Four-Room House used for? Clearly it was not a domestic structure. It was a major public building. When Glueck abandoned his smelter theory, he suggested that the building was used for storage. Unfortunately, the few artifacts found in the building—a clay stopper and a few pottery sherds—provide little help in determining its function.

031

From an architectural perspective, it does not seem that storage was a primary function of this large public building. The Israeli archaeologist Yigal Shiloh has argued that this building and other large buildings like it probably functioned as citadels.⁸ He is probably right, if we conceive of a citadel as a stronghold or fortified place for defense or refuge. But this identification should not be taken to imply any specific military, political, commercial or domestic use. In short, we simply do not know the specific use to which this building was put.

At the time of his excavations, Glueck first thought that the four-room building stood alone. He designated this first level as Period IA. We now know, however, that the building was an integral part of the site’s earliest occupational level, a fortress enclosed by a casemate wall.^b The building never stood as an isolated structure but, instead, was located in the middle of a square, about 135 feet on each side, formed by the casemate wall.

The space between the casemate perimeter wall and the central structure was apparently devoid of other buildings and served simply as a courtyard. As is clear from the aerial photographs, the casemate fortress was only excavated on the western, southern and eastern sides. The northern section of the enclosure remains buried beneath Glueck’s northern dump.

Later, a solid wall with offsets and insets was built along an entirely different line from the casemate wall. In this new solid wall enclosure, the Four-Room House stood in the northwest corner. Part of the southern and eastern sides of the old casemate wall remained inside the solid enclosure wall, but the rest of the old casemate wall was outside, and probably abandoned. This later fortress, which we have called the solid-wall fortress, represents the last major architectural level at the site.

When Glueck first excavated the rooms within the parallel casemate walls, not realizing that they were part of a casemate wall, he described them as a row of industrial workshops associated with the nearby “smelting” complex (the Four-Room House). In several of the casemate rooms he found structures he interpreted as hearths, thus giving credence to the smelting hypothesis. After he abandoned the smelting hypothesis, Glueck recognized the row of rooms as a casemate fortification wall. Despite his revision with regard to the function of the Four-Room House and the casemate complex as a whole, however, Glueck continued to date this earliest level to the time of King Solomon, the tenth century B.C.

Glueck’s primary dating tool for the supposed Solomonic level was a pottery assemblage he described as “crude, handmade, friable, smoke-blackened pots, many of which were built up on a mat, and most of which have various simple types of horn or ledge handles, or combinations of both.” Originally, he interpreted these vessels as crucibles and associated them with the smelting of copper ores, but later he recognized them as ordinary domestic pottery, handmade rather than wheelmade.

When Glueck excavated this pottery, archaeologists were just beginning to become familiar with it. Today, we know it as “Negevite” pottery. We also know that it cannot be used as a chronological indicator. The same handmade pottery was used (primarily in the Negev) for hundreds of years, embracing the entire Iron Age (1200–586 B.C.) at the very least. Nor can it be regarded as any indication of ethnicity. Instead of being used to date an excavation level, Negevite pottery must itself be dated by associated wheelmade pottery forms often found with it.

Because Glueck relied primarily on this “Negevite” pottery, which was considered at the time diagnostic of the tenth century B.C., to date the Four-Room House, he saved only these handmade wares to document the 032chronology of the building. The wheelmade pottery he threw out. Therefore we are now deprived of the only reliable evidence for dating the structure.

Studies in comparative architecture sometimes help to date archaeological levels, especially when solid dating evidence is lacking, as in the case of Tell el-Kheleifeh’s casemate fortress. Unfortunately, Four-Room Houses and casemate walls are well documented throughout the Iron Age, and the Tell el-Kheleifeh fortress reminds us of an architectural tradition that is well attested from the central Negev southward to the site of Tell el-Qudeirat (the site that may be Kadesh-barnea). Israeli archaeologists, including the late Yohanan Aharoni, Rudolph Cohen and Ze’ev Meshel, have surveyed and, in some instances, excavated an extensive network of these sites in the Negev highlands. They feature casemate fortifications that enclose open courtyards; there are often settlements with dwellings constructed on the Four-Room House plan nearby. These sites differ from one another in shape and size, and the casemate rooms differ in number and size; the plan, position and construction of the gateways also differ. As at Tell el-Kheleifeh, the entrance to the compound usually consists of an opening in the line of casemate units.

Cohen has suggested a four-fold classification system based on groundplan for these Negev fortresses: (1) roughly oval, (2) rectangular, (3) square and (4) fortresses with towers.⁹ However, none of these features can be used to date these sites. Indeed, based on the evidence from nearly 40 of these sites, we can say that the variations in ground plan—oval, round, square, etc.—seem to be not only chronologically irrelevant, but typologically and functionally irrelevant as well. It may be that topography was the primary consideration in determining the plan of these sites.

In “The Iron Age Sites in the Negev Highlands: Military Fortresses or Nomads Settling Down?” BAR 12:04, Israel Finkelstein even challenged the generally accepted function of these sites in the Negev highlands as fortresses. Many of Finkelstein’s observations are well taken. Nevertheless, Finkelstein’s alternative interpretation does not seem adequate to explain all of these complexes, especially those of substantial construction and strategic location. Certain of the sites may, as Finkelstein suggests, reflect the process of sedentarization of nomadic groups, but this is not the only possible interpretation for every structure. Similarity of design does not 033demand uniformity in function. In my view, the “fortress” hypothesis best accommodates the total picture and remains the most plausible interpretation, although we must allow for the possibility of exceptions.

Tell el-Kheleifeh’s earliest phase with casemate walls does appear to be related architecturally to the central Negev fortress tradition. Although this casemate fortress is significantly larger than the square central Negev fortresses, its dimensions are comparable to those of other designs.¹⁰ The size and quality of our fortress were probably determined by the site’s strategic location and the use of mudbrick, rather than stone, construction.

One of the major differences between the Tell el-Kheleifeh fortress and the central Negev fortresses is the location of the Four-Room House. In the Negev fortresses, domestic Four-Room Houses were found outside the casemate wall; at Tell el-Kheleifeh a large public Four-Room House was located in the center of the casemate square.

Despite the similarities between the Tell el-Kheleifeh fortress and the central Negev fortresses, we have a problem with dating. Cohen believes that many of the Negev fortresses should be dated to the time of Solomon (tenth century B.C.) and that they relate to Solomon’s efforts to defend his southern border, although Cohen recently re-dated three of these square fortresses to the Persian period (“Solomon’s Negev Defense Line Contained Three Fewer Fortresses,” BAR 12:04). Meshel, on the other hand, assigns the fortress network to “one of the kings who defeated the Edomites and Amalekites,” most likely Saul or David. At Tell el-Kheleifeh, however, none of the wheelmade pottery from the site, whether from the casemate fortress level, or the later solid wall fortress level, dates earlier than the eight century B.C., though a few forms could be dated to the ninth century. No 11th- or 10th-century pottery is attested at Tell el-Kheleifeh. Therefore, the parallel between the Tell el-Kheleifeh casemate phase and the plan of the central Negev fortresses remains a relationship of comparative architecture, and not chronology.

Accurate dating of our casemate fortress, including the four-room building in the center, remains a problem. A brief season of excavation at the site could probably solve the problem. By removing Glueck’s northern dump, the unexcavated northern third of the casemate square could be investigated and its pottery horizon isolated. This section of the casemate wall represents one of the site’s few remaining areas of undisturbed stratigraphy.

We noted above that after the destruction of the casemate fortress, the plan of Tell el-Kheleifeh was radically changed. The fortress was replaced by a significantly larger settlement with a solid offset/inset wall and a four-chambered gateway in the southwestern corner. The new fortification is approximately 180 feet on a side, compared to approximately 135 feet in the casemate wall. The new fortified settlement with a solid wall, retained inside it a portion of the earlier casemate fortress, creating something like an inner enclosure or courtyard in the northwest quadrant of the new solid-wall plan. The northern and western sides of the earlier casemate wall were now outside the new solid wall, which also destroyed a portion of the northern wall of the Four-Room House.

The new solid-wall fortification also included a poorly preserved outer wall, described by Glueck’s architect Jacob Pinkerfeld as “a thin low outside wall, whose purpose was to delay the assailant a little before he could reach the main walls.” A 150-foot length of this outer wall is preserved in the southeast corner of the site. Another section is preserved on the western side of the settlement, about 03412 feet outside the solid wall.

Both Glueck and Pinkerfeld maintained that the earliest offset/inset plan was devoid of any buildings within the two courtyards created by the solid and casemate walls. According to Glueck, the later phases of this larger settlement witnessed continued building within these two enclosures, eventually producing the concentration of structures that is visible in the 1940 aerial photograph. Our reappraisal suggests, however, that, from its beginning, the earliest phase of the new plan was a fortified settlement with interior architecture. Undoubtedly, the construction of buildings within the offset/inset wall was a process that continued for a long time.

The date of the fortified settlement with a solid wall can be established with reasonable certainty from the wheelmade pottery Glueck found and saved and also from some very important seal impressions. The wheelmade pottery of Tell el-Kheleifeh’s solid-wall settlement can be isolated from Glueck’s records so that we can establish the earliest and latest dates for this level. Clearly, the wheelmade forms date between the eighth and early sixth century B.C. The vessel types are all diagnostic of the late Iron Age: the so-called “Edomite” cooking pots; inverted-rim kraters; “Assyrian” bowls, carinated cups, bottles and censers; juglets, decanters and saucers, to name only a few. On the basis of this pottery, we can say with confidence that Tell el-Kheleifeh’s fortified settlement dominated the region of the southern Aravah between the eighth and early sixth century B.C. This chronology is also suggested by nearly two dozen stamp impressions that appear on the handles of certain wheelmade vessels. The small oval stamps read “belonging to Qaws‘anal, servant of the king” (lqws‘nl ‘bd hmlk). The script is datable to the late seventh or early sixth century B.C.

035

There is evidence of occupation at Tell el-Kheleifeh after the sixth century B.C., though it is very fragmentary. The building remains are scanty at best, consisting of a few mudbrick walls that were constructed on top of the fortified settlement but on a different alignment. A few Phoenician and Aramaic ostraca of the fifth and early fourth centuries B.C., together with a handful of fifth-century Greek body sherds, constitute the most reliable dating criteria for this level. Some sixth- to fifth-century bowls, jars and storage vessels were also uncovered.

Some important questions remain unanswered: Who was responsible for building the casemate fortress? What was the site’s specific function? If the architecture of the earliest level is related to the so-called fortress tradition of the Negev highlands, where are the material remains, especially the pottery, from the eleventh or tenth century B.C.?

This last question raises the problem of identification, which is perhaps the most perplexing of all. Tell el-Kheleifeh provides no clear archaeological evidence, either ceramic or architectural, for its identification with the Ezion-Geber of Israel’s wilderness traditions (Numbers 33:35–36) or even of King Solomon’s reign in the tenth century (1 Kings 9:26–28); indeed the site provides only questionable evidence for the ninth century (1 Kings 22:47–48). Do the ruins of Tell el-Kheleifeh preserve the story of Biblical Ezion-Geber?

In short, is Tell el-Kheleifeh Ezion-Geber?

Glueck’s interpretation of the site was unfortunately based on a Biblical framework for which there was no archaeological justification. The archaeological data were made to conform to the historical contours provided by the Bible. Glueck’s identification of the site as Ezion-Geber remained the underlying premise of his interpretation of Tell el-Kheleifeh’s stratigraphy and architecture, and ultimately for its occupational history.

Tell el-Kheleifeh must be allowed to tell its own story in its own language. For the moment at least, based on the available archaeological data, Tell el-Kheleifeh cannot be identified as Ezion-Geber. This does not detract from Glueck’s achievement, but simply recognizes how far we have come in the last 50 years.

King Solomon built a fleet of ships at Ezion-Geber which is near Elath on the shore of the Red Sea, in the land of Edom … and they went to Ophir and brought from there gold … and they brought it to King Solomon” (1 Kings 9:26; see 2 Chronicles 8:11–18 for parallel passage). Ophir is endlessly elusive. But where is Ezion-Geber? The geographical markers concerning Ezion-Geber in the Bible are so numerous and so locatable that one might suppose it would be easily found. It was the southernmost port city of the Solomonic empire; from there royal ships […]