The barren sand track leads south from the ancient silt fields of Marib through the sands of the Rub’ al-Khali, along the edges of the rugged mountains of central Yemen, over precipitous passes, and finally to Wadi Beihan, where Hajar. Bin Humeid and Timna’ are located (see Fig. 2). Timna’ was the capital of pre-Islamic Qataban, and Hajar Bin Humeid was a point for frankincense and myrrh transshipment from production areas even further to the south in Arabia. Fragrance-laden camel caravans moved north from the Wadi Beihan on their long and rewarding journey to the markets of Egypt, Palestine, and Mesopotamia (Groom 1981). They traveled well-worn paths along the edges of the Arabian desert, passing Marib, the capital of Sheba (Saba’) and the most powerful of the ancient South Arabian kingdoms, as they plied their route north. Apparently with the rise of sea trade for spices in the first centuries A.D., and especially with the advent of Islam in the 7th century A.D., the route came to be virtually abandoned. It was only with the introduction of powered vehicles in the 1960s that the route between Marib and the Wadi Beihan again regularly carried a variety of international commodities.
Halfway between Marib and Wadi Beihan lies the Wadi al-Jubah, a mountain-ringed, defensible piece of arable land which is about 22 ki lometers long and between 3 and 8 kilometers wide (Figs. 3, 4). Since 1982, the American Foundation for the Study of Man (AFSM) has sponsored three seasons of archaeological research within the confines of the wadi. At the present time two further seasons are planned in the current phase of operations.
Prior to the arrival of the AFSM team, no archaeological research had ever taken place in the wadi. In 1976, Father Albert Jamme of the Catholic University of America visited the wadi, collecting pre-Islamic Arabic inscriptions (see Jamme in Toplyn 1984:70-71). In addition, over the years a number of pre-Islamic Arabic artifacts, including some artistic pieces, made their way from Wadi al-Jubah to the National Museum in Sanaa. Thus, the initial problem facing the project in 1982 was how to tackle archaeological research in a largely unknown and unmapped area.
Fortunately, some important research had been done in both Wadi Beihan (Timna’, Hajar Bin Humeid) and Marib in the early 1950s. Wendell Phillips and William F Albright had led AFSM archaeological teams that had explored the pre-Islamic Arabic occupation, cemetery, and agricultural installations at these sites. The results of these projects have been published and are well known (Phillips 1955; Bowen and Albright 1958; Cleveland 1965; Jamme 1962; Van Beek 1969). Unfortunately, no clear cut chronology, either epigraphic or ceramic, has come to be accepted for use in dating these or other archaeological remains in Yemen. Scholars have generally been divided between those who argue for a fairly ‘early’ date (ca. 1100 B.C.) for the beginning of this pre-Islamic Arabic culture (e.g., Jamme 1962; Van Beek 1969), and those who argue for a much later (ca. 500 B.C.) starting date (e.g., Pirenne 1956).
The AFSM team members decided to conduct a reconnaissance in the Wadi al-Jubah in order to determine the archaeological potential of the preserved remains in a cost effective manner (Toplyn 1984). Eight weeks of survey in 1982 and 1983 yielded the discovery of 101 new sites. These sites were found through visual reconnaissance during a circuit of the wadi, and through information gathered from the local inhabitants. Three cultural horizons were found during the reconnaissance: ‘Neolithic,’ Pre-Islamic Arabic, and Modern.
Neolithic Period
Neolithic structures were found in various places around the edges of the wadi. At all times they were found at an elevation higher than the wadi floor, and usually in the lower channels of subsidiary wadis where they debauch into the main wadi. Four types of structures which tentatively have been identified as ‘Neolithic’ have been found (Fig. 5): round houselike structures, round cairns, rectangular cairns, and lines of stone found in the vicinity of the houselike structures. They are all constructed Of stone and are covered with `desert varnish’ (patina acquired through time on exposed surfaces). In no case has any artifact been found in conjunction with these structures. Their identification as ‘Neolithic’ is based on two lines of evidence. First, the French archaeological mission in Yemen found similar structures that they dated to the Neolithic period partially on the basis of identifiable lithics (Bayle des Hermens 1976; Bayle des Hermens and Grebenart 1980). Second, independent evidence for an early date was found at our site AFA 56G, where an ash deposit was found next to one of these structures. Beta Analytic Inc. analyzed the ash and determined a radiocarbon date of 5270 ± 90 b.p. (Beta-7823). The uncalibrated date determined from this radiocarbon date is 3320 ± 90 b.c., using a 5568-year carbon-14 half-life and one-sigma precision. (Alternatively, using the new CRD at a two sigma standard deviation, the ash should date between 4395 and 3855 B.C. [Klein et al., 1982:129].) This date seems to be in general agreement with the French dating of their material. We hesitate to call our structures Neolithic, however, because we have excavated no stratified deposits and have no material evidence to indicate that cultural period. Also, at the present time we are not sure that the term ‘Neolithic’ is appropriate for the date indicated by the radiocarbon analysis, since so little is known about this time period in South Arabia.
Pre-Islamic Arabic Period
By far the greatest number of sites in the wadi date to the pre-Islamic Arabic (Qatabanian and Sabaean) period. Four types of evidence for this cultural horizon have been found: epigraphic, cemetery, irrigation/agricultural, and occupational.
Epigraphic material was the goal of Father Jamme when he reached Wadi al-Jubah in 1976. Most of the pre-Islamic epigraphic evidence was found on blocks of building stone that had been incorporated into Modern buildings (Fig. 6). The problem of interpretation is compounded by the recent moving of building stones by vehicles. It is clear that some blocks found in the Wadi al-Jubah originated elsewhere since the inscriptions on the blocks were recorded at other sites prior to their removal during the past 30 years (Jamme in Toplyn 1984:70). Previously unknown inscribed blocks cannot be assumed to have originated in our wadi unless they are found in a stratified context or are in situ. One inscribed block found in the Wadi al-Jubah can certainly be called in situ. It was found in the walls of an ancient well where it could easily have been read while water was being drawn.
About 200 pre-Islamic graffiti, which usually contain only personal names, have been found lightly carved into various vertical mountain faces around the edges of the wadi. Also, about 20 inscribed potsherds have been collected; these typically contain fragments of letters and names.
Remains of two or three pre-Islamic Arabic cemeteries on the floor of the wadi are known to us. We have not examined the cemeteries in detail, but informants describe typical pre-Islamic Arabic pottery as coming from these burials. In one case, we found large quantities of burned human bones on a small rise, which was apparently a burial installation.
Many of the sites in the wadi are non-occupational; rather they relate to the ancient dam/irrigation/canal/silt agricultural system. The agricultural methods employed by the ancient Qatabanians and Sabaeans were one of the few things known as we started the work in Wadi al-Jubah, since the AFSM expedition to Wadi Beihan in 1950-51 explored that system (Bowen and Al-Night 1958), and compared it with the massive dam and silt deposits at Marib. A German team has recently explored the Marib remains and tried to reconstruct how the system actually worked (Schmidt 1982; Brunner 1983). The massive dam at Marib, with its channels and silt fields, served as a model to explain what we first saw on a smaller scale in the Wadi al-Jubah: massive amounts of weathered silt, some silt tracts, remnants of channels, and subsidiary dams. Problems exist, however. A fairly large dam might have been expected in order to parallel Marib, but no evidence for such a dam has yet been found. Many small localized dams exist, however, around the edges of the wadi. This situation is causing problems of interpretation that we hope to solve in our next two campaigns.
Channels which are covered with decomposed ‘concrete’ led to ancient field systems. The fields are now large tracts of deeply weathered silt. A probe through the silt beds suggests that they are about 10 meters deep and that all of the accumulation is the result of man’s agricultural deposition. Also, the bedding of the silt and the location of the visible channels atop the silt beds suggest that the system grew and changed over time, with the channels and human occupation sites rising as the silt deposition in the fields grew (Fig. 7).
Another interesting facet of the ancient agricultural system is the frequency of wells found in and around it. This may be explained as a form of water recycling. Rain comes to this part of Yemen in the form of mountain cloudbursts during monsoon seasons twice a year, in the summer and in the winter. The irrigation systems caught the water and directed it to the fields during flood conditions. The wells may be explained as a means of bringing the water back to the surface after it had percolated through 15 to 20 meters of silt and wadi gravel. This would supply water during the long dry seasons when no rain could be expected. The purpose of the well water is less clear. No mechanism for large-scale water raising has been found, but the use of well water for some kind of irrigation would still appear to have been possible (e.g., for growing palm trees). It could also have been used as a source of drinking water for both humans and animals.
Ancient mounds (hajran) containing the remains of human occupation are common, throughout the wadi (Fig. 8). Small dwellings, villages, and one major ancient city have been found. Again, most seem to be found around the edges of the wadi where the ancient agricultural fields are best preserved. The numerous small occupation sites usually are found on the silt flats. Probes have been excavated at two of these sites. At Hajar al-Kanus (HK 25), the site was found to be sitting on top of about 8 meters of silt (Fig. 9). Typical pre-Islamic Arabic pottery (Fig. 10) was found at this site in association with obsidian debitage. Farther south at Hajar at Tamrah (HT 12), a larger probe discovered over 4 meters of occupational remains resting above the underlying sand and silt. Here, four-occupational strata were isolated. The top layer contained the collapse of a roofed structure. Broken pottery was found along with the carbonized roof collapse. Radiocarbon dating of three of these beams yielded the following information, which, again, is uncalibrated and presented at a one-sigma confidence level:
Beta-7173 2460 ± 60 b.p.
510 ± 60 b.c.
Beta-7174 2380 ± 50 b.p.
430 ± 50 b.c.
Beta-7175 2630 ± 70 b.p.
680 ± 70 b.c.
This indicates an abandonment of the site around 400 b.c. at the very latest, and very similar radiocarbon dates came from the uppermost (terminal) layers of two other sites in the wadi, Jubah al-Jadidah and Hajar al-Husn (Blakely 1983). The bottom layer at Hajar at-Tamrah contained a few scraps of pottery and a chunk of charcoal. This piece (Beta-7178) yielded a date of 328C ± 110 b.p., (1330 ± 110 b.c.) uncalibrated. This dating evidence clearly supports the ‘early’ chronology that was noted above, and ii probably even pushes it back somewhat further in time to the 13th century B.C.
Hajar at-Tamrah also yielded identifiable samples of wood, seed, bone, and obsidian. The wood and seeds suggest that Zizyphus spina christi was very common in the area. This is a small bushlike tree that produces an edible fruit. It has been called lotus and jujube in English. The faunal collection consists mainly of sheep, goat, and camel, but other species such as cow and various equids are present. The obsidian collection represents a microlithic industry. At the present time the analysis of this material is incomplete, and the full results will be presented in the forthcoming report on the site.
A number of medium-sized sites exist in the wadi. Some of them are concentrated at the north end where the mountains almost join. The area is naturally defendable, and three fortlike sites at this point may indicate a defensive barrier. In comparison, two other natural and less defensible entrances to the wadi in the southeast have modest defensive walls or barriers, but no major occupational remains nearby.
The most important occupational site in the wadi is located at the modern village of Jubah al-Jadidah. The name of the site is Hajar ar-Rayhani (HR 3), but it is usually referred to by the village name (Jubah al-Jadidah). The site is between 5 and 10 acres in size, and it is surrounded by a large defensive wall. Potsherds and finely worked stones litter the mound, as does the Modern period garbage dump. Modern stone robbing has opened a number of cuts in the defensive wall, and in January and February 1984, we excavated a small probe in one of these cuts (Figs. 11, 12a—b). This probe went down about 6 meters and was then halted, when safety and limited work space necessitated that decision. Five occupational strata were isolated here. The top stratum appears to have been deposited after a major destruction of the defensive wall. Pottery from this top deposit seems to be of a later variety than that found at Hajar at-Tamrah, and it may date to the last third of the first millennium B.C. (note, however, that the radiocarbon evidence given above argues for similar terminal dates, ca. 400 to 300 b.c., at Hajar at-Tamrah and at Jubah al-Jadidah). The second through fourth strata are all layers that are associated with the active usage of the defensive wall. Pottery similar to that from Hajar at Tamrah suggests a date in the mid-first millennium B.C. for usage. The most exciting discovery was in the lowest (fifth) stratum that we encountered. Here the remains of a copper foundry were uncovered. Crucibles, slag, and charred wood were found along with what appeared to be foundry floor layers. Currently these remains are under study in MASCA (Museum Applied Science Center for Archaeology). In future work we hope to explore a larger area of the foundry and to reach its bottom. Also, a search for copper mines in the surrounding mountains needs to be conducted.
During the excavation at Jubah al-Jadidah, many artifacts and pieces of material culture evidence were found: bone, beads, seeds, wood, steatite bowls, obsidian, and other worked stones. Currently these items are being studied and prepared for publication.
The evidence for pre-Islamic Arabic culture in the Wadi al-Jubah is thus very extensive and very complex. The most significant data thus far are the radiocarbon dates which have come from Hajar at-Tamrah, Jubah al-Jadidah, and other sites in the wadi. Although the dates must still be considered tentative until confirming evidence can be obtained (such as from the lowest unexcavated layers in the probe at Jubah al-Jadidah, or from other excavated sites elsewhere in Yemen or Arabia), it is nonetheless clear that the dates line up strongly in favor of the ‘early’ chronology of Jamme, Van Beek, and others. The dates suggest that the South Arabian cultures began to form as early as the 13th century B.C., which is about the same time that the camel seems to have been introduced in the Near East as a common beast of burden (see Bulliet 1975). The camel bones in our probes may agree with this reconstruction, and may constitute evidence for a breakthrough in transportation that helped to facilitate the development of these ‘remote’ cultures. The dates also support the biblical tradition of the Queen of Sheba (ca. 950 B.C., contemporary with Solomon), as well as the later Assyrian and Babylonian references to the Sabaeans and other Arabs. Although the Wadi al-Jubah may have been abandoned after ca. 400 to 300 B.C. for reasons which are not yet understood, it is clear that pre-Islamic Arabic culture continued to flourish into later centuries at other sites, such as Marib and el-Fau (Al-Ansary 1981). Since the Islamic periods are not represented at Wadi al-Jubah or other nearby areas, it seems likely that these inland caravan and agricultural regions declined and remained backwater areas during those periods, when Islamic Arabic culture flourished elsewhere (especially to the west).
Modern Period
Numerous examples of Modern occupation were also found in the course of the reconnaissance in the Wadi al-Jubah. Pottery (Fig. 13), artifacts, structures, and cemeteries all attest to Modern occupation. All of these remains, however, seem to be at most about 100 years old. No earlier examples of Islamic pottery or coins were found. Informants tell us that “everybody’s great grandfather” came from Jerishah, the small village at the north end of the wadi. This may indicate that no one resided in the wadi before that time, but this topic has not yet been studied.
Future Work
A reconnaissance survey and two stratigraphic probes have brought us to this point. We now gasp the types of remains to be found in the wadi and we have ideas as to what they mean. In our next two campaigns, we will address three major areas of study:
First, we plan to continue excavating our probe at Jubah al-Jadidah. We feel that the ceramic and radiocarbon evidence is readily available there, and we need to get more and earlier remains to complete the sequence.
Second, we plan to complete an epigraphic survey of all known inscriptions in the wadi so that they can be published. This should be completed at the end of our next season.
Third, and of ultimate importance, we plan to bring in specialists to conduct a systematic study of the irrigation/agricultural remains. This will entail both geomorphological and botanical investigation of the preserved remains. At the conclusion of that work, systematic mapping, collection, and study of remains in the best preserved area will take place. From this evidence we hope to learn in detail how and when our irrigation/agricultural system actually worked.
Many other studies, such as an investigation of the ‘Neolithic’ remains, the cemeteries, or mountain occupations, must wait for the future. It also remains to be seen if large-scale excavation of some of the occupational sites in the wadi might be carried out in the future.