Visit Us | Contact Us | Membership | Make a Gift | Calendar | Order Online | What's New

Print this Page

Home > Research > Publications > Annual Reports > 1998-99 Annual Report

Hadir Qinnasrin Excavations


Donald Whitcomb

This report presents a new excavation on behalf of the Oriental Institute in Syria. Work at Hadir Qinnasrin began as a truly international project. Professor Marianne Barrucand of the Sorbonne was the principal organizer of the project with the active assistance of the Institut français d'études arabes de Damas (IFEAD) and Institut français d'archéologie du proche-orient (IFAPO); our thanks especially to MM. Dominique Mallet and Jean-Marie Dentzer, the respective directors. Our valued guide to excavation and survey in northern Syria was Professor Claus-Peter Haase of the University of Copenhagen. The project enjoys the support of the Directorate general of Antiquities and its director, Professor Dr. Sultan Muhesen; Ms. Fedwa Abido from Aleppo served as our department representative. I functioned as the field director for the project. Two French topographers and French and Syrian students completed the international team. The next season, scheduled for fall 2000, we hope will have much larger American - and Oriental Institute - representation.

Hadir Qinnasrin is located about 25 km (10 miles) south of Aleppo, just east of the main road connecting Aleppo with Hama, Homs, and Damascus. Being so close to Aleppo, one may note first some of the archaeological aspects of that city. Aleppo was known as Beroia in the classical period; its classical structure, as Sauvaget demonstrated in the seminal study of its topographic transformations (1941), was still evident at the time of the Muslim conquest (ca. 638). The Seleucid foundation probably initiated the street pattern with insulae of 100 x 45 m and juxtaposition of the town and a citadel, a common Seleucid settlement pattern; indeed this pattern is probably much older as witnessed in the citadel itself. Little is known about the extent of destruction inflicted on Aleppo during the Persian attack of Khosrow in 540 or its reconstruction under Justinian. The apparent lack of attention paid to Aleppo after the Muslim conquest suggests that Aleppo was a minor center during the early Islamic period and revived only as the Hamdanid capital in the tenth century.

The Muslims selected another settlement as their administrative center (jund capital) for northern Syria. This was Qinnasrin, long identified as Chalcis ad Belum. The layout of Chalcis with its citadel and lower town (see below) made of white limestone suggests another Seleucid foundation. As Tate notes, "the wall of Chalcis … delimited a space … measuring about 1000 x 900 m, remarkable for a city of the interior." The town functioned as a Byzantine military center on the road from Antioch toward the Euphrates, the key of the eastern limes (frontier). During the Muslim conquests in northern Syria, Chalcis was the scene of a treaty (in 637, according to Kaegi), which allowed the Byzantines to evacuate civilians and forces from the region. The site of Chalcis (later known as Eski Halab) is a mound beneath the shrine of Nabi ºIs; it was first mapped by Brossé in 1919 and again by Lauffray in 1942/43, and both descriptions and photographs were included in the historical topography of Mouterde and Poidebard in 1945. While there is little doubt that this long tradition of French scholarship has correctly identified the classical site of Chalcis ad Belum, the continuation of this site in the early Islamic period as the jund capital of Qinnasrin is less studied and assured.

Qinnasrin was founded as a new jund by Mu'awiya while he was governor of Syria in 643. There are other reports that Mu'awiya's son, Yazid I, both "destroyed the walls of the city" and "built the capital of the jund" in 680. A cursory examination of Chalcis in 1990 begins to explain this apparent confusion. The classical site had very few Islamic artifacts, suggesting that Islamic Qinnasrin should be sought elsewhere. Yazid had dismantled Byzantine defenses of the old city and built a new military headquarters and capital (i.e., a misr). The solution came from a survey by Claus-Peter Haase, who noted a low, mounded site 4 km to the east by the name of Hadir. This name excited our interest since Aleppo and many other classical cities of Syria had pre-Islamic Arab camps outside the city walls. Such camps are frequently described as transient assembly points for the great commercial caravans and their inhabitants as bedouin organized as caranvaneers. Irfan Shahid suggests that these camps soon became established settlements with permanent architecture, called parembolais in Greek and hadir in Arabic. The hadir was an ethnic suburb inhabited by Arab tribesmen; as the historian Baladhuri notes, "the hadir of Qinnasrin had been settled by the Banu Tanukh since they first arrived in al-Sham and pitched their tents there. Later they built their houses in it." The relationship of these suburbs and their adjoining cities seems similar to the association of Islamic urban foundations and an older pre-Islamic city: for example, Fustat with Babylon; al-Basra with Khuraiba; and Kufa with al-Hira. Our working hypothesis is that the Muslim conquerors avoided the alien, classical cities and settled in the hadir among fellow Arabs. The resulting settlements developed into cities of a distinctly Arabian type; the incipient Islamic city may be recognized in this urban form. The hadir became a madina (the traditional term for city) in a land where there had been little Byzantine building and an overall stagnation of urban life.

Armed with this hypothesis on the historical importance of Hadir Qinnasrin, we arrived at the site for a brief season of survey and sondages. One of the first things we noticed was that Hadir was no longer a village and the town of Hadir had expanded to encompass much of the low mounded area of the former city. The initial survey, or better a reconnaissance, of the town and its periphery was necessarily a matter of chance observations within empty lots, gardens, and fallow fields. This situation and the limited resources available made a systematic survey unfeasible at this preliminary stage. The oldest portion of Hadir appears to be centered on the mosque, which may overlie an earlier structure (see fig. 2). Immediately west of the mosque is the cemetery, measuring about 250 x 150 m; its contours and dense accumulations of sherds resembles occupation mounding and suggests an earlier residential function. A long stone in the cemetery, possibly a lintel, carries an apparent Kufic inscription.

The survey began with the most obvious antiquity, the mound of Tell Hadir (area A). The tell covers an area about 500 m in diameter and towers some 15 m above the springs, which lie immediately on its north side. The artifacts collected on the mound and to its west (areas B and C) belong to the Bronze Age and indeed these early materials were scattered throughout the survey and excavations. Field scatters to the west, north, and south were investigated; the latter suggested the possibility of occupation and indicated the need for further survey south of the mound, as local informants had suggested. The fields north of the tell (areas H, H1) produced sherds that, when added to the fragmentary architectural fragments near the present Baladiya offices, suggest a Byzantine or transitional occupation in this vicinity. Further collections were made along the west slope of modern occupation (roughly between the 265 and 270 m contours). This slope produced a consistent range of collection as far as the Wadi Turab (area P), where modern fields hindered further investigation; the area is currently known as Rasm al-Ahmar. There was no indication of architectural remains visible on the surface; nevertheless, this area was selected to attempt two soundings, areas L and K.


Area L was a very promising mound composed of a distinctive red-yellow soil; in fact, it proved too promising, being made of mudbrick of a most difficult type to delineate. This exercise in frustration was alleviated only by the excellent ceramics and other artifacts recovered. The sondage in area K proved to be luckier; this was a flat area, hemmed in by houses, revealing walls made of small stones immediately beneath the surface. During the 10 days available, a pattern of stone walls emerged and in the last few days, antecedent walls made of mudbrick were found.

While the ceramic types are very similar in both areas K and L, the structural remains in area K present two distinct phases: the earlier phase was a series of mudbrick walls forming a two-room building while the later phase saw these walls replaced by stone walls in the same configuration. The first ceramic phase is characterized by cooking pots, storage vessels, and other ceramic objects well known from north Syria and typified in the late Byzantine site of Déhès. However, a number of ceramic types common at Déhès appear to be absent from Hadir Qinnasrin, suggesting the chronological overlap is not complete. At present, it appears that the correlation of these and other sites in north Syria point to a very late Byzantine, or better, a transitional Byzantine/early Islamic date, roughly late seventh or early eighth century.

It seems that the architectural renovation of phase 2 follows directly with little or no gap in the occupation. The second ceramic phase is characterized by the introduction of glazed wares. The style of these first glazed wares is perhaps the most interesting and important result of this preliminary excavation. This ceramic type is a splashed polychrome style of yellow or white base with green and brown accents. The forms are invariably simple bowls with a low ring base. This has been labeled "yellow ware" by Oliver Watson on the basis of production collections from Tell Aswad near Raqqa. It is also found at Rusafa, Madinat al-Far, Qasr al-Hayr al-Sharqi and other early Islamic sites in north Syria.

This common glazed ware is also well known from excavations at Abu Gosh and Khirbat al-Mafjar in Palestine. More recently, the materials from Tell Yoqne'am (Qaymun, near Haifa) have clearly demonstrated the transitional nature of this ware between Coptic glazed ware (a late Umayyad Levantine tradition) and the introduction of Samarran wares (an eastern tradition of the early ninth century). Furthermore, the evolution of Coptic glazed ware (or slip-painted lead-glazed ware, according to Scanlon) suggests an Egyptian antecedent for this new style, which was subsequently developed in Bilad al-Sham. As often as this ware has been found, its stratigraphic relationship with late Byzantine/early Islamic ceramic styles has not been clear until the trenches at Hadir Qinnasrin.

The preliminary interpretation of the ceramic collections from Hadir Qinnasrin is that this innovation of ceramic production, the technique of glazing that becomes a diagnostic indicator of early Islamic culture, was a common feature of pre-Samarran culture in Syria. Moreover, the relationship of "yellow ware" to brittle wares, molded juglets, and early sgraffiato becomes clearer as these fall into a developmental sequence in these excavations. One may note that these excavations produced no evidence of occupation after the tenth century. This was predictable from historical accounts of the destruction and decline of Qinnasrin. Thus the present collections of ceramics offer a clear separation of early Islamic from the common Ayyubid/Mamluk types in north Syria. While many indications must remain preliminary, these ceramics seem to offer chronological diagnostics for the eighth century (a minimal period of ad 700 to 800), an assemblage of glazed and common types that may refine the dating of excavations and surveys in Syria for the early Islamic period.

From Tent to House to Arab City

If we return to the plan of area K (see fig. 4), one sees an apparently freestanding structure composed of two rectangular rooms. Moreover, the plan seems to have been made first in mudbrick and then duplicated in stone cobbles. The smaller of the rooms has a couple of ovens and storage vessels. One of the more curious features is the south (actually southwest) wall, which had a series of column bases and was mostly open. These features add up to a very specific house form, a type derived from the "black" tent used by Arab tribes in the recent past. The transition of a nomad tent into more permanent material is not so uncommon. Ethnographic study of nomad tribes in Syria has documented a temporary house called a sibat, used for seasonal occupation and very similar to the remains that we uncovered.

The literature on Arab tribes in Syria is extensive and much of this documentation has concerned typologies and processes of settlement. This is hardly surprising in that much of the impetus to study was from mandate and later administrations with an interest in settlement of "nomad" groups. In his study of the archaeological evidence of nomads, Cribb makes an important point that the nomad camp is always, despite its appearance to occasional outsiders, a highly structured spatial arrangement of residential units. In the case of Qinnasrin, it seems likely that the form of the hadir, the original camp structure, influenced the development of the settlement. Such camp settlements were part of the common experience of tribes in Arabia and brought with the Tanukh and Tayy tribes when they entered Syria. The early Muslims must have found this a conducive environment when they entered Syria and founded their amsar. The amsar were more than camps, however, but founded as urban administrative centers upon Arab urban models. The sites of 'Anjar in Lebanon or Aqaba in southern Jordan may contain reflections of these models and their adaptations. When the early Muslims rejected settlement in Aleppo and Chalcis in favor of Hadir Qinnasrin, they might have selected the familiar ethnic pattern of Arab tribes and at the same time initiated a fundamental step in the development of the early Islamic city, an urban type to have wide influence in the formation of Muslim communities in Syria and throughout the Middle East.

Revised: July 30, 2007

Home > Research > Publications > Annual Reports > 1998-99 Annual Report