The construction phases of Khor Rori's monumental gate

Arab. arch. epig. 2001: 12: 249–259 Copyright C Munksgaard 2001Printed in Denmark. All rights reserved

ISSN 0905-7196

The construction phases of Khor Rori’smonumental gate

ALESSANDRA AVANZINI AND ROBERTO ORAZIPisa, Italy

The choice of the location for the foundation of the city of Sumhuram wassuggested to the Hadrami colonisers by numerous factors (presence of drinkingwater, perfect visibility of the location from a distance, etc.). In order to survivefor more than three hundred years the city needed to defend itself and reinforceits defensive system throughout the whole course of its history. In 1999–2000MID undertook studies of the defensive system of the suburban gate complexof Sumhuram. This study which is based on architectonic and epigraphic data,led not only to a better understanding of the construction phases of the gate,but also to the elaboration of an hypothesis about the beginning of the historyof the city. It is our belief that at the end of the first century BC a first settlementwas founded with mainly a defensive function. In the first century AD it wastransformed into a city.

The foundation of Khor Rori, South Ar-abian Sumhuram (1), by the king of Hadra-mawt, is linked to the development of seatrade at the end of the first century BC be-tween the Mediterranean and India. As iswell known, the political-economical de-velopment of the Hadrami kingdom (2) inthis period depended on its technical andmilitary ability to control the coastal regionand arm a fleet.

Dating the foundation of Sumhuram, orof a partially-walled previous settlement,to the second half of the first century BCleads us to at least two general obser-vations. First of all, it undermines one ofthe arguments brought by N. Groom (3)against the identification of Sumhuramwith the city of Moscha mentioned in thePeriplus. According to Groom, around thefirst half of the first century AD, at the timeof the journey that the author of the Peri-

249

plus took along the Arabian coasts, Sum-huram had not yet been built.

Secondly, since a harbour situated atsuch a distance from its mother countrywould not have been able to survivewithout any sea link with other harbourson the Arabian coast, we can affirm confi-dently that, at least starting from the firstcentury BC, a commercial route existedthat linked South Arabia to India. The im-portance of Arabian trade as a trait d’unionbetween the Indian Ocean, the Red Sea andthe Mediterranean for the beginning ofcommercial exchanges between Arabia andIndia has often been undervalued byscholars, who tend to overestimate the roleplayed by Rome or the late Ptolemies (4).

The choice of the site at which to foundSumhuram (Fig. 1 and 2) is certainly linkedto numerous factors. Firstly the ‘geo-graphical’ factors: on the one hand the rela-

A. AVANZINI AND R. ORAZI

Fig. 1.Map of the region (drawing byVincenzo Labianca).

tive proximity to the area where incensewas produced, to which Sumhuram wasconnected by a network of settlements, onthe other the convenient distance, by sea,from the Hadrami harbour of Qana. Sec-ondly the ‘territorial’ factors: Sumhuram,like Qana, is very clearly visible from thesea thanks to the two rocky promontoriesthat protect the mouth of the wadi Darbat.The whole area along the coast seems toabound in precious sources of drinkingwater.

At this stage in our research, it is moredifficult to trace a more precise ‘historical’context. It is very likely that the city per-formed different functions: as ‘harbour’,

250

‘fortress’, Hadrami ‘colony’ in a distantand scarcely inhabited land, military ‘out-post’, proof of a real ‘territorial expansion’,or at least of the intention to expand theHadrami kingdom, east of a region inhab-ited by people with whom it was some-times possible to establish an alliance, butfrom whom it was always necessary to de-fend oneself.

What clearly emerges, however, is thatthroughout the whole of its history the in-habitants of Sumhuram continued to adddefensive structures in order to protect thecity, until it possessed one of the most com-plex defensive systems of South Arabianculture.

THE CONSTRUCTION PHASES OF KHOR RORI’S MONUMENTAL GATE

Fig. 2.Map of the Khor Rori region.

The choice of the site for the foundationof the city, thus, proves the existence of anincreasing conflict between convenienceand security that lasted throughout thewhole history of Sumhuram.

Sumhuram, a small town surrounded bywalls (Fig. 3), was built on a rocky hill notfar from the ocean and was defended onthree sides by the natural inclination of theground. Thus all its defensive structures

251

are on the northern side which looks overthe interior and a relatively flat area. It ison this side that the city gate opens out:three different doors blocked a rising paththat was constantly controlled by a centraltower and protected from the outside by astrong isolated wall.

The MID focused its attention on this lastbuilding during the 1999–2000 expedition.The study of the architectonic structure of


Fig. 3.Plan of Sumhuram.

Sumhuram’s monumental gate and its epi-graphic documentation has provided uswith some interesting clues that have ledto a better understanding of the history ofthe city.

We were essentially wondering whetherthe construction of the whole complex hadbeen carried out all at once or whether ithad been realised in the course of differenthistorical periods. Initially both the inscrip-tions and the analysis of the architectonicstructures seemed to point to a unitary con-struction phase for the whole complex.

Before cleaning the city gate complex of

252

the back fill left on the site by previous ar-chaeological expeditions, it seemed thatonly a small wall (Fig. 4, M13) had beenadded to the tower in order to build theeastern jamb of the more external door,seemingly an afterthought in a unitarybuilding plan. Four inscriptions (our in-scriptions 3–6) seemed to confirm a unitarybuilding phase for the city gate.

The corpus of gate inscriptions consistsof eight monumental inscriptions (5). Onlytwo of them are still in situ. It is possible tolocate the position of the inscriptions alongthe entrance path to the city thanks to Al-


Fig. 4.Schematic drawing of the gate area (by Vincenzo Labianca).

bright’s map. On the new map of the gatedrawn by the MID, which from a plani-metric point of view is considerably dif-ferent from the one published by Albright,we have indicated the inscriptions dis-cussed here (Fig. 5).

Inscriptions 3–6 form a single corpus andare characterised by the presence of thesame characters (the king of Hadramawt Il-ıcazz Yalutw, and the chief of the troop ofHadramawt in the country of Sakalan:Abıyathac` Salhw an son of Dhamarcalı). Inthese inscriptions the foundation and thecolonisation of Sumhuram by the inhabi-tants of Shabwa is mentioned. If the kingmentioned is the Eleazos of the Periplus,this corpus can be dated to the first half ofthe first century AD.

253

They also follow the same textual model:

inscription 2: NP cbd $lcdd Yltw mlk Hw dw rmt[...] qtdm hgrhn S1mrm grbt-td [...] bn rbbm$d s2qrm [...] b-mwsw t w-thw rg mr$-s1 $bytd w c

S1lhw n bn Dd mrcly qdm gys2 Hw dw rmt b-$rdwS1$klnand inscription 3: NP cbd $lcdd Yltw mlkHw dw rmt [...] qtdm grbtn w-bny hgrhn S1mrm[...] bn rbbm $d s2qrt [...] b-mwsw t w-thw rgmr$-s1 $bytd c S1lhw m bn Dd mrcly qdm gys2

Hw dw rmt b-$rdw S1$klninscription 4: NP [...] tbc w-s2yc mr$-s1mn$bytd c bn Dd mrcly [...] w-hw wr cm-s1 s2ltd t$hw yym bnmw ktbc $d s1tfhwand inscription 5: NP [...] tbc w-s2yc $bytd c

bn Dd mrcly [...] w-hw wr cm-s1 s2ltd t $hw yymbnmw ktbc $d s1tfhw


Fig. 5.City gate and location of inscriptions (drawing byVincenzo Labianca).

Since inscriptions 3, 4, 5 and 6 are con-temporary we tended also to include in thisphase the construction of the tower and thethird door (Fig. 5).

Inscriptions 7 and 8, however, do not fitwell in this picture.

On the very eastern part of the same wallwhere inscriptions 5 and 6 were located (M10), inscription 7 mentioned the king Yad-ac$ab son of Ghaylan. If the king mentionedis the same attested at Hajar bin Humeid,after the collapse of Qataban, this inscrip-tion could be dated to the beginning of thethird century. It is evident that an inscrip-tion on its own is not enough to determinethe date of a monument: on the wall builtin the first half of the first century AD,many years later, during the kingdom of

254

Fig. 6.Southern wall of the tower. Vertical cut between theeastern and southern walls.

Yadac$ab, an important character of the cityleaves an eternal memory of his name.

But more serious problems arose with in-scription 8 located on the western wall ofthe tower. In inscription 8 the name of theking is not mentioned (6), but paleographicand textual analysis compels us to thinkthat this inscription is of a considerablylater date than inscriptions 3–6 (7). Also in-scription 8, like 7, could have been writtenin a later phase, but the text clearly men-tions the construction of the tower and thethird door in order to improve the defenceof the city. Frankly, it is difficult to acceptthe hypothesis that this inscription wasadded decades after the construction of thetower and the third door.

For a long time an answer provedelusive, until a closer analysis of thesouthern wall of the tower (M10) providedus with the key to solving the problem. In-deed, we could see how the lowest part ofthe wall, and the foundation on which ithad been built, penetrated inside theeastern wall without being connected to itin any way. This meant not only that theeastern wall had been built in a laterperiod, but also that the other walls of thetower, which were perfectly bonded toeach other (and thus the tower itself), had


Fig. 7.City gate, plan of phase 1.

been built when the southern wall alreadyexisted.

An equally important observation wasmade near the eastern corner of theinternal side of the southern wall. A ver-tical cut, 3 or 4 cm wide, separated thelower part of the eastern and southernwalls, precisely where they were made ofperfectly squared ashlars (Fig. 6). Albrighthad noticed this indication of separation,but on the basis of this he had hypoth-esised that the construction of the towerhad started in the south-eastern corner andthat the brickwork, after moving in an anti-clockwise direction arrived back at thestarting point (8).

255

Fig. 8.City gate, reconstruction of phase 1.


Fig. 9.City gate, plan of phase 2.

Fig. 10.City gate, reconstruction of phase 2.

256

During the October–December 1999campaign, when we started the cleaning ofthe area between the buttresses of themiddle door and the southern wall of thetower, we were able to arrive at a more pre-cise explanation. Near the eastern corner ofthat wall we unearthed some sort of a rec-tangular platform which appeared to beperfectly connected to the brickwork of thesouthern wall, and as we found out later,this was the partly demolished remains ofthe wall itself (Fig. 4).

Thus, during the first phase the gate issimilar to the one found at Naqab al-Hajar(9) – it is an L-shaped gate – but it alreadyhas an external wall, like the gate of the


final phase; the eastern and western but-tresses and the path that leads from themiddle door to the more northern internalone clearly belong to the same period (Figs7 and 8). The last phase seems to be essenti-ally linked to the construction of the tower(A6), the third door and the external wall(Figs 9 and 10).

It is not possible to determine the lengthof the interval that separated the twophases but, since we think that inscriptions7 and 8 date back to this second phase, wecan hypothesise that the interval betweenthe two phases was fairly long. The con-clusions drawn on the basis of the analysisof the architectonic structure coincide withthe ones drawn on the basis of the inscrip-tions.

Once the two main building phases hadbeen determined, we were still left withone contradiction to be solved. In a struc-ture which was undoubtedly unitary, likethat of the L-shaped entrance path to thecity, inscription 1 and inscriptions 3 and 4,which are not contemporary, described indifferent words the same event.

In inscription 1 (see Fig. 5), even thoughit is very corrupted, we can read the nameof the mukarrib of the Hadramawt Yash-hur$ıl and of one of the buttresses of thegate. The inscription, thus, can be dated tothe second half of the first century BC. Cel-ebrating the construction of one of the but-tresses of the only city gate was equivalentin effect to celebrating the very foundationof the city.

Despite all the difficulties met in thetranslation of the individual words andsome expressions, inscriptions 3–6, whichbelong to the Eleazos phase (a few decadeslater), undoubtedly mention the construc-tion of the city of Sumhuram and its colon-isation by the Hadrami inhabitants.

An archaeological datum and com-parison with two architectonic structureshave helped us to formulate an hypothesis

257

that, even though it still needs to be veri-fied by future research, seems to be ex-tremely credible. First of all we have totake into consideration the fact that somesoundings carried out by the MID in 1998on the southern side of the city, dated thepaving of a house to the last quarter of thefirst century BC on the basis of C14analysis. This phase was characterised bythe same type of ceramics used in the fol-lowing phase. At that time we were notable to determine whether that buildingbelonged to a settlement preceding thefoundation of Sumhuram, or Sumhuram it-self, the foundation of which had to bedated to the end of the first century BC. Itwas also impossible to ascertain whetherthis settlement was walled or not.

Fig. 11.Fortress of al-Bina$.


The second fundamental element thatled us to our hypothesis is provided bytypological comparison with the al-Bina$

fortress, a strong wall built 30 km north ofthe harbour of Qana. The fortress of al-Bina$ (Fig. 11) seems to consist of a longwall, with an entrance gate, but without acity wall in the proper sense of the term. Inthe inscription RES 2687 on the gate of theal-Bina$ fortress (10), we find mentioned –according to an interpretation proposed byRobin which is perfectly acceptable – thesame mukkarrib Yashhur$ıl of our inscrip-tion 1 (11). In this inscription the term hgr‘city’ is never used to refer to the construc-tion, rather cqt ‘fortress’.

Finally, a similar structure which is stillvisible on the Inqitat Mirbat had long beenknown to us: here a long wall, which alsoin this case is provided with a gate, blockedout the whole northern side of the promon-tory, and thus, defended the side thatlooked over the interior, ie. the only sidethat could be attacked, since the other ones,that looked over the ocean, were totally im-pregnable. This comparison is based onlyon typological elements since evidence re-garding the chronological aspects of the In-qitat Mirbat’s structure is not yet available.

Thus, in the second half of the first cen-tury BC, when the kingdom of Hadramawtexperienced a phase of exceptional pros-perity, one of its kings, Yashhur$ıl, whotook up again the old and prestigious titleof mukarrib, is mentioned in the inscrip-tions of the gates of two fortresses whichhe had built to protect the harbours androutes that connected the city to the in-terior. This is a clear sign of a renewed in-terest of the kingdom in the coast and seatrade.

Fortresses not cities. As far as Sumhuramis concerned, we can imagine that at theend of the first century BC it had a de-fensive wall on the northern side withtowers and an L-shaped, fortified gate.

258

In the period of Eleazos the function ofthe fortress of Sumhuram changed: it wastransformed from a military outpost into areal city that was colonised by the inhabi-tants of Shabwa (in inscription 4 line 5, hgris not the noun ‘city’, but a verb – a veryexceptional case –, probably in the II form:mt bny w-hgr Smrm ‘when he built andmade of Sumhuram a city’).

We think we have come to a betterunderstanding not only of the constructionphases of the city gate of Sumhuram, butalso of the history of the city. This is a goodexample of the results that can be obtainedthanks to the collaboration of scholarsworking in different fields. The inscriptionsof the gate of Khor Rori, if seen outsidetheir context and without the under-standing of the architectural phases, wouldbe almost completely robbed of their his-torical meaning. Without the inscriptionsthe two building phases at the beginningof the history of Sumhuram would have re-mained for the most part obscure.

References1. The city was partially excavated by a US mission

in the fifties, under the direction of Wendell Phil-lips, with Frank Albright as chief archaeologist(Albright F. The American Archaeological Expeditionin Dhofar, Oman, 1952–1953. Washington: AFSM,1982). An expedition organised by the Universityof Pisa (MID – Missione Italiana in Dhofar), di-rected by Alessandra Avanzini and with RobertoOrazi as Field Director, has been working on thesite since 1996 (see Avanzini A. La Missione Ital-iana nel Dhofar. EVO 19: 1996: 181–241). Duringthe first three years short campaigns of one monthwere arranged and the archaeological surveyswere coordinated by Daniele Morandi. Since 1999the MID has been working on the site for a muchlonger period (five months). The core element ofMID’s scientific project is the restoration of thecity and the study of its architectonic structure.

2. The history of Hadramawt in this period is fairlywell known. In the first century BC Hadramawttook part in some wars against the Qatabaniankingdom, allied itself with Saba and with some


tribes that in the following century would formthe nucleus of the Himyarite power (RES 4336).At the end of the first century BC Hadramawt en-joyed a period of prosperity. One of its rulers,Yashhur$ıl, for the last time in South Arabia’s his-tory, took the title of mukarrib. At the end of thefirst century AD Hadramawt was ruled by $lcz Yltwson of Ydc$l, identified with the king Eleazos men-tioned in the Periplus of the Erythrean sea. At theend of the first century Hadramawt foughtagainst Saba in Jawf. For the first part of thesecond century AD we do not have any historicaldocumentation, and this might not be due tochance. We cannot say whether Hadramawt wentthrough a period of political crisis or whether theequilibrium reached with Himyar and Saba led toa period of peace; what can be stated without anydoubt, however, is that the conflicts betweenSouth-Arabia’s states re-ignited in the second halfof the second century and led at the end of thecentury to the collapse of the kingdom of Qatabanwhich was largely annexed by the kingdom ofHadramawt. Ydc$b Gyln surrounded with walls –and thus took possession of – Hajar bin Humeid,the last Qatabanian capital. However, wars weredestined to continue; in trying to establish an al-liance with Saba, Hadramawt identified in the Hi-myarite kingdom the real danger to its future. Atthe end of the third century AD after alternatingmilitary fortunes, Hadramawt was annexed byHimyar.

3. Groom N. The Periplus, Pliny and Arabia. AAE 6:1995: 180–195.

4. De Romanis F. Cassia Cinnamomo Ossidiana. Roma:Saggi di Storia Antica, 9: 1996: 143–144.

5. The inscriptions of Khor Rori have been editedseveral times, from the complete corpus by A.Jamme to the excellent editions of some of the in-scriptions by A.F.L. Beeston and W.W. Müller.

6. The god Sin dhu-Ilum is mentioned. In the gate’scorpus the name of the god is mentioned also ininscription 2 (now in the museum of Salalah). Atthe end of the inscription we can clearly read: b-td cdS1yn, translated by Jamme ‘in the irrigated land of

259

Sin’, a translation that led archaeologists to debatethe question of which ‘irrigated land’ was men-tioned here. This is a good example of how mis-leading epigraphic texts can be when taken tooliterally. In fact, td cd certainly has a much more gen-eric meaning than the one proposed by Jammesince it also signifies ‘a piece of land’. Moreoverwe could also suppose that the scribe might havemade a mistake for: b-mcd S1yn ‘following the willof Sin’.

7. To realise this, we only need to look at the verbsof construction used in inscription 8 and inscrip-tions 3–6: br$ instead of bny and qtdm and thestereotypical expression ‘from the basis to thetop’: bn mwtd rm $d s2qrm instead of bn rbbm $ds2qrm.

8. Albright, American Archaeological Expedition: 18:‘Construction of the room walls apparently wasbegun at the Southeast corner running counter-clockwise around the room until it butted into itseast wall at the starting point.’

9. Breton J.-F. Les fortifications d’Arabie meridionale du72 au ler siecle avant notre ere: Mainz: Archäologi-sche Berichte aus dem Yemen, 8: 1994: 135–137.

10. Robin Ch. Yashhur$ıl Yuharcish, fils d’Abıyasac,mukarrib du Hw adw ramawt. Raydan 6: 1994: 101–111.

11. It is interesting to note that Ydd t$n, the name of thetower mentioned in Sumhuram inscription 1, isthe name of one of the towers of al-Bina$.

Address:Alessandra AvanziniDipartimento di scienze storiche del mondo anticoVia Galvani 156126 – PisaUniversity of Pisae-mail: avanzini/sta.unipi.it

Roberto OraziCNR- Istituto per le Tecnologie Applicate ai

Beni CulturaliVia Salaria km. 29.30000016 Monterotondo Scalo – Romae-mail: orazi/mlib.cnr.it

The construction phases of Khor Rori's monumental gate

Documents