Towards the earliest history of Kinda

Towards the earliest history of Kinda

IntroductionIn 106 AD the Nabataean kingdom was annexed byRome. Although the southernmost limit of theformer Nabataean territory under direct Romancontrol is still disputed (Graf 1988: 173–174, 182),Roman military garrisons of the Legio III Cyrenaicawere stationed in the mid-second century AD in themost important trade centres of Dedan, Mad�a’inS @alih@ and al-’Ul�a. According to the Greek, Latin andNabataean graffiti from these sites, the garrisons hadto protect the caravans, passing through H@ ij�az, fromthe raids of nomads (Graf 1978: 3, 11; Isaac 1989:247). However, Rome’s relations with the nomads of

north-west Arabia were not always hostile: the‘kings and tyrants of the Arabs’, as the Periplus ofthe Erythraean Sea says (§20: 7.9–11), tried to subduethe people of Kanraitai, who represented theprincipal danger for the Roman sea traders of themid- ⁄ late first century AD along the central part ofthe Arabian shore of the Red Sea. Later — in166 ⁄ 167–169 AD — the confederation of Tham�ud(Nab. srkt tmwdw = Greek œhmoV Qalotdgmîm) erecteda temple in honour of the Roman emperors MarcusAurelius and Lucius Verus at Ruww�afa (for theGreek-Nabataean bilingual inscription from thetemple there see Bowersock 1975: 513–522).

In the second century AD Rome could not protect the caravan trade in north-west Arabia to the south of Dedan with its own military detachments,stationed in Mad�a’in S @alih@ and al-’Ul�a. Rome seems to have attracted Aksumfor the subjugation of the Arab nomads of north-west Arabia and protectionof the ‘Incense Road’. At the end of the second century AD Rome, Aksumand Saba’ shared common interests: Aksum helped Rome to restore peace onthe caravan routes, in which Saba’ seemed to have been interested as well,and Rome provided a market for East African and South Arabian products.During the west Arabian campaign, described in the inscription RIE 277(Monumentum Adulitanum II), which is to be dated to the very end of thesecond century AD, the unnamed Aksumite king, presumably Gad�urat,conquered the lands of Kinaidokolpites and Arrabites. The former namerepresents a combination of two names, the first of which seems to be derivedfrom the name of Kinda (identification of the entire name with Kinda is takenfor ‘somewhat unlikely’ in Retso 2003: 450, n. 60). The Kinda seem to havebeen also known in the earlier classical tradition under the name of Kanraitai— the inhabitants of Ghamr dh��-Kinda — and thus were the most aggressiveand dangerous obstacle for the Roman traders on the land routes in north-west Arabia and in the northern Red Sea in the first–second century AD. TheAksumite invasion forced the resettlement of the Kinda on the other branchof the ‘Incense road’ — through modern Qaryat al-F�aw — and its blockade,consequently followed at the beginning of the third century by thecampaigns of the Sabaean king S�a’ir ‘Awtar against them.

Keywords: Kinaidokolpites, Kinda, Aksum, Saba’, caravan trade, Qaryatal-F�aw, Monumentum Adulitanum II, Claudius Ptolemy

M.D. BukharinInstitute of World History,Russian Academy of Sciences,Leninski pr. 32a, 119334Moscow, Russia

e-mail: misha_nina@ yahoo.com

Arab. arch. epig. 2009: 20: 64–80 (2009)

Printed in Singapore. All rights reserved

64

One of the most informative sources for thehistory of western Arabia in the first centuries ADis Monumentum Adulitanum II, an inscription copiedby Cosmas Indicopleustes at Aksum in 525 AD(Cosm. Ind. 2. 56; also referred to as OGIS 199 andRIE 277; for the latest edition see Bernand, Drewes &Schneider, 2000: 33). It describes the campaign of anunknown king of Aksum through East Africa to theborders of Egypt, the crossing of his arm to theopposite shore of the Red Sea, the submission ofKinaidokolpites and Arrabites, the restorationof peace on the sea and land routes, and the returnof the Aksumite army. It makes clear that Romecould not protect caravans to the south of Dedan byitself. This inscription never seems to have beenconsidered in the context of relations between Romeand the nomads of western Arabia.

The dating of RIE 277The dating and authorship of the inscription aredisputed. Currently, the text is variously dated‘earlier than the 3rd cent.’ (Desanges 1984: 258), tothe first quarter of the third century (Robin 1989: 150;Cuvigny & Robin, 1996: 710) or to the middle of thethird century AD (Shitomi 1994: 73–114; 1997: 81–102). However, if the combat described in RIE 277 isseen in the light of other Aksumite and SouthArabian inscriptions, then a more precise dating ofthe inscription seems possible.

According to RIE 277, the starting point of thecampaign must have been Adulis (not Aksum assuggested by Fiaccadori 2004: 130). Having com-pleted their campaign, the Aksumite army returnedto Adulis where the king made a sacrifice to Ares asthanks for a safe and successful return. Havingstarted his campaign in East Africa, the unnamedAksumite king had to conquer peoples extendingfrom the area of modern Somalia in the south andfrom the coast of the Red Sea in the east to theborders of Egypt. The first of these peoples were theG�afg œhmoV (RIE 277 ⁄ 3), the Ge’ez speaking Aksu-mites. In all probability the anonymous king onlybecame king of Aksum as a result of his East Africancampaign.

This does not contradict the information in thePeriplus of the Erythraean Sea although this describesAksum as a united kingdom with a royal residencedeep in the interior (4: 2. 8; 5: 2. 20). The political

situation in East Africa could change quickly andonce-united (or seemingly united) kingdoms mightsplit into a number of smaller chieftainships, oranother ruler, e.g. any negus of Gabaz (this title isknown from RIE 186 ⁄ 10) — the port of Adulis — asin the case of RIE 277, could rise to power and thenattempt to confirm that power throughout the entireland.

Several ancient South Arabian inscriptions pro-vide parallels to the information contained in RIE277. For example, according to CIH 308 ⁄ 12, the kingof Saba’, ‘Alhan Nahf�an, concluded a treaty withGad�urat and the Aksumite army (ms @r ⁄ ’h@bsn). Ja 631and Ja 635 mention only the Abyssinians (‘h@bsn) inSouth Arabia. According to Ja 631 ⁄ 13, S�a’ir ‘Awtar— the son of ‘Alhan Nahf�an — maintained friendlyrelations with Gad�urat, already calledmlk ⁄ h@bst ⁄ w’ksmn, the king of Abyssinia and theAksumites. Interestingly, h@bst is mentioned before‘ksmn. This might suggest that the conquest ofAksum by Gad�urat took place between the conclu-sion of the treaty with ‘Alhan Nahf�an and thebeginning of contacts with S�a’ir ‘Awtar.

RIE 277 does not tell us for certain whether theAksumite army came back to Aksum or Adulis afterconquering the East African peoples, or continueddirectly to the borders of Egypt, having crossed theRed Sea and reached Arabia. The first possibilityseems the least probable and it would have beenmore natural, having arrived in Egypt, to reach theeastern shore of the Red Sea.

According to Ja 631 ⁄ 21–22, the Abyssinian armythat invaded H@ imyar was under the command ofBYGT — son of the negus (Gad�urat) — by-gt ⁄ wld ⁄ ngsyn. Since RIE 277 does not mention anAksumite presence in South Arabia, but doesmention the subjugation of the Bec�a (RIE 277 ⁄ 12),the Aksumite conquest of East Africa and invasionof western Arabia (RIE 277) must have preceded thatof H@ imyar mentioned in Ja 631. It follows thereforethat BYGT was not a native son of the negus, but achieftain of the conquered Bega and that wld in Ja631 was part of his royal title. This does not, ofcourse, exclude the possibility that wld coulddesignate truly paternal relations as, for example,in RIE 186 ⁄ 1: wldm ⁄ ’lm ⁄ ’mdm, in RIE 189 ⁄ 1–2 (cf. theanalogous expression bn ⁄ mh@ rmm in l.3 of thetrilingual inscription of ‘�Ez�an�a [Uhlig 2001: 15])and in RIE 185bis I ⁄ 3).

TOWARDS THE EARLIEST HISTORY OF KINDA

65

This inference is also based on the followingexamples from the inscriptions of ‘�Ez�an�a:wld ⁄ mh@ rmm (RIE 185II ⁄ 3), wld ⁄ lmh@ rmm (RIE 185bisII ⁄ 3-4), wldm ⁄ mh@ rmm (RIE 186 ⁄ 4) and walda ⁄ mah@ rem(RIE 187 ⁄ 3–4, RIE 188 ⁄ 5). The negus is calledmlk ⁄ mlkn (RIE 185 I ⁄ 2, RIE 185bis I ⁄ 3 etc.), i.e. Kingof Kings, who united under his power the kings ofthe conquered peoples. They were ‘children’ (wald)of the king of Aksum, in the same way as the negushimself was a ‘child’ of Mah@ rem.

As Ja 631 seems to show, the reasons behind theAksumite invasion of H@ imyar under Baygat was achange of power. The Sabaean campaign againstAksum in support of H@ imyar had been conductedby the tribe SMHRM YHWLD without the sanctionof the king of Saba’ (Korotayev 1994: 50–51). Such aninitiative could only be undertaken when there wasno king of Saba’, i.e. during an interregnum.

The treaty between Aksum and Saba’ (CIH 308)was concluded between the negus Gad�urat and‘Alhan Nahf�an and S�a’ir ‘Awtar, the kings of Saba’.The Aksumite invasion of H@ imyar, consequentunification of the armies of the warring dynastiesof Saba’ and dh�u-Rayd�an — the latter representedby Li’az��z Yuhan��f Yuhas@diq (Ja 631 ⁄ 26) — againstAksum, and the dating of Ja 631 to the reign ofLuh@ ay’athat Yarkhum, king of Saba’ and dh�u-Rayd�an, all point to a date for the invasion in aninterregnum. On the one hand, S�a’ir ‘Awtar was nolonger king, but he had not yet been replaced byLuh@ ay’athat Yarkhum. On the other hand, thetreaty between Saba’ and Aksum was no longervalid and Gad�urat was already free of formerobligations.

Sequence of eventsThe most likely sequence of events seems to be thefollowing: having consolidated his own power onthe western shores of the Red Sea, the Aksumiteking — author of the RIE 277 and presumablyGad�urat — concluded a treaty with ‘AlhanNahf�an, king of Saba’ and dh�u-Rayd�an, therebygaining an ally in the southern part of the ArabianPeninsula, and thereafter sent his army to LeukeKome on the coast of western Arabia, eventuallyreaching Najr�an (the Aksumites did not invadeSaba’). ‘Alhan Nahf�an, the Hamdanite, havingcome to the throne, sought support to consolidate

his power and found it in the person of theAksumite negus Gad�urat.

The situation described in RIE 277 finds certainparallels in Robin-Umm Layl�a 1 ⁄ 5–6, 10, dated tothe period c. 200–270 AD (Robin 1989: 147, 152),which commemorates the decision of the tribeKhawl�an Gud�ad�an and its allies to strengthentheir fortification against the coming Abyssinians.The text was found c. 50 km to the north-west ofS @a’da (Robin 1982: 2) at practically the samelatitude as Najr�an and the port of J��z�an. Itmentions ‘chiefs’ (‘mr’) of Khawl�an, the kings ofSaba’ (‘mlk ⁄ sb’) and the Abyssinian army whichseems to have been coming from the north. Onemay suppose that the strengthening of fortifica-tions against the oncoming Aksumite army wasprompted by unfriendly relations with Saba’ andtheir allies the Aksumites. It is possible that thecampaign of Saba’ against Khawl�an described inCIH 308 ⁄ 19–24 was also waged against Khawl�anGadud�an.

If correct, these considerations would permit us todate RIE 277 to the beginning of the reign of ‘AlhanNahf�an.

Aim of the campaignThe Aksumite king who authored RIE 277 says thathe subdued Kinaidokolpites and Arrabites and in sodoing restored peace on the trade routes on sea andon land. His campaign was aimed at ‘ÐdeÚerhai les’e„q»mgV jaˆ pkšerhai’ (RIE 277 ⁄ 30). It follows thatthe caravan traffic through the entirety of westernArabia was endangered and that Aksum wasgreatly interested both in the peaceful continuationof the caravan trade through Arabia and of the seatrade along the eastern shore of the Red Sea.Nothing is known of the Aksumite caravans inArabia, or of potential Aksumite aid to SouthArabian caravans. As such, the reference must beto Romans caravans.

The presence of a coin of the third centuryRoman emperor Gallienus (257–258) in the H@ ij�az(Potts 1994: 218) and of an Aksumite militaryrepresentative at Najr�an (Ja 577 ⁄ 10–13) makeAksumite-Roman cooperation for the protection ofthe trade routes in Arabia plausible. Some laterByzantine sources are also of interest in this regard.According to the Chronographia of Theophanes the

M.D. BUKHARIN

66

Confessor, trade along the land routes betweenAksum, H@ imyar and Rome was conducted byByzantine traders:

oƒ d�e sîm `Pxla…xm pqaclasetsaˆ di�a soà `Olgq…sote„ršqvomsai ™pˆ sÕm AÙnotl…sgm jaˆ s�a ™mdoseqalšqg sîm 'Imdîm jaˆ A„hiÒpxm

[Romans conduct trade through H@ imyar tillAksum and the innermost parts of Indians andEthiopians (223)].

This information is confirmed by Johannes Malalas:

oƒ d�e pqaclasetsaˆ `Pxla…xm di�a sîm `Olgqisîme„ršqvomsai e„V s¾m AÙnoÚlgm jaˆ ™pˆ s�a ™mdÒseqabar…keia sîm 'Imdîm… sîm oâm pqaclasetsîme…rekhÒmsxm e„V s¾m vèqam sîm 'Aleqisîm ™pˆ sÕ

poi»rarhai pqaclase…am,… D…lmoV Ð barikeÝV sîm'Aleqisîm, ™uÒmetrem aÙsoÝV jaˆ p�amsa s�a aÙsîm¢ue…keso…

[The Roman traders come to the Aksumitesthrough the Amerites in the inner kingdoms ofthe Indies … so, some of the traders have comeinto the land of the Amerites for trade… Dimnos,the king of the Amerites, has killed them andtaken all their [possessions] (433)].

Here the Greek name D…lmoV must be derived fromdh�u-yamanat. Hence it was not a proper name, butpart of a royal title. The adherence to Judaism, whichfollows from the next sentence in Malala’s text(…kšcxm Ósi oƒ `Rxla‹oi oƒ vqirsiamoˆ jajîV poioàriso‹V 'Iotda…oiV ™m so‹V lšqerim aÙsîm…), points todh�u-Nuw�as.

The background: Kamqa~€isai, Kimaidojokp…sai andthe KindaIt is interesting to note that the threat to the caravantrade, which prompted the Abyssinian incursion,was already present in the middle of the first centuryAD. According to the Periplus of the Erythraean Sea,the people of Kamqa~€isai, who obviously occupiedthe coastal regions of the southern H@ ij�az, were one ofthe main dangers facing sea traders in the centralRed Sea. The ‘kings and tyrants’ (shaykhs?) ofArabia tried to subdue these robbers:

DiÕ jaˆ rtmevîV ¢pÕ sîm stq�ammxm jaˆ barikšxmsÁV 'Aqab…aV a„vlakxs…fomsai· kšcomsai d�e Kam-qa~€isai

[For this reason they are constantly being takenprisoner by the governor and kings of Arabia;Kanraitai is their name (§20: 7.9–11) — trans.Casson 1989: 63].

The identification of the Kamqa~€isai has long beendisputed. Emendations, such as Karram…sai (fromGhass�an), Kamam…sai (from Kin�ana) or even'Aqqab…sai, have been proposed in order to refutea connection with the inhabitants of MinaeanQRNW (e.g. Glaser 1890: 165–166; a link betweenthe Kamqa~€isai and QRNW was suggested by Pirenne1961: 77).

An emendation appears unnecessary, however, ifthe correspondences between Greek and localSemitic consonants are considered. The most pro-blematic in the name Kamqa~€isai is the initial j. Thiscan correspond to q, g, gh and more rarely to h@ and kin Arabian ethnonyms and toponyms rendered inancient Greek: thus K�awima pÒkiV (Ptolemy, Geogr.6.7.14) < H@ abs��n (U.A.E.) (53�37’19’’E; 23�49’29’’N);KoqolamˆV pÒkiV (Ptolemy, Geogr. 6.7.19) < aram.h@ r�um�an�a1 (‘forbidden’) > 'IeqÕV jÒkpoV (Ptolemy,Geogr. 6.7.19) = Gulf of Kuwait; KoqÒdalom ¥jqom(Ptolemy, Geogr. 6.7.11) = R�as al-H@ add (22�31’59’’N;59�47’09’’E) with dissimilation dd > rd (qd). Thename of R�as al-H@ add seems to be connected with thename of Cadaei (Pliny, Nat. Hist. 6.149).

Only the shift q > j can be considered commonacross the entire Arabian Peninsula. Ptolemy’sGeography gives only a few Greek ethnic andgeographic names with j in H@ ijaz. The shift j < ghappears, for example, in Majoq�aba (Ptolemy, Geogr.6.7.32) which has traditionally been understood as aderivation from Arabic mikr�ab (‘sanctuary’) andconnected with Mecca (most recently Cuvigny &Robin 1996: 716). There is, however, anothertoponym — portus Mochorbae (Pliny, Nat. Hist.6.149) — that represents the Latin translation ofPtolemy’s `IeqÕV jÒkpoV (Ptolemy, Geogr. 6.7.19) andKoqolamˆV pÒkiV (Ptolemy, Geogr. 6.7.19), i.e. the

1 On the derivatives of Eastern Aramaic h@ rm in the sense ‘todedicate for sacred use’, ‘to ban’; see Sokoloff 2002: 483–484.

TOWARDS THE EARLIEST HISTORY OF KINDA

67

Greek rendering of Aramaic h@ r�um�an�a (‘forbidden’).As implied by the Latin name, the Greek formshould have been spelled with v instead of j(*Mavoq�aba) if it was derived from mikr�ab.

Since the Greek j in the west Arabian toponymson Ptolemy’s map can correspond to Arabic g (e.g.KÒpaq < J�ar [Ptolemy, Geogr. 6.7.5]; KšmsoV < Judda[Ptolemy, Geogr. 6.7.5])2 or gh (e.g. Karramisîm vèqa[Ptolemy, Geogr. 6.7.6], Karamdqe‹V [Agatharchides,De mari. 96], Casani [Pliny, Nat. Hist. 6.150],Karram‹sai [Stephen of Byzantium, Ethnica 372.14);< Old South Arabian _GSN), another interpretation ofMajoq�aba seems probable: Majoq�aba < Arabicmaghrib. A derivation from the root _GRB (‘to lie inthe west’) is more credible than one from QRB (‘to beclose to’, ‘to offer’ as qarraba) or GRB (cf. jirbiyya’,‘northwest wind’, Lane 1863: 403). Two otherethnonyms — K�aqbai (Agatharchides, De Mari. 97)and Cerbani (Pliny, Nat. Hist. 6.154) — can also beconnected with Majoq�aba.

Another example of the j < gh correspondence isfound in the name Ghass�an > Karramisîm vèqa(Ptolemy, Geogr. 6.7.6) = Karamdqe‹V (Agatharchides,De mari. 96), Casani (Pliny, Nat. Hist. 6.150);Karram‹sai (Stephen of Byzantium, Ethnica 372.14).Note also Gar�amdoi (Diodorus Siculus, Hist. 3.45.6).

The shift Sem. m > Greek m (Ghamr > Kamqa~€isai) isseen on Ptolemy’s map of Arabia where 'Ialb…ajèlg (13; 6. 7. 3) < Yanb�u’ al-Bah@ r and vice versa —Kaql�am bar…keiom (226; 6. 7. 34) < Old South ArabianQRNW.

It follows, therefore, that Kamqa~€isai is to bederived from a local name with initial gh, mostprobably from Ghamr (dh��-Kinda), a place c. 90 kmto the north-north-west of modern Mecca and thatthe Kamqa~€isai themselves can be connected withKinda, thus constituting the first written reference tothem. It is interesting to note that both peoples — theKamqa~€isai and the Kimaidojokp…sai — appear inconnection with the ‘Arabs’, ‘kings and tyrants ofArabia’ in the Periplus (§20: 7.10) and with 'Aqab…saiin RIE 277 ⁄ 27. The reference in the Periplus to thestruggle of the ‘kings and tyrants of Arabia’ againstKamqa~€isai is thus to be seen in the context of

attempts by local rulers to secure a safe passage forcaravans through the southern H@ ijaz and for shipsalong the western coast of the Red Sea.

Kimaidojokp…sai (Ptolemy, Geog. 6.7.5)Until now this name has defied satisfactory explana-tion. It had already been connected with Kinda andunderstood as ‘Kinda, living on the shore of the gulf(jÒkpoV)’ (Conti Rossini 1921: 23, fn. 2). However, ifthis were the case then the form *Kokpisojima‹daiwould be expected. Moreover, the Kimaidojokp…saioccupied the entire coast of the H@ ij�az, not that of aparticular gulf.

Later, von Wissmann interpreted Kimaidojokp…saias the union ‘Kin�ana wa Kalb’ (1957: 297, n. 27), i.e.‘perverts from the Gulf’ or adherents of the Kinda.As Bowersock noted, ‘This bizarre appellation seemsto denominate dwellers along the gulf who wereeither perverts (if the Greek is to be understoodliterally) or adherents of Kinda, now known to becentred in Qaryat al-F�aw. It is now time to askwhether the Kindite Empire embraced parts ofnorth-west Arabia after the decline of the Naba-taeans’ (1996: 563).

More recently the Kimaidojokp…sai have beenidentified with the tribal group Daw’at withoutoffering any explanation of the name (Cuvigny &Robin 1996: 714–717). There is, however, a goodphonetic and geographic correspondence withDaw’at in ancient Greek ethnography — namelythe Dšbai mentioned by Agatharchides as inhabi-tants of the H@ ij�azi Tih�ama (De mari. 95). Theethnonym Dšbai is not to be confused with QÁbaipÒkiV on Ptolemy’s map of Arabia Felix (24; 6. 7. 5),which is to be identified with modern Dhahb�an(21�56’48’’N; 39�04’55’’E).

The identification of Kimaidojokp…sai with anethnonym written with initial k has been rejected,since ancient Arabic q was pronounced *k andconsequently Greek j rendered q, not k. Semitic kcorresponds to Greek v (Cuvigny & Robin 1996: 716–717). Examples from Latin cannot serve as supportfor a correspondence between Sem. k and Greek j:Capeus (Pliny, Nat. Hist. 6.147) < Kuwayt; Caminacum(Pliny, Nat. Hist. 6.160) < Old South ArabianKMNHW; cancamum (Plin. NH. 1.12; 12.98–99) <ASA kmkm. Until the beginning of the first centuryBC the Greek aspirates were usually rendered in

2 For the identifications see below. Ptolemy’s map is citedaccording to Humbach and Ziegler (1998; 2002).

M.D. BUKHARIN

68

Latin as unvoiced: Lat. *k (c) corresponded to Greek* k (v), as well as to *k (j), and Latin c could alsocorrespond to *h. As Seelmann noted, ‘Griechischesk und x haben die Romer in ubernommenenLehn- und Fremdwortern...regelrecht durch c ev. chwiedergegeben’ (1885: 333); cf. Lat. Eutyces for GreekEÙstv»V, or Lat. Capeus (Pliny, Nat. Hist. 6.147) forGreek Cabot�asa (Ptolemy, Geog. 6.7.32), both ofwhich are derived from the name of Kuwait, or Lat.Icarum (Pliny, Nat. Hist. 6.147) for 'Iv�aqa (Ptol. 25;6.7.47) < Arabic Faylaka: Lat. c = Greek v < Sem. k.

An ostracon from mons Maximianon mentioningCimedajokpisîm (not Cimedojokpisîm; see the photo-graph in Cuvigny & Robin 1996: 699, fig. 1) dating to118 AD makes the earliest reference to theCimedajokpisai, about half a century before Ptol-emy’s Geography. The reading Cimedajokpisîm with

initial v nullifies the argument of Robin and Cuvignyagainst identifying the first part of Kimaidojokp…saiwith Kinda and removes any connection betweenthat name and purely Greek notions that could liebehind this ethnonym, such as j…maidoV (‘pervert’) orj…madoV (‘fox’, ‘sly’). The reading proposed here alsomakes it clear that the name could only be derivedfrom a Semitic name with initial k, followed by n andd: *KND. This spelling corresponds precisely to thename Kinda. The appearance of i after the secondconsonant is, possibly, due to the building of thename Kimaidojokp…sai after the jimaidokÒcoV-pattern,reflecting an interpraetatio Graeca of a non-Greekname.

Moreover, there are other Greek names onPtolemy’s map of Arabia that illustrate the corre-spondence of Greek j with Semitic k. These include:

Fig. 1.

Aksum and Kinda in west Arabia

TOWARDS THE EARLIEST HISTORY OF KINDA

69

'Arj‹sai (Geog. 6.7.26) = the inhabitants of R�asH@ as��k in ‘Om�an; cf. however the rendering'Ar…vom in the Periplus (§33: 11.9);

'Obq�aja [Geog. 6.7.28; earlier identified with al-‘Assaf��ya ⁄ Abraq ar-Ruh@ayma (von Wissmann1963)] = Mabrak an-N�aqa (26�55’N; 37�53’E) c. 15km to the north of Mad�a’in S @�alih@ ;

Diorjoq…dotV mÁroV jaˆ pÒkiV (Geog. 6.7.45) = OldSouth Arabian *D-SKRD = Dioscuridu (Pliny, Nat.Hist. 6.153; Periplus §30: 10.3) = Suqut @r�a. The nameDiorjoq…dotV consists of two parts: Dio +rjoq…dotV, the first of which renders the posses-sive pronoun dh�u. According to W.W. Muller, theGreek and Latin names go back to di-Sak�urid(2001: 151–152; cf. Biedermann 2006: 11–12). Ear-lier interpretations in the sense of an interpraetatioGraeca of Sanskrit dv��pa sukh�adh�ara — ‘the blessedland’ (e.g. Tubach 1993: 102; for critics seeBiedermann 2006: 11) — are unsatisfactory, sincethis Sanskrit name is nowhere attested. Othernames on Ptolemy’s map of Arabia and in Pliny’sNatural History include an additional vowel torender interdental pronunciation of the precedingconsonant, e.g. Qiak»lah (Qi�akgkka) (Ptolemy,Geog. 6.7.10) < Old South Arabian TLMYT <Greek Psokela~€iV (S@al�ala), Ti�acaq (Ptolemy, Geog.6.7.33) = Fujayra (with change t > f), Suelleni(Pliny, Nat. Hist. 6.157) < Tu’al. The name St�acqoV< R�as S�ajir (4; 6. 7. 1) also seems to show thispattern.

'Iv�aqa (Geog. 6.7.47) = Icarum (Pliny, Nat. Hist.6.147) = 'IjaqoV (Arrian, Anab. 7.20.5; Strabo, Geog.14.1.6, 16.3.2; Dionysius Periegetes 608ff) < Arab.Faylaka.

Another name can be added to this list without,however, any certitude about its identification:KqtpsÕV kil»m (Ptolemy, Geog. 6.7.12) earlier identi-fied with Masqat@ (von Wissmann 1977: 9; Salles1995: 122) < Kalba (25�04’53’’N; 56�20’41’’E) in thelagoon Khawr Kalba with the shift l > r. As theexample of MÒrva kil»m (Ptolemy, Geog. 6.7.10)shows (MÒrva kil»m < mus�ahila), the modifierjqtpsÕV must not necessarily be a translation(‘hidden’) but can represent the Greek rendering ofa local name or notion, e.g. q�arib — ‘a man

journeying to water... one doing so by night’ (Lane1863: 2508). Relevant information in the Periplusappears in the report on the inhabitants of theislands of Kaka…ot:

Peqijokp…fomsi d�e s¾m ™volšmgm ½peiqom e„V aÙs¾ms¾m ¥qjsom ½dg peqˆ s¾m e„rbok¾m sÁV PeqrijÁVhak�arrgV je‹msai mÁroi pke…omeV, aƒ Kaka…otkecÒlemai (mÁroi), rvedÕm ™pˆ rsad…otV dirvik…otVpaqasesallšmai sÁ vèqvÆ pomgqoˆ d�e oƒ

jasoijoàmseV aÙs�aV ¥mhqxpoi jaˆ ¹lšqaV oÙ pokÚ

si bkšpomseV

[After coasting due north along the next stretchof the shore, in the vicinity by now of theentrance to the Persian Gulf, you meet numer-ous islands, called the Isles of Kalaios, strungout along the coast for almost 2000 stades. Themen who populate them are rascals who donot come in sight much during the daytime(§34.20–24) — trans. Casson 1989: 71 withcorrections].

These islands are identified with the islands ofDaymaniyat to the north-west of Musqat (Potts1990: 315, n. 222). The name KqtpsÕV kil»m can beconnected with the night-time robbing activity of theinhabitants of these islands, as mentioned in thePeriplus.

Thus, there are names (e.g. Kimaidojokp…sai,'Arj‹sai and ”IjaqoV) with a double rendering ofthe initial Sem. k as Greek j and v. The mostprobable interpretation of Kimaidojokp…sai is that itsfirst part is derived from the name of Kinda,rendering this name in Greek in a pejorative form,which coincides with the Arabic diminutive of Kinda> *Kunayda. The interpretation of the second part ismore problematic. It might be derived from Arabicghalaba (‘to win’). If so, it would be yet anotherexample of the correspondence of Semitic gh withthe Greek j and the entire name Kimaidojokp…saicould be then understood as ‘the victorious Kinda’.Alternatively, it could be compared with tribalnames such as Ban�u Kil�ab or Ban�u Kalb. In lateantiquity the Ban�u Kalb lived near Tab�uk. The Ban�uKil�ab b. Rab��’a b. ‘ �Amir was one of the branches of‘ �Am��r b. S @a’s @a’a, one of the most powerful tribalgroups in the central part of northern Arabia. Theirterritory extended to the east, north and north-west

M.D. BUKHARIN

70

of Mecca. Since the territory of Ban�u Kil�ab is close tothat of Kinda, the identification of -jokp…sai withBan�u Kil�ab appears more credible.

According to the Arab historical tradition asanalysed by G. Olinder (1927: 34), a certain princeof the Kinda escaped from H@ ad@ ramawt and settled inGhamr dh��-Kinda, two days’ journey (90 km to thenorth-north-east) from Mecca. Olinder datedthe appearance of the Kinda in the region of Meccato the fourth century AD (1927: 36). According toal-Hamd�an��, the appearance of the Kinda here had tohave preceded their resettlement in the H@ ad@ ramawt,a notion that fits well with the idea of a Kindapresence, under the name of Kamqa~€isai, in the H@ ij�azalready by the middle of the first century AD:

...wa f�ı H@ ad@ ramawta sakanat Kinda ba’da ‘an ‘ajilat‘ani l-bah@ rayn wa-l-Musaqqar wa-Ghamr dh�ı-Kindaf�ı-l-j�ahiliyya ba’da qutila ‘Ibn al-Jawn

[The Kinda have settled down in H@ ad@ ramawt afterthey have been driven out from Bah@ rain, Mush-aqqir and Ghamr dh��-Kinda in the time of J�ahil��yyaafter Ibn al-Jawn was killed (S@ifa. 88. 1–2)].

In this respect quite interesting information issupplied by the following fragment of Pliny’sNatural History (5.65). Pliny informs us about theCanchlei, who live between the Cedrei and theNabataei:

Vltra Pelusiacum Arabia est, ad Rubrum marepertinens et odoriferam illam ac divitem terram etBeatae cognomine inclutam. haec Catabanum etEsbonitarum et Scenitarum Arabum vocatur, sterilis,praeterquam ubi Syriae confinia attingit, nec nisi Casiomonte nobilis. his Arabes iunguntur, ab orienteCanchlei, a meridie Cedrei, qui deinde ambo Nabataeis.

[Beyond the Pelusiac mouth of the Nile is Arabia,extending to the Red Sea and to the Arabia knownby the surname of Happy and famous for itsperfumes and its wealth. This bears the names ofthe Cattabanes, Esbonitae3 and Scenitae tribes ofthe Arabs. The soil is barren except where itadjoins the frontier of Syria, and its only remark-

able feature is El Kas mountain. The Arabian tribeof the Canchlei adjoin those mentioned on the eastand that of the Cedrei on the south, and both ofthese in their turn adjoin the Nabataei — trans.Rackham 1989: 269].

It would be quite tempting to regard the formCanchlei as being derived from KHL — the secondpart of the Sabaean name of Qaryat dh�at-K�ahilum,modern Qaryat al-F�aw and capital of the Kinda. Theinsertion of an additional nasal is not rare in Greekand Latin renderings of Semitic names (e.g. Kaqlam[Ptolemy, Geogr. 6.7.34]) < Old South ArabianQRNW). If Kinda, whose future centre was Qaryatdh�at-K�ahilum, can really be seen in the Canchlei,then we might interpret this passage of Pliny’s workto mean that, at the time this information wasprovided to Pliny, the Kinda occupied territories innorth-eastern Arabia on the frontiers of Syria andthe Nabataean lands. Hence this would supportHamd�an��’s statement that the Kinda inhabitedBah@ rain and Mushaqqir (i.e. territories in north-eastern Arabia).

Aksum and the Arabs in the first–secondcentury ADThe Arabs seem to have played a dominant role inthe sea trade from Africa to the northern H@ ij�az viaMuza. As the Periplus reports, the tyrant of Ma’�afir(prince, vassal of H@ imyar) leased sea trade withRhapta, the southernmost market known in Africa,to the traders of Muza who mostly used Arab pilotsand crews, since the Arabs had close contacts andmixed families with the Rhaptans (§16: 6.9–13). TheArabs also had close relations with Leuke Kome, thesouthernmost Nabataean port (§19: 6.30). Muza wasfull of Arabs — experienced sailors and shipowners— who played an active part in trade with the ‘otherside’, i.e. east Africa (Aksum), and even with India(§21: 7.21–23). On the other hand, the author of thePeriplus knew Suqut@r�a was an ‘earlier’ meeting placeof merchants from Muza and Indian traders fromsouthern and western India:

TÕ l�em Ókom 'Aq�abxm matjkgqijîm ¢mhqèpxm jaˆ

matsijîm pkeom�afei jaˆ so‹V ¢p’ ™lpoq…aV pq�aclarijime‹saiÆ rtcvqîmsai c�aq sÁ sot pšqam ™qcar…v jaˆ

Baqtc�afxm „d…oiV ™naqsirlo‹V.

3 The name Esbonitarum is derived from the nisba plural ofOld South Arabian *SB’YN.

TOWARDS THE EARLIEST HISTORY OF KINDA

71

[The whole place teems with Arabs — shipownersor charterers and sailors — and is astir withcommercial activity. For they share in the tradeacross the water and with Barygaza, using theirown outfits (§21: 7. 21) — trans. Casson 1989: 63].

It should be remembered that, according to Pliny theElder, Muza was not a port of trade with India (Nat.Hist. 6.104). Hence, the merchants of Muza neededan intermediary place where they could meet Indiantraders and this was Suqut @r�a, as implied by thePeriplus:

Stmevq»ramso d�e aÙsÁ jaˆ ¢pÕ MoÚfa sim�eV jaˆ sîm™jpkeÒmsxm ¢pÕ LiltqijÁV jaˆ Baqtc�afxm Óroijas�a sÚvgm e„V aÙs¾m ™pib�akkomseV.

[Trade with it used to be carried on by some of theshippers from Muza and also by those sailing outof Limirike and Barygaza, who by chance put in atit (§31: 10. 20–25) — trans. Casson 1989: 69].

It follows that the Arabs controlled a large segmentof the sea trade in the Red and Arabian Seas, i.e. thecoastal trade between East Africa, Aksum, SouthArabia and Nabataea, with its centre at Muza (al-Mukh�a). The ‘kings and tyrants of Arabia’ werenaturally very interested in maintaining this trade,which gave them so much profit. In the time of thePeriplus (mid-first century AD) they tried to main-tain it by themselves, sending their armies againstKamqa~€isai (possibly Kinda). At the end of the secondcentury, the situation was slightly different. AsPtolemy’s map and RIE 277 show Kimaidojokp…sai(probably Kinda) were pushed from the mainland(Ghamr dh��-Kinda) to the coast and the ‘Arabs’ wereseemingly not very happy with Aksumite interfer-ence in western Arabian affairs4. The Aksumitesmust have been the closest allies of the Romans andthe party most interested in the maintenance ofpeace on the Incense Road. The alliance of the‘Arabs’ with Aksum, noticeable in the first centuryAD, had collapsed and Aksum now fought againstboth the Kinda on the shores of the Red Sea and the‘Arabs’ on the mainland. The Arabian campaign of

the Aksumites only continued the earlier actions ofArabians princes, allied to the Romans, who stroveto restore peace on the western Arabian branch ofthe Incense Road in favour of Rome.

The Abyssinian navyAccording to RIE 277, the Abyssinian army had tocross the Red Sea twice: ‘from the borders of Egypt’,i.e. presumably from Berenice to Leuke Kome, and‘from the Sabaean borders’ to Adulis. The mostprobable starting point of the second crossing seemsto have been J��z�an (Old South Arabian G’ZN[‘Abdall�ah 2004: 2–5])5, a port that lies practicallyopposite Adulis. A Latin inscription found on theFaras�an archipelago, lying between both ports,confirms that in the mid-second century AD thispart of the Red Sea was under Roman control(Phillips, Villeneuve & Facey 2004: 239–250; Ville-neuve, Phillips & Facey 2004: 143–192; for adiscussion see Bukharin 2005–2006: 135–140). How-ever, practically nothing is known about the Aksu-mite navy. It seems that both crossings of the RedSea — ‘from the borders of Egypt’ to Leuke Komeand on the way back to Adulis — were achievedusing Roman ships and even the foundation of theRoman port in the Faras�an islands c. 140 AD mayhave been directly connected with the Aksumitecampaign through East Africa and western Arabia.Moreover, the later crossing of the Red Sea by theAbyssinian army sent against dh�u-Nuw�as (T@ abari,Ta’r��kh I2 929, 16), with the help of Byzantine ships,confirms that the Abyssinians themselves had nofleet capable of crossing the Red Sea and that forsuch crossings the Abyssinians had to rely on theassistance of a foreign navy.

The crossing of the Red Sea ‘from the borders ofEgypt’ to Leuke Kome was not without precedent.The Periplus seems to point to the use of thisitinerary by the Aksumites already before themiddle of the first century AD:

'Ej d�e sîm eÙxmÚlxm Beqm…jgV ¢pÕ MtÕV Óqlotdtrˆm dqÒloiV ½ sqirˆm e„V s¾m ¢masok¾mdiapkeÚramsi sÕm paqaje…lemom jÒkpom ÓqloV ™rsˆm›seqoV jaˆ uqoÚqiom, Ókšcesai Letj¾ jèlg

4 The subjugation of both Kinaidokolpites and Arrabites bythe Aksumite army does not necessarily mean that thesetwo groups were allies.

5 I owe this reference to Professor W.W. Muller to whom Iwish to express my sincere thanks.

M.D. BUKHARIN

72

[To the left of Berenice, after a voyage of two orthree runs eastward of Myos Hormos past thegulf lying alongside, there is another harbor andfort, called Leuke Kome (§19: 6. 26–28) — trans.Casson 1989: 61].

The port of Berenice at the southern border of Egyptwas the starting point of this itinerary. It seems thatneither Greeks — and later Romans of Egypt — northe Arabs used this itinerary. The author of thePeriplus does not give a list of goods traded in LeukeKome, as he did for other ports where the Romanswere active, while the Arabs followed the east coastof the Red Sea.

It remains to suggest that the use of the sea routefrom Berenice to Leuke Kome by the Aksumites wascommon at least two centuries before the Arabiancampaign of the Aksumite negus who authored RIE277. If this is correct, then it follows that theAksumite army defeated the Kinda in the southernH@ ij�az and that the appearance of the Kinda aroundmodern Qaryat al-F�aw needs to be seen as a result ofthis defeat. The presence of a military representativeof the Abyssinian negus ‘qb ⁄ ngsyn at Najr�an men-tioned in Ja 577 ⁄ 10–13 (c. 250 AD) may also havebeen a result of the Aksumites’ Arabian campaignfor the control of another branch of the IncenseRoad. The h@ bst mentioned in a graffito at JabalKawk�ab (Ryckmans 1957: 561) may also reflect theAksumite presence on the way from Najr�an toQaryat al-F�aw, either during or shortly after themarch recorded in RIE 277.

The route of the Abyssinians in western ArabiaThe route of the Abyssinians through East Africaand western Arabia is of interest. The king whoauthored RIE 277 says that his army reached theborders of Egypt, having gone by the regular road:

Øpšsana … œhmg … s�a lšvqi sîm sÁV A„cÚpsotÐq…xm o„joàmsa, Øpos�anaV pefeÚerhai ™po…gra s¾mÐdÕm ¢pÕ sîm sÁV ™lÁV barike…aV sÒpxm lšvqiA„cÚpsot (2–14).

He says that he had conquered the Kinaidokolpites.One may assume that at the very least their capitalmust have been taken. Ptolemy’s map of Arabiaallows us to reconstruct the route of the Abyssinians

through western Arabia from Leuke Kome to thefrontiers of Saba’.

Letj¾ jèlg (Ptolemy, Geog. 6.7.3)The location of Leuke Kome has exercised scholarsfor over a century. Aside from RIE 277, only thePeriplus and Strabo mention this Greek toponym:

pokk�a d�e pahëm jaˆ sakaipxqgheˆVpemsejaidejasa‹oV Âjem e„V Letj¾m jèlgm sÁVNabasa…xm cÁV, ™lpÒqiom lšca

[After many experiences and hardships he arrivedin fourteen days at Leuke Kome in the land of theNabataeans, a large emporium (Strabo, Geog.16.4.23) – trans. Lloyd Jones 1995: 357].

At first sight, the absence of this name on Ptolemy’smap of Arabia appears enigmatic. Leuke Kome hasbeen identified with various places on the Arabiancoast of the Red Sea, principally al-’Ayn�una (e.g.Kirwan 1399 AH ⁄ 1979: 55–61; Bowersock 1983: 48–49; Casson 1989: 188), ‘Umm Lajj (e.g. Warmington1928: 11; Pirenne 1961: map 1), Yanb�u’ al-bah@ r (vonWissmann 1957: 317–322; 1976: 466, Abb. 1; Busch-mann 1991: 88–89), R�as Kark�uma (von Maltzan 1865:92 gave ‘Ras Rharkuma’), al-Buw�anat (von Wiss-mann 1963), somewhere to the south of al-Wajh(Moritz 1925: 2262), at the mouth of W�ad�� al-H@ amd)but not at al-Wajh (Caskel 1952: 20) or near al-Wajh(Gatier & Salles 1988: 186–187).

There are several counter-arguments against anidentification with al-’Ayn�una or Yanb�u’. Al-’Ayn�una was known to Ptolemy under its ownname: A‡mot (Ptolemy, Geog. 6.7.43) and ”Ommg(Ptolemy, Geog. 6.7.2) while Yanb�u’ al-bah@ r can berecognised in 'Ialb…a jèlg (Ptolemy, Geog. 6.7.3) =insula Iambe (Pliny, Nat. Hist. 6.168). Letj¾ jèlg alsofigures on Ptolemy’s map under the name `Ratm�ahotjèlg (Ptolemy, Geog. 6.7.3), between al-’Ayn�una andYanb�u’ al-bah@ r. Several Greek and Latin renderingsof Semitic toponyms in Arabia demonstrate the shiftof l > q, e.g. BamoÚbaqoi (17) < B. Ball�� (?); KqtpsÕVkil»m (85) < Khawr Kalba (?); ”A 'ppg (191) < H@ �a’il >Arreni (Pliny, Nat. Hist. 6.1576; 'Iv�aqa (295) < Arab.

6 Already suggested in Bowersock 1996: 561.

TOWARDS THE EARLIEST HISTORY OF KINDA

73

Faylaka; Arab. sal��kha > Lat. serichatum (Pliny, Nat.Hist. 12.99). Similarly, q in the 'Ratm�ahot seems tosubstitute Sem. l7. As such, 'Ratm�ahot may bederived from Arab. laban�at, translated then intoGreek as Letj¾ jèlg (‘white village’).

The coordinates of Letj¾ jèlg on Ptolemy’s map(67�15’E; 25�40’N), as compared with those of'Ialb…a jèlg (68�00’E; 24�00’N), suggest that Letj¾

jèlg is to be sought c. 185 km to the north ofmodern Yanb�u’, i.e. near al-Wajh (26�14’05’’N;36�27’55’’E). This is supported by the relevantpassage in the Periplus quoted above (`Ej d�e sîmeÙxmÚlxm Beqm…jgV ¢pÕ MtÕV Óqlot dtrˆm dqÒloiV½ sqirˆm e„V s¾m ¢masok¾m diapkeÚramsi sÕmpaqaje…lemom jÒkpom ÓqloV ™rsˆm ›seqoV jaˆ

uqoÚqiom, Ó kšcesai Letj¾ jèlg).The Aksumite sailing voyage started at Berenice

and ran ‘to the left’ (`Ej d�e sîm eÙxmÚlxm Beqm…jgV),i.e. to the north of Berenice. The next stage was MtÕVÓqloV, i.e. Old Qusayr (Redde & Glovin 1987: 61, 63;Peacock 1993: 226–239) where the fleet turned to theeast (e„V s¾m ¢masok¾m) and crossed the adjacent gulf(diapkeÚramsi sÕm paqaje…lemom jÒkpom). This gulfwas tangential to the aim of the voyage, and neitherthe Gulf of Suez nor the Gulf of ‘Aqaba, but onlySharm al-Wajh, can be meant. The entire voyagefrom Berenice to Sharm al-Wajh lasted two or threedays (dtrˆm dqÒloiV ½ sqirˆm): one day for thecrossing from Berenice to Myos Hormos, anotherfor the journey from Myos Hormos to Leuke Kome.An additional day seems to have been reserved forany delay in crossing. Also in favour of anidentification of Letj¾ jèlg with al-Wajh is the factthat the Minaean ‘Lists of Hierodules’ mention WG’(Ma’��n 95 ⁄ 3 = M 394; Ma’��n 96 ⁄ 9 = M 395), identifiedby von Wissmann with al-Wajh (1970: 975, Abb. 7). Ifthe identification of Letj¾ jèlg with al-Wajh iscorrect, then this was an important trading centre,familiar to Minaean traders and obviously orientedtowards Egypt.

veqrÒmmgroV ¥jqa (Ptolemy, Geog. 6.7.3)This has previously been identified with ‘EilandUm-er-Ruma’ (von Maltzan 1865: 92) and R�as ‘Ab�uMadd (von Wissmann 1963) = R�as Khurayya (c. 30km to the south of al-Wajh@ . The coordinates of both

'Ratm�ahot jèlg (67�15’E; 25�40’N) and CeqrÒmmgroV¥jqa (67�00’E; 25�40’N) on Ptolemy’s map implythat the places were near each other.

'Ialb…a jèlg (Ptolemy, Geog. 6.7.3)Identifiable as Yanb�u’ al-Bah@ r (24�06’10’’N;38�03’22’’E). It is interesting to note that initial y isnot rendered by k, unlike some other Arabian namesrendered in Greek or Latin.

Qaltd‹sai (Ptolemy, Geog. 6.7.4)This is the tribal confederation Tham�ud.

Sidgmo… (Ptolemy, Geog. 6.7.4)Cf. B. S’ad b. Zayd M�alik.

D�a 'ppai (Ptolemy, Geog. 6.7.4)This is identifiable with B. ad-D�ar b. H�ani’ b. H@ ab��bab. Num�ara b. Lakhm (?)8. The identification ofD�a 'ppai with al-’Azd b. al-Ghawth, also known asDarr�a, must be rejected, since this tribe lived inal-Yam�ama (Wustenfeld 1853: 99). Pliny the Eldermentions the ethnonym gens Darae (Nat. Hist. 6.150)following his description of the insulae Alaea (Dahlakislands), suggesting that the gens Darae are to besought in South Arabia. These islands are identicalwith DailÒmxm (Ptolemy, Geogr. 6.7.43) = mÁroi'Akaka…ot kecÒlemai (Periplus §4.2: 13–15) = insulaequae Aliaeu vocantur (Pliny, Nat. Hist. 6.173) = 'Eka…a(Strabo, Geog. 16.4.8). All of these names are derivedfrom ‘l (‘god’), plur. ‘l’lt. According to the Periplus,the `Akaka…ot-island lay to the south of the bay ofZula (i.e. Adulis) and must be identified with theislands Huwak��l and Baka. These are in turnidentical to the islands Bachhias and Antibacchias(Pliny, Nat. Hist. 6.173) ⁄ B�ajvot jaˆ 'Amsib�ajvot(Ptolemy, Geogr. 4.8.38). The Latin names of theseislands recall the name of the god Bacchus, and canbe semantically compared with the Greek DailÒmxm.Arab geographers (e.g. ad-Dimashq��, K. nukhbat. 162)mention the island al-Qut@rubiya, whose name isderived from Arab. qut@rub (‘evil spirit’, ‘demon’).Ad-Dimashq�� says only that this island lies to thesouth of the Burnt Island (Kasajejatlšmg; 283; 6.7.44on Ptolemy’s map), Jabal at @-T@ �a’ir.

7 The r < l shift is quite common in ancient Arabic, in par-ticular in the Qays dialect (see Kofler 1940: 77–78).

8 For the place of B. ad-D�ar in traditional Arabic genealogysee Caskel 1966: 246 ⁄ 16.

M.D. BUKHARIN

74

In the time that elapsed between the compositionof these two works (i.e. Pliny’s Natural History andPtolemy’s Geography), the B. ad-D�ar (D�a 'ppai) mayhave moved from the south of the Arabian Peninsulato the northern Tih�ama.

BamoÚbaqoi (Ptolemy, Geogr. 6.7.4)Possibly *Ban�u Barr�a ⁄ Kin�ana (?). Barr�a bt. Murr ‘wasthe first wife of Khuzayma b. Mudrika and motherof’As’ad b. Khuzayma, whose son Kin�ana marriedher...’ (Wustenfeld 1853: 107). The location ofBamoÚbaqoi on Ptolemy’s map corresponds well tothe territory of Kin�ana, one of the branches of the B.Kalb. On the other hand, the name BamoÚbaqoi alsorecalls the name of the region of B. Ball�� b. ‘Amr (B.Bal�� b. ‘Amr, ‘the ancient tribe Qud@�a’a…to the Northof Juhaina and to the West of ‘Udhra’ (Caskel 1966, i:233). This identification requires, however, a shift ofq < l (see above). Does BamoÚbaqoi = Barrasasei(Pliny, Nat. Hist. 6.155)?

”Aqrai (Ptolemy, Geogr. 6.7.4)This is B. H@ �arith. Several tribes may be reflected inthis name including al-H@ �arith b. Buhtha in theTih�ama (Wustenfeld 1853: 208; cf. Caskel 1966, ii:122 ⁄ 12) or al-H@ �arith b. Ka’b b. ‘Abd b. M�alik b. Nas@rb. al-’Azd, known to have lived in the last centuriesbefore Islam in the region of Najr�an (Wustenfeld1853: 210; cf. Caskel 1966, ii: 210 ⁄ 15). On the otherhand, ”Aqrai could be a translation from Arab. ‘ar��s(‘tent’), which could, as in the case of Sjgm…sai,denote a nomadic way of life. ”Aqrai = Arsi (Pliny,Nat. Hist. 6.155).

Kimaidojokp…sai (Ptolemy, Geogr. 6.7.5)See the discussion above.

KÒpaq jèlg (Ptolemy, Geogr. 6.7.5)This has been identified previously with al-J�ar (vonWissmann 1963). In the H@ ij�azi dialect g waspronounced as q (Kofler 1940: 115; Rabin 1951:125–126) and was consequently reflected in Greek j,while p (Latin p) often reflects Semitic w, e.g.NapgcoàV (Ptolemy, Geogr. 6.7.7) < Mawsij; QapaÚa(Ptolemy, Geogr. 6.7.27) < Tawba; ”Arpa (Ptolemy,Geogr. 6.7.34) < Nazwa; Capeus (Pliny, Nat. Hist.6.147) < Kuwayt; Param(alacum) < Old SouthArabian W’RM (Pliny, Nat. Hist. 6.157); or Pallon(Pliny, Nat. Hist. 6.159) < Old South Arabian W’LN.

Thus, the following chain can be suggested: J�ar <*Gawar > *Qawar > KÒpaq. This kind of shortening(KÒpaq – J�ar) may be illustrated by another examplefrom Ptolemy’s map of Arabia, thus ZÚcaima(Ptolemy, Geogr. 6.7.43) < Saw�akin > bed@awyeu-Sok (Hofheinz 1997: 87).

”Aqca jèlg (Ptolemy, Geogr. 6.7.5)This has previously been identified with R�asMast�ura (von Maltzan 1865: 169) and Jidda (vonWissmann 1963) or as a derivation from Arab.‘irq ⁄ ’urq (‘dune’) (von Wissmann 1968a: 1306) =Rayyis (23�34’45’’N; 38�36’27’’E) with the change y ⁄ g(?). It is probably the equivalent of ”Ecqa jèlg(Strabo, Geog. 16.4.24), the starting point for thecrossing of the Red Sea by Aelius Gallus on his wayfrom Arabia to Egypt.

Zabq�al Bar…keiom (Ptolemy, Geogr. 6.7.5)This has been identified with Rab��gh (von Maltzan1865: 198) and Mars�a Ibr�ah��m (Ge’ez za-Abreham)near al-L��th (von Wissmann 1963; 1968a: 1304, 1307).However, Mars�a Ibr�ah��m lies in modern Eritrea(18�52’81’’N; 37�24’83’’E) = Mast�ura (23�06’27’’N;38�50’00’’E). The reading Zadq�alg in the Ethnika ofStephen of Byzantium (Zadq�alg, bar…keiom sîmKimaidojokpisîm [372. 14]) points to the possibilityof the identification of Mast�ura with the capital ofthe Kinaidokolpites. The change s > z is traceable inthe western Arabian dialects, e.g. in that of Kalb(Rabin 1951: 195). The reading Zadq�alg is also givenin Herodianus, De prosodia catholica (3.1.325), whichconfirms the correctness of the reading Zadq�alg,retained in the Middle Ages.

KšmsoV jèlg (Ptolemy, Geogr. 6.7.5)This has been identified with Mersa Deneb (vonMaltzan 1865: 198) or al-Qunf��dha (von Wissmann1963). The absence of f in Qunf��dha, as well as thefact that interdental dh was never rendered in Greekas s, argue against the identification of KšmsoV withal-Qunf�udha (19�08’27’’N; 41�04’19’’E). Moreover,al-Qunf�udha lies 3� to the south of Dhahb�an andsuch an identification would have broken the logicof Ptolemy’s description. It may be suggestedinstead that KšmsoV jèlg = Judda, according to localdialectal pronunciation Jidda (21�32’47’’N;39�09’50’’E). The above-mentioned H@ ij�azi dialectalpronunciation of g as q, reflected in Greek as j, as

TOWARDS THE EARLIEST HISTORY OF KINDA

75

well as the dissimilation dd > nd > ms need to betaken into consideration: KšmsoV < *qnd < �GDD >Judda (for dissimilation cf. also gandal < *gndl < *gddl< �GDL.).

QÁbai pÒkiV (Ptolemy, Geogr. 6.7.5)This has been identified with Dahlimar (von Mal-tzan 1865: 198, 205), W�ad�� Tayya (Moritz 1923: 108),S @abya (Pirenne 1955: 159, n. 3), H@ �al�� or Dhahb�an (vonWissmann 1957: 299–300), Dhahb�an (von Wissmann1963) or Mahd adh-Dhahab (23�29’43’’N,40�52’24’’E) (von Wissmann 1964a: 157; 1968a:1310). The identification with Mahd adh-Dhahab isimpossible since this lies on the mainland. The mostprobable is Dhahb�an (21�56’48’’N; 39�04’55’’E).

Bais…ot posaloà ™jboka… (Ptolemy, Geogr. 6.7.5)This has been identified with Meeresarm Obhor (pl.from bah@ r) near R�as Dahlimar (von Maltzan 1865:206–210) or the mouth of the W�ad�� Bays ⁄ Bayd@(17�30’N) (Sprenger 1875: 49; von Wissmann 1963;1968a: 1305). The W�ad�� Bays flows into the Red Seatoo far to the south to be for Bais…ot posaloà

™jboka…. A more likely alternative is the mouth ofW�ad�� B�at @in.

aƒ pgcaˆ soà posaloà (Ptolemy, Geogr. 6.7.5)This is the source of W�ad�� B�at @in. W�ad�� B�at @in hasbeen called a ‘major feature of Arabia and can betraced right across the peninsula. It appears beyondthe narrow belt of Dahana sands as the W�ad��Rumma (or Rima), which curves westwards aboutlatitude 26� as it passes through Nejd and leadseasily into the Hejaz by the W�ad�� Hamdh and itstributaries thus linking up the central Arabiantowns… both with the Red Sea and the head of thePersian Gulf under three different names: B�at @in,Rumma and H@ amd@ ’ (Mason, Sherwin-White &Hyamson 1987: 116–117).

ConclusionRIE 277 was composed by Gad�urat, king ofAbyssinia, early in the reign of ‘Alhan Nahf�an, kingof Saba’. In light of other Old South Arabian texts(CIH 308 and Ja 631) and a Latin inscription fromthe Faras�an island, as well as in the context of thePeriplus of the Erythraean Sea, RIE 277 reflects both thealliance between Rome and Aksum and between

Aksum and Saba’. The motivation behind thisalliance was the protection of Roman trade interestsin the Red Sea and in western Arabia with the helpof Aksum.

The principal danger for Roman sea traders in theRed Sea and for caravans, travelling towards theMediterranean through western Arabia, came fromthe tribal union of Kinda, known in the first centuryunder the name of Kamqa~€isai (Periplus) and in thesecond century as Kimaidojokp…sai (as part of aunion between Kinda and B. Kil�ab). The Kindablocked the sea for the cabotage sea trade and theland branches of the Incense Road and wereconsequently driven away by the Aksumites, forcingthem to resettle in the southern part of centralArabia around Qaryat al-F�aw (ancient Qaryat),outside the sphere of interest of Rome and Aksum.The subsequent Kinda blockade of the centralArabian branch of the Incense Road became thejustification for the campaign of S�a’ir ‘Awtar, king ofSaba’, against the Kinda at Qaryat.

It would be interesting to connect the campaign ofS�a’ir ‘Awtar against TML (Ja 635 ⁄ 36–37) with theprocess of Kinda migration through western Arabia.TML must be identical to Pliny’s Thomala (Nat. Hist.6.154), an important centre for the trade in aromatics(‘et, quo merces odorum deferunt, Thomala’).Thomala is normally identified with modern B��sa(20�00’12’’N, 42�44’49’’E), an oasis lying on the wayfrom Saba’ to the H@ ij�az (see most recently de Maigret1997: 318; 1999: 222). S�a’ir ‘Awtar fought there withbnw ⁄ yw[n]m ⁄ wqrytm in the land of TML. YWNM isknown from Ma’��n 93(A) ⁄ 45 = M 392 as Ionia orGreece in general. ‘The sons of QRYTM’, apparentlythe Kinda, still retained control of one of the mostimportant trade centres in western Arabia, albeit tothe south of their previous territory, namely TML.

Ja 635 seems to describe a situation that resultedfrom the movement of the Kinda from the H@ ij�az tosouth-central Arabia. The Kinda are said to havefought against Saba’ in modern B��sa with their alliesthe ‘Greeks’, who, if they really represent tradersfrom Roman Egypt, must have previously beeninterested in defeating the bnw qrytm. The situationreflected in Ja 635, is this: the ‘Greeks’ are said to beallies of Kinda (bnw qrytm) on the ‘Incense Road’against Saba’.

It seems that ywn (‘Ionians’) does not necessarilymean traders sensu stricto in Ja 635. We read in the

M.D. BUKHARIN

76

biography of the Roman Emperor Macrinus: pug-navit tamen et contra Parthos et contra Armenios etcontra Arabas, quos Eudaemones vocant, non minusfortiter quam feliciter (Script. Hist. Aug. Macrin. 12. 6).Thus, Macrinus waged war against South Arabiansand since his reign (217–218) coincides chronologi-cally with the presumed dates of S�a’ir ‘Awtar, wemay suppose that the Sabaeans had to fight in TMLagainst Kinda (bnw ⁄ qrytm) and Roman forces(bnw ⁄ ywnm).

Previously (e.g. Pirenne 1961: 193, 197–198; vonWissmann 1964b: 478–484) the Arabian campaign ofMarinus was connected with the presumed destruc-tion of Arabia Felix by the unnamed Caesar referredto in the Periplus:

Nàm d�e oÙ pqÕ pokkoà sîm ¹lesšqxm vqÒmxmKa‹raq aÙs¾m jasersqšwaso

[… not long before our time Caesar sacked it (§26:8. 31–32) — trans. Casson 1989: 65].

Romans could be called ‘Ionians’, i.e. Greeks, sinceGreek remained the spoken language of the RomanEast and Egypt. However, the reasons behind thechange of Roman preferences remain unclear. Itseems that some twenty years after the Axumitecampaign in western Arabia and forced resettlementto the south, the Kinda might have establishedcontacts with Rome and convinced a Roman army tolaunch a campaign against Saba’. This remains,however, a mere supposition but, in any case, thegroundwork for the later interference of the Kinda inRome’s policy in the East must have already beenlaid in the early third century AD.

Appendix. Qaryat al-F�aw in the classicaltraditionThe Roman world came to know inner Arabia withthe campaign of Aelius Gallus in 25 ⁄ 24 BC. Theitinerary of the Roman army, especially the wayback to Egypt from M�arib, has never been analysedfrom a geographical point of view. Only a few factsare securely known. The Roman army turned backat M�arib (Strabo, Geog. 16.4.24) and the furthestpoint of Roman penetration in Arabia was Caripeta(Pliny, Nat. Hist. 6.160). The itinerary of theRomans to the shores of the Red Sea led through

Nšcqama, `Eps�a uqšasa, C�aakka jèlg, MakÒha and”Ecqa jèlg.

Nšcqama is surely identifiable with Najr�an. There-fore, `Eps�a uqšasa, which, according to Strabo, laytwo days’ march from Najr�an (c. 80 km), can beidentified with the region between B��r Khadr�a (c. 30km from Najr�an) and B��r H@ im�a (70 km from Najr�an).C�aakka jèlg can be reconstructed as follows: thefirst consonant of the local name must have been k,which is regularly rendered as j in Greek. Thecombination of aa points to the falling of anintervocalic consonant, as in Ti�alaa (RIE 277 ⁄ 6) <THMT. The combination of the consonants *KHLsuggests that the Greek name of C�aakka must bederived from the Old South Arabian (QRYT DHT)KHLM. Though KHLM occurs only in Sabaeaninscriptions and never in the texts from Qaryat al-F�aw itself, the fact that the Roman army marched toC�aakka from M�arib makes a derivation of this Greekname from Old South Arabian KHLM probable.

In light of the possible presence of Romans underthe command of Aelius Gallus at Qaryat al-F�aw, theCaripeta in Pliny’s Natural History requires reconsi-deration. Various places have been identified withthis name, e.g. ancient H@ RB in W�ad�� Raghw�an (vonWissmann & Hofner 1953: 31; for previous identifi-cations see Pirenne 1961: 107–108), but this cannot beconsidered the furthest point of Roman penetrationin Arabia, and neither does h@ correspond to Greek jin South Arabian toponyms. Caripeta has also beenidentified with H@ ar��b Khawl�an on the way fromM�arib and Raghw�an to the highland of Suma’y (vonWissmann 1964a: 213–214), but the Yemeni high-lands cannot have been the furthest point of Romanpenetration. The most popular identification ofCaripeta has been with ancient HRBT in W�ad�� H@ ar��b,modern Hajar H@ in�u az-Zurayr (von Wissmann1968b: 1282), which lies to the south-east of M�arib,although Strabo says that Aelius Gallus turned northat M�arib. For these reasons an identification ofCaripeta with ancient H@ RB seems unlikely. Asalready noted above, Latin p in Pliny’s work as wellas Greek p on Ptolemy’s map of Arabia quite oftenrender the semivocal w. Therefore, the initial form ofLatin Caripeta may be reconstructed as *QRWT. Thew could change with y (Beeston 1984: 10; Stein 2003:29–34) as illustrated by the existence of the formsTaveni (Pliny, Nat. Hist. 6.157) and Tagmo… (Uranius,Fragment 11) for T@ ayy��’. Thus Caripeta, the furthest

TOWARDS THE EARLIEST HISTORY OF KINDA

77

point reached by the Romans in Arabia, can bederived from Old South Arabian QRYT and directlyconnected with the Greek name C�aakka, both ofwhich render the ancient name of Qaryat dhat-Kahilum.

The toponym MakÒha may possibly be connectedwith M�aqiha on Ptolemy’s map of Arabia (Ptolemy,Geog. 6.7.20). If so, this points to the march of theAelius Gallus’ army through T@ uwayq al-’ �Ar��d@ ontheir way to the coast of the Red Sea.

Two other names on Ptolemy’s map may alsorender QRYT: Ga…a pÒkiV (Ptolemy, Geog. 6.7.29) andGiq�aha ⁄ Goiq�aha (Ptolemy, Geog. 6.7.32). Ga…a pÒkiV isgiven by Ammianus Marcellinus as Geapolis (23.6.47)in the list of the six most important cities of Arabia:Geapolim et Nascon et Baraba itidemque Nagara etMaepham et Tapphara et Dioscurida [Geapolis, Nasq,M�arib, Najr�an, MYF’ (Sayb�an) and Suqut @ra]. Therendering of q as g in the Bedouin dialects of Arabiais a well-known feature (Thilo 1958: 21; Holes 1996:37–39). The rendering of a local name with initial q >c in the case of Caripeta may represent the domina-tion of the Minaeans at Qaryat al-F�aw in the late firstcentury BC. As archaeological excavations at Qaryatal-F�aw showed, the lower strata of the site (third–first century BC) attest to a Minaean presence (al-Ansary 1997: 371–373), while the ‘Bedouinisation’ ofthe site had already begun at the time of Ptolemy’sGeography. The Old South Arabian form of thetoponym must have changed in the following way:

Qarya > *Qayya > *Gayya > Ga…a. In the same way,Giq�aha ⁄ Goiq�aha, earlier identified with Qaryatayn tothe north-west of ar-Riy�ad) (von Wissmann 1963),renders local QRYT but without the assimilation of rto y.

The data of archaeology are supplemented by theMinaean inscription, published on a poster of KingSaud University (ar-Riy�ad@ ) in 1407 AH: …sqnyt ⁄dhbn ⁄ … ⁄ bhdyht ⁄ w’krb ⁄ ‘h d ⁄ ‘h@ lysm ⁄ ns’h ⁄ knhr ⁄‘d ⁄ slky ⁄ wgb’h ⁄ ‘d ⁄ qryt ⁄ wqrnw…9. This is adedication made by a Minaean for his safe returnfrom Seleukeia-on-the-Tigris to Qaryat al-F�aw andQarn�aw and can presumably be dated to the firstcentury BC.

AcknowledgementsI would like to express my most profound gratitude to ProfessorW.W. Muller (Marburg) for having kindly read the manuscript of

this article and for important remarks and corrections. Respon-

sibility for the conclusions remains my own. The main ideas here

were presented and discussed in a seminar at the Institut fur

Sprachen und Kulturen des Vorderen Orients (Friedrich Schiller

University, Jena; Lehrstuhl fur semitische Philologie) in 2007. For

this opportunity I am grateful to Professor N. Nebes and the

other participants in the seminar, especially Dr P. Stein, A.

Multhoff, Dr M. al-Sal�am�� and F. al-’ �Al��.

References‘Abdall�ah, Y.M. 2004. Naqs gad��d min

al-’as@r as-saba’�� al-qad��m. Al-Musnad 2:

2–5.

al-Ansary, A.R.T. 1997. Qaryat al-Fau.

Pages 371–373 in Meyer, E.M. (ed.), TheOxford Encyclopedia of Archaeology in theNear East. Vol. 4 (New York & Oxford,Oxford University Press).

Beeston, A.F.L. 1984. Sabaic Grammar.

Manchester: JSS Monograph 6.

Bernand, E., Drewes, A.J. & Schneider, R.,

2000. Recueil des inscriptions de l’Ethiopiedes periodes pre-axoumite et axoumite,

Vol. 1. Paris: Boccard.

Biedermann, Z. 2006. Soqotra: Geschichteeiner christlichen Insel im IndischenOzean vom Altertum bis zur fruhenNeuzeit. Wiesbaden.

Bowersock, G.W. 1975. The Greek-Naba-

taean Biligual Inscription at Ruwwafa,

Saudi Arabia. Pages 513–522 in Bingen,

J. et al. (eds), Le monde grec: Hommages aClaire Preaux (Bruxelles, Universite

libre de Bruxelles, Faculte de Philoso-

phie et Lettres 62).Bowersock, G.W. 1983. Roman Arabia.

Cambridge, MA: Harvard University

Press.

Bowersock, G.W. 1996. Exploration in

North-West Arabia after Jaussen-Sa-

vignac. TOPOI 6 ⁄ 2: 553–563.

Bukharin, M.D. 2005–2006. Romans in the

Southern Red Sea. Arabia 3: 135–140.

Buschmann, K. 1991. Motiv und Ziel des

Aelius-Gallus-Zuges nach Sudarabien.

WO 22: 85–94.

Caskel, W. 1952. Die Beduinenstamme inNord- und Mittelarabien und im ‘Ir�ak.

Wiesbaden: Harrassowitz.

Caskel, W. (ed.). 1966. Gamharat an-Nasab.Das genealogische Werk des His�am ibnMuh@ ammad al-Kalb��, 2 volumes. Leiden:

Brill.Casson, L. 1989. The Periplus Maris Ery-

thraei. Text with Introduction, Translationand Commentary. Princeton: Princeton

University Press.

Conti Rossini, C. 1921. Expeditions et

possessions des H@ abas�at en Arabie. JA18: 5–36.

Cuvigny, H. & Robin, C. 1996. Des

Kinaidokolpites dans un ostracon grec

du desert oriental (Egypte). TOPOI6 ⁄ 2: 697–720.

9 Reading by Professor W.W. Muller, Marburg.

M.D. BUKHARIN

78

de Maigret, A. 1997. The frankincense road

from Najr�an to Ma’�an: A hypotheticalitinerary. Pages 315–332 in Avanzini,

A. (ed.), Profumi d’Arabia. Atti delConvegno (Rome, Saggi di Storia Antica

11).

de Maigret, A. 1999. The Arab nomadic

people and the cultural interface be-

tween the ‘Fertile Crescent’ and ‘Ara-

bia Felix’. AAE 10: 220–224.

Desanges, J. 1984. Rome et les riveraines

de la Mer Rouge au IIIe siecle de notre

ere. Apercus recents et nouveaux pro-blemes. Ktema 9: 249–260.

Fiaccadori, G. 2004. Sembrouthes ‘Gran re’

(DAE IV 3 = RIE 275) per la storia del

primo ellenismo aksumita. La Paroladell’Passato 335: 103–157.

Gatier, P.-L. & Salles, J.-F. 1988. Aux

frontieres meridionales du domaine

nabateen. Pages 173–190 in Salles, J.-F.

(ed.), L’Arabie et ses mers bordieres(Lyon, TMO 16).

Glaser, E. 1890. Skizze der Geschichte undGeographie Arabiens von den altestenZeiten bis zum Propheten Muh@ ammad,nebst einem Anhange zur Beleuchtung derGeschichte Abessyniens im 3. und 4. Jahrh.n. Chr. Auf Grund der Inschriften, derAngaben der alten Autoren und der Bibel,vol. 2. Berlin: Weidmannsche Buch-

handlung.

Graf, D.F. 1978. The Saracens and the

defense of the Arabian frontier.

BASOR 229: 555–569.Graf, D.F. 1988. Qura ‘Arabiyya and

Provincia Arabia. Pages 173–211 in

Gatier, P.-L., Hellz, B. & Rey-Coquais,

J.-P. (eds), Geographie historique auProche Orient, Actes de la Table Ronde deValbonne, 16–18 Septembre 1985 (Paris,

Notes et Monographies techniques 23).

Hofheinz, A. 1997. Saw�akin. Encyclopaediaof Islam 9: 87–89.

Holes, C. 1996. Dialects of South Eastern

Arabia in a socio-historical perspec-tive. Zeitschrift fur arabische Linguistik31: 34–57.

Humbach, H. & Ziegler, S. (eds). 1998.

Ptolemy. Geography, Book 6. Middle East,Central and North Asia, China, Part 1.

Wiesbaden: Ludwig Reichert.

Humbach, H. & Ziegler, S. (eds). 2002.

Ptolemy. Geography, Book 6. Middle East,Central and North Asia, China. Part 2.

Wiesbaden: Ludwig Reichert.

Isaac, B. 1989. Trade Routes to Arabia andthe Roman Presence in the Desert.

Pages 241–256 in Fahd, T. (ed.), L’Ara-bie preislamique et son environnementhistorique et culturel. Universite desSciences Humaines de Strasbourg, Actesdu Colloque de Strasbourg 24–27 Juin1987 (Leiden, Travaux du Centre de

recherche sur le Proche-Orient et la

Grece 10).

Kirwan, L.P. 1399 AH ⁄ 1979. Where to

Search for the Ancient Port of Leuke

Kome. Pages 55–61 in Abdulla, A.M.

et al. (eds), Studies in the History ofArabia II: Pre-islamic Arabia (Riyadh,University of Riyadh).

Kofler, H. 1940. Reste altarabischer Dia-

lekte. WZKM 47: 61–130; 233–262.

Korotayev, A. 1994. The political role of

the sha’b of the first order. Rayd�an 6:

47–52.

Lane, E.W. 1863. An Arabic-English Lexicon.

London: Williams and Norgate.

Lloyd Jones, H. (transl.). 1995. Strabo,Geography, Books 15–16. Cambridge &

London: Loeb Classical Library.Maltzan, H. von. 1865. Meine Wallfahrt

nach Mekka. Reise in der Kustengegendund im Inneren von Hedschas, vol. 1.

Leipzig: Dyk.

Mason, K., Sherwin-White, A.N. &

Hyamson, A.M. 1987. The Middle EastIntelligence Handbooks 1943–1946, Vol. 1,Iraq and the Persian Gulf. Slough:

Archive Editions.

Moritz, B. 1923. Arabien: Studien zur physi-kalischen und historischen Geographie desLandes. Hannover: Orient-Buchhan-

dlung Heinz Lafaire.

Moritz, B. 1925. Letj¾ jèlg. RE 12 ⁄ 2: 2262.

Muller, W.W. 2001. Antike und mittelal-

terliche Quellen als Zeugnisse uber

Soqotra, eine einstmals christliche In-

sel. Oriens Christianus 85: 139–161.

Olinder, G. 1927. The Kings of Kinda of theFamily of �Akil al-Mur�ar. Leipzig: Lunds

Universitets Arsskrift, N.F. Avd. 1. Bd

23. Nr. 6.Peacock, D.P.S. 1993. The site of Myos

Hormos: A view from space. Journal ofRoman Archaeology 6: 226–239.

Phillips, C., Villeneuve, F. & Facey, W.

2004. A Latin inscription from South

Arabia. PSAS 34: 239–250.

Pirenne, J. 1955. La Grece et Saba: Unenouvelle base pour la chronologie sud-arabe. Paris: Memoires presentes par

divers savants a l’Academie des In-

scriptions et Belles-Lettres 15.

Pirenne, J. 1961. Le Royaume sud-arabe deQataban et sa datation d’apres l’archeolo-gie et les sources classiques jusqu’auPeriple de la Mer Erythree. Louvain:

Bibliotheque du Museon 48.

Potts, D.T. 1990. The Arabian Gulf inAntiquity. Vol. 2. From Alexander theGreat to the Coming of Islam. Oxford:

Clarendon Press.

Potts, D.T. 1994. Augustus, Aelius Gallus

and the Periplus: A re-interpretation of

the coinage of S@ana’�a’ Class B. Pages

212–222 in Nebes, N. (ed.), Arabia Felix.Beitrage zur Sprache und Kultur desvorislamischen Arabien. Festschrift WalterW. Muller zum 60. Geburtstag (Wiesba-

den, Harrassowitz).

Rabin, C. 1951. Ancient West-Arabian. Lon-

don: Taylor’s Foreign Press.

Rackham, H. (transl.). 1989. Pliny, NaturalHistory in Ten Volumes. II. Libri III–VIII.Cambridge & London: Loeb Classical

Library.

Redde, M. & Glovin, J.-C. 1987. Du Nil a laMer Rouge: documents anciens et

nouveaux sur les routes du desert

oriental d’Egypte. Karthago 21: 5–64.

Retso, J. 2003. The Arabs in Antiquity. TheirHistory from the Assyrians to theUmayyads. London & New York: Rou-

tledge Curzon.

Robin, C. 1982. Les Hautes-terres du Nord-Yemen avant l’Islam. I. Recherches sur lageographie tribale et religieuse de Hawl�anQud�a’a et du pays de Hamd�an. II.Nouvelles inscriptions. Leiden: Neder-

lands Historisch-Archaeologisch Insti-

tuut te Istanbul.

Robin, C. 1989. La premiere intervention

abyssine en Arabie meridionale. Pages

147–162 in Beyene, T. (ed.), Proceedingsof the Eighth International Conference ofEthiopian Studies, vol. 2 (Addis Ababa,

Institute of Ethiopian Studies).

Ryckmans, G. 1957. Graffites sabeens

releves en Arabie sa’udite. RSO 32:557–563.

Salles, J.-F. 1995. The Periplus of the

Erythraean Sea and the Arab-Persian

Gulf. Pages 115–146 in Boussac, M.-F.

& Salles, J.-F. (eds), Athens, Aden,Arikamedu. Essays on the InterrelationsBetween India, Arabia and the Mediterra-nean (New Delhi, Manohar).

Seelmann, E. 1885. Die Aussprache desLatein nach philologisch-historischenGrundsatzen. Heilbronn: Henninger.

TOWARDS THE EARLIEST HISTORY OF KINDA

79

Shitomi, Y. 1994. Une nouvelle interpreta-

tion du Monumentum Adulitanum.Journal of East-West Maritime Relations3: 73–114.

Shitomi, Y. 1997. A New Interpretation of

the Monumentum Adulitanum. Memoirsof the Research Department of the ToyoBunko 55: 81–102.

Sokoloff, M. 2002. A Dictionary of JewishBabylonian Aramaic. Baltimore: Johns

Hopkins University Press.

Sprenger, A. 1875. Die alte GeographieArabiens als Grundlage der Entwick-lungsgeschichte des Semitismus. Bern:

Huber.

Stein, P. 2003. Untersuchungen zur Phono-logie und Morphologie des Sabaischen.

Rahden ⁄ Westf.: Epigraphische For-

schungen auf der Arabischen Halbin-

sel 3.

Thilo, U. 1958. Die Ortsnamen in deraltarabischen Poesie: Ein Beitrag zur vor-und fruhislamischen Dichtung und zurhistorischen Topographie Nordarabiens.Wiesbaden: Schriften der Max Freiherr

von Oppenheim-Stiftung 3.

Tubach, J. 1993. Die Anfange des Chris-

tentums im Sudarabien. Eine chris-

tliche Legende syrischer Herkunft in

Ibn His�am. Parole de l’Orient 18: 101–

111.Uhlig, S. 2001. Eine trilinguale ‘Ezana

Inschrift’. Aethiopica 4: 7–31.

Villeneuve, F., Phillips, C. & Facey, W.

2004. Une inscription latine de l’archi-

pel Faras�an (sud de la mer Rouge) et

son contexte archeologique et histo-

rique. Arabia 2: 143–192.

Warmington, E.H. 1928. The Commercebetween the Roman Empire and India.

Cambridge: Cambridge University

Press.Wissmann, H. von. 1957. ‘De Mari Ery-

thraeo’. Pages 289–325 in Wilhelmy, H.

(ed.), Hermann Lautensach-Festschrift(Stuttgart, Stuttgarter Geographische

Studien 69).

Wissmann, H. von. 1963. Faltkarte in

Grohmann, A. Arabien. Munich: Kul-

turgeschichte des alten Orients

III.1.1.3.4.

Wissmann, H. von. 1964a. Zur Geschichteund Landeskunde von Alt-Sudarabien.Vienna: Sitzungsberichte der Osterrei-

chische Akademie der Wissenschaften,

phil.-hist. Kl. 246 [= Sammlung Eduard

Glaser 3].

Wissmann, H. von. 1964b. Himyar, An-

cient History. Le Museon 77: 429–499.Wissmann, H. von. 1968a. Zaabram. RE 11:

1304–1312.

Wissmann, H. von. 1968b. Uranios. RE 11:

1286–1292.

Wissmann, H. von. 1970. Ophir und

Hawila, das westarabische Goldland,

Dedan und Hegra. RE 12: 909–979.

Wissmann, H. von. 1976. Die Geschichte

des Sabaerreiches und der Feldzug des

Aelius Gallus. ANRW 9 ⁄ 1: 308–544.

Wissmann, H. von. 1977. Das Weihrauch-land Sa’kal�an, Sam�arum und Mos-cha.

Vienna: Sitzungsberichte der Osterrei-

chische Akademie der Wissenschaften,

phil.-hist. Kl. 324 [= Sammlung Eduard

Glaser 14].

Wissmann, H. von & Hofner, M. 1953.

Beitrage zur historischen Geographie desvorislamischen Sudarabien. Mainz: Aka-

demie der Wissenschaften und der

Literatur in Mainz (88.4).

Wustenfeld, F. 1853. Register zu den genea-logischen Tabellen der arabischen Stammeund Familien. Mit historischen und geo-graphischen Bemerkungen. Gottingen:

Dieterichschen Buchhandlung.

M.D. BUKHARIN

80