Hill, (Sir) George Francis The ancient coinage of southern Arabia THE BRITISH ACADEMY e Ancient Coinage of Southern Arabia G. F. Hill om the Proceedings ofthe British Academy, Vol. VII} London Published for the British Academy ~ jL By Humphrey Milford, Oxford University Press .Q, Amen Corner, E.G. * Price Two Shillings net 4- Mb 0/61 6 NVI THE ANCIENT COINAGE OF SOUTHERN ARABIA BY G. F. HILL Read May 5, 1915 THE ancient coinage ofSouthern Arabia is one ofthe mostobscure branches ofnumismatics. In origin it is Greek; but in development it is Semitic. For the proper study of it a numismatist who is equally well equipped on the Greek and Semitic sides is required; and such a scholar has yet to be discovered. What is more, the study of South Arabian epigraphy is at present in a somewhat inchoate stage; vast quantities of inscriptions have been discovered, but only partly published; an extraordinarily bitter personal feud, with wide ramifications, has done anything but quicken progress; and of the scholars who have devoted themselves to this branch of archaeology, only one, Mordtmann, has paid serious attention to the coins. In fact, the divorce between numismatics and archaeology is as painfully evident here as in any other place. That must be the excuse for me, as a numismatist with hardly the most rudimentary Semitic equipment, if I venture to deal with the subject; my wish is merely to put the numismatic material together in a form in which Semitic scholars may be able to deal with it effectively. According to Strabo,1 whose information is based on Eratosthenes, there were four leading tribes in occupation of Southern Arabia, or rather of that portion which may be described as lying over against Aethiopia. First there are the Minaeans, in the part near the Red Sea, with their chief town Kama or Karnana; adjoining them, the Sabaeans, with their metropolis Mariaba; third, the Katabanians, who extend to the Straits and the crossing of the Arabian Gulf; their royal city is called Tamna; and farthest to the East, the Chatramotites, whose city is Sabata. The absence ofthe Himyarites from this list is due to the fact that they did not rise to power until after the time ofEratosthenes. 1 xvi. 768. vii c 2 PROCEEDINGS OF THE BRITISH ACADEMY Hitherto the coins of Southern Arabia have always been classed together as 'Himyarite'. It will be seen that the greater part of them must be divided between Sabaeans and Himyarites, and also that there is ground for distinguishing two small groups ofcoins, one attributable to the Minaeans, the other to the Katabanians (people of Kataban), although this latter group can only be regarded as subordinate to the main Himyarite series. Three out of the four tribes mentioned by Strabo are thus provided with a coinage. Since the rise of the Himyarites to power probably did not take place before the middle of the second century B.C., when their capital at Sapphar regia (Sqfar, near Yerim) superseded the old Sabaean capital at Mariaba (Marib\ the earliest series of the coins with which we are concerned should strictly be regarded as Sabaean rather than Himyarite. Nevertheless, the chronology is so uncertain, and the series are interlaced in so curious a way, that it is very difficult to draw any line between them. Roughly speaking, we may assume that the earliest coins, which are direct imitations of the earlier Attic coinage, belong to the Sabaean period, while the later, flat coins (of that which we may for convenience call the San'a class),1 with a reverse type derived from the Attic coinage of the 'New Style', the small coins with names and heads of various kings, and the 'bucranium' series must certainly belong to the Himyarite period. But there are certain single coins, or small groups, which, although in fabric and types they look fairly early, seem by their monograms and inscriptions to be intimately connected with the San'a class, apparently so much later. The following is an attempt at a provisional classification of the various series. I. IMITATIONS OF THE OLDER ATTIC TYPES a. Obv. Head of Athena. Rev. Owl, with olive-spray, crescent, ,and AGE, more or less blundered; traces of incuse square on some specimens (Plate I. 2). The largest coins which appear to belong to this class are repro- ductions of the Athenian tetradrachm; the only specimens known to me are at Berlin.2 One is countermarked on the obverse with X 1 Since the great majority, if not all, of the known specimens in silver seem to have come from the great hoard discovered there and described by Schlumberger,Le TrfsordeSaria(Paris, 1880). 2 D. H.MiillerandJ.W. Kubitschek, SudarabischeAltertumer(Vienna),1889, p.76, 1, nos.474and 183-7. I citethis work henceforward as 'M.u. K.'. All thesecoinswere broughtfrom South ArabiabyGlaseror Mordtmaim. ANCIENT COINAGE OF SOUTHERN ARABIA 3 (Sabaean n) and A ; another has something likeaSabaean monogram scratched on the reverse. But these large coins are quite exceptional, and we are justified in regarding as the ordinary unit the smaller coins of 5-55 grm. maximum.1 These units all show the Sabaean a on the cheek of Athena; the halves, when legible, are similarly marked with 3 ; the quarters with O;2 the eighths with &. The same system appears to be followed on the series next to be described. On the San'a coins, however, we find-the halves marked sometimes with a, sometimes with 3, while a cross (which may be meant for D, though that is not certain) occurs on the reverse ofsome ofthe San'a units which have 3 on the obverse. The P (b) which is found on one half(Schlumberger, PL III. 56) mayperhaps be really f (a). Schlumbergerhas suggested that the a is the initial of Nejran (Neypava). But if he is right, it would seem to follow that the other letters mentioned above are also mint-initials,3 and that, at least in the earliest period, the four different denominations were issued from four different mints. If this seems improbable, it is, for the following reasons, equally difficult to accept the view, which suggests itself upon the con- sideration of the earlier series, that the letters are the initials of denominations. Schlumberger records (p. 22) an early Attic tetra- drachm which has been countermarked with a Sabaean a,4 and, We as already stated, the a is found on halves of the San'a class. have also seen that the Berlin Museum possesses5 a piece of about the weight of the Attic tetradrachm (16-95 grm.), imitated from the earlier Attic types, with two countermarks, viz. X and A, of which the former may be the Himyarite n ; and this letter, as we have seen, is found on the quarters. The coins of the class with which we are dealing bear nothing Sabaean or Himyarite about them save their style and the letter on the cheek ofAthena (the *tetradrachms' at Berlin being without even the latter distinguishing mark of Arab origin). They still retain traces of the incuse square, and were dated by Head6 about 1 The standard isdiscussedbelow(pp. 23-4). 2 ExceptonepublishedbyMordtmann,Num. Zeit.,1880, p.293,Taf.V, no. ii,. which appears tohave O (y). 3 For a, the city of 123, associated in an inscription with Nejran, has been suggested (see C.I. S., iv. 7). * It must be remembered that the Sabaean a is hardly distinguishable from, thesame letter in some other Semitic scripts. 6 M.u. K.,p. 76, no. 474. ' Num. Chron., 1880, p. 310. 4 PROCEEDINGS OF THE BRITISH ACADEMY 400 B.C. It is, however, clear from the treatment of the eye that they are imitated from the comparatively late Attic coins (Plate1. 1) whichmaythemselvesbedatedtothefourthcentury(circa393-322B.C. according to Head).1 The earliest imitations themselves are scarcely earlier than the third century.2 /8. Similar to series a, but slightly broader in fabric and later in style; on the reverse, Sabaean letter or monogram; traces of incuse square rarely ifever present (PI. I. 3). The units, halves, and quarters are marked with the same letters as in series a; no eighths seem to be known. These coins must cover a fairly long period of time ; for in proportion to the number of specimens known (the collections in London and Vienna provide all or nearly all of them), the number of varieties is comparatively large, at least nineteen different letters or monograms being represented; while to strike the thirty-eight specimens ofthe unit in the British Museum alone about twenty-nine obverse and twenty-seven reverse dies were required. The series may be dated to the second century B.C. The lower limit is furnished by the fact that some coins which resemble this series in fabric and style are intimately connected by monograms and inscriptions with the flat coins ofthe San'a class. Some eighteen or nineteen letters or monograms occur on coins of this series; they range from single Sabaean letters to elaborate monograms. The most remarkable is a group of two signs, one exactly resembling the Sabaean n, the other being forked above, with a wavy tail (PL I. 4, 5). The latter has been the subject of considerable discussion. The points to be remembered are: first, that the two signs are frequently found together in a position of importance in lapidary inscriptions; second, that the forked sign is always much larger than the other, which is almost always of the same proportions as the ordinary letters, and may therefore be taken as the letter n. An examination of all the evidence makes it fairly clear that the forked sign has some symbolic significance. It is possibly a degenerate pictograph derived from the bucranium and associated with 'Athtar;3 but still more probable appears to me the derivation from the Babylonian twin-dragon sceptre. The earliest example4 of the twin-serpent-sceptre motif is found on 1 Hist. Num.', p. 374. a I donotspeak ofthe 'tetradrachms', having seen none of those at Berlin, butofthe unitsandsmallerdenominations. 8 Thissuggestion isnotnew; see C.I. , iv, no. 366, p. 12. * I owewhatfollows tomy colleague, Professor L. W. King. ANCIENT COINAGE OF SOUTHERN ARABIA 5 a libation vase in the Louvre of dark green steatite dedicated by Gudea, patesi of Lagash, to Ningishzida, his patron deity, about 2450 B.C.1 Ningishzida in his chief aspect was a war-god and a Sumerian prototype of the god Ninib in his later character, whose emblem was the twin lion-headed sceptre; so that the twining serpents with natural heads are the directancestorsofthelion-headed serpents ofthe later emblem, as we getit, forinstance, on aboundary- stone of Nazimaruttash,2 about 1330 B.C. I take it that the wavy form generally assumed by the tail of the Sabaean sign in question is a relic ofthe spirals ofthe serpents' tails. The other sign, as I have said, is usually if not always represented of the same size as the ordinary letters of the inscription, and is doubtless only n. Weber and the editors of the C./.*S., however, regard it as a special symbol, the former elaboratinga mostingenious theory, which identifies it with the double curved symbol which occurs so frequently upon the later coins (see below). If he were right in this last identification, then (1) the voided and solid forms of the curved symbol must be distinguished, because (2) the solid form, at least, of the curved symbol occurs occasionally in connexion with and addition to the group of signs which we are discussing. But that the solid and voided forms of the curved symbol cannot be distinguished in significance is clear from the fact that both are used indifferently in the same context on coins ofthe San'a class. y. In a small group ofcoins, comprising two specimens at Vienna,3 one at Paris, and one in the British Museum, we find on the obverse, instead ofthe head ofAthena, a beardless male head with curly hair, in which the Viennese scholars see a resemblance to the portrait of Philetairos on Pergamene coins. There is no letter on the cheek. The owl is more erect than on the series a and )8, and A0E is absent. The monograms are more elaborate, and there are two on each coin. In spite of the very different drawings of the monograms on the two coins given in the Viennese publication, they appear from the photographs to be meant for the same; that on the right is also the same as appears on the Paris and London coins, which are incomplete on the left. The Paris coin shows a a below the right- hand monogram, whichthusconsists ofn+y+ +l|+S. The left-hand monogram,judgingfrom the photographs, consistsofD"1 with 3 above; 1 L. Heuzey, Catal. des Antiquites chaldeennes, 1902, p. 280; the same, De- couvertesen ChaldeeparE. deSarzec, vol. ii, 1912, PI. 44. 2 HilprechtAnniversary Volume, p. 274, Fig. 7. 3 M.u. K.,p. 68, II, nos. 1,2. 6 PROCEEDINGS OF THE BRITISH ACADEMY but Miiller and Kubitschek draw it as a more elaborate combina- tion.1 8. The latest of the coins imitated from the older Attic coins (Plate I. 5) retain the old types, the J on the cheek of Athena for the units, the broken down A0E and the pair of signs discussed above on the reverse; but they introduce certain new features, viz. the Yanaf monogram, the curved sign (see p. 10), and the very puzzling inscription j [(vyjvlfv vV2 One of these coins in the British Museum appears to have a bare male head on the obverse, instead of the head of Athena; but in its present condition this is not certain. The monogram jf (=SPS Yanaf) represents a regal surname ('exalted'). Mordtmann has remarked3 that this name occurs as the surname of three kings of Saba, all called Samah'ali; of a king whose name is missing on an inscription of Safar,4 and elsewhere; while in the form IANAA4> it is inscribed on one of the later Aethiopic coins. Mordtmann further notes that since the word has no significance in Aethiopic, the equation IANAA<t> = ^ helps to confirm Gutschmid's theory that of the two names which occur on the Axumite coins one represents the under-king of Yemen. It is obvious that all the Sabaean or Himyarite coins with the Yanaf monogram are not necessarily to be attributed to one ruler on account of that monogram only. Nevertheless, it would be un- reasonable on the ground of fabric alone to separate the coins of the group now under consideration from those of the San'a class'which are connected with them by the Yanaf monogram, the Aramaic We inscription, and the pair of signs (PI. I. 7, 8). may therefore attribute them to the same ruler, to whom must be due the intro- duction ofthe coinage imitated from the 'New Style' Attic coins. To the elucidation of the Aramaic inscription I am unable to contribute anything definite. Mordtmann (foe. cit.) holds that it shouldbe inverted,andreads it V1agash, i.e. Volagases,an Arsacid 1 Theapparentlowerpartofthemonogram on theViennesecoin,Taf. XIV. 13 (Babelon, Traite, PL CXXV1. 21), isevidentlyonlyduetodoublestrikingofthe monogram. The 3 doubtless has the same significance here aswhen itoccurs onthecheek ofAthena. a Forconvenience I call thishenceforward theAramaic inscription. M.u.K., p. 67, no. 14, give an additional letter on the right, which is, however, the remainsofthe A of AGE. 8 Num. Zeit., 1880,pp. 296f. ; Z.D.M.G., xxxi, p. 90. * 'Amdau Bayyin, who struck coins at Safar (Kaidan), was also called Yanaf (seebelow, p. 18); mayhisthen bethemissingname?