Research on this topic was aided by the resources of the American Numismatic Society and the American Academy in Rome. I am indebted especially to Professor Edwin Brown of the University of North Carolina at Chapel Hill and to Professor Calvert Watkins of Harvard University, for information and comments.

The following special abbreviation will be used: Weidauer = L. Weidauer, Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975).

¹ These are fully published by Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975). 25–7, nos. 91-113Google Scholar: twenty-three coins in all, but the legend is completely missing from 104, 106 and 107 as a result of minting procedures (see n. 2 below). However, these three coins are reverse die-linked with coins of the WALWE. series. There are also four coins (Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) nos. 76-8, 84Google Scholar) of a different lion-head series which are reverse die-linked with coins of the WALWE. series.

² The explanation of this phenomenon here cannot be that the dies were designed to mint staters, and thus on smaller coins the type would only partly be reproduced. For as Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) observes (46-7)Google Scholar, on smaller denominations the lion heads are in fact smaller. Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975)Google Scholar concludes from this that the incomplete types are not the result of any minting process but had ‘a specific purpose’—which is left unspecified. One explanation might be suggested. Having planned to strike coins with two lion heads, it was discovered that if both heads were to appear on the round flan the heads would have to be quite small in relation to it. The mint adopted the odd but more impressive solution of fully reproducing only one. This problem was resolved by the time of the earliest gold and silver coinage (the ‘Croesids’) by producing the oblong flans (stamped with lion and bull) that are characteristic of it.

³ Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) p. 60Google Scholar, and see Jongkees, J. H., Acta Orient., xvi (1938) 254–5Google Scholar.

⁴ In advance of the publication of the BM symposium (March 1984) on the date of the Ephesian Artemision, I shall not discuss that controversial issue. Most scholars have accepted a date c. 600 BC or a little later for the initial construction and the Basis Deposit; Price, M. J. (Studies in numismatic method presented to Philip Grierson [Cambridge 1983] 1–4)Google Scholar suggests a date possibly as late as c. 575 for the Deposit, which implies a date somewhat earlier for the WALWE. series. Price also believes (Studies in honor of Leo Mildenberg, edd. Houghton, A. et al. [Wettcren 1984] 221 n. 25Google Scholar, and see also Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) 107Google Scholar) that ‘both on typological and on stylistic grounds’ the WALWE. series belongs late among the various issues of electrum coinage (which he thinks continued through the reign of Croesus [c. 561/o-c. 547//6]). Price's basis for this argument (and see Studies Grierson 2) is a stylistic and typological resemblance between this scries and the Lydian bimetallic ‘bull and lion’ coinage, dated ‘not much before the middle of the century’ (ibid. n. 9), or after Croesus’ fall (Studies Mildenberg, passim). However, the inconsistency between a date close to 550 or after 546, and a date before c. 575, must discourage the use of typological and stylistic observations as a criterion for dating. Since the question of chronology is largely irrelevant to this article, I shall not discuss the arguments of M. Vickers (NC cvl [1985] 1-44) that coinage may have begun not earlier than the 540s. (However, see my comments in AJA 1987 [nn. 1, 42].)

⁵ See Jeffery, L. H., The local scripts of archaic Greece (Oxford 1961) 325–7, 345, 359-61, 289Google Scholar and Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) 60–61Google Scholar. The point was first made by Six, J. P., NC³ x (1890) 205Google Scholar.

⁶ See Pedley, J. G., Sardis in the age of Croesus (Norman, Okla. 1968) 72Google Scholar and n. 56 below; for an overview of lions at Sardis, see Hanfmann, G. M. A. and Ramage, N. H., Sculpture from Sardis (Cambridge, Mass. 1978) 20–22.Google Scholar

⁷ See in particular the well-known hoard containing coins only of this type, found at Gordion in Lydian territory: Bellinger, A. R., Essays in Greek coinage presented to Stanley Robinson, edd. Kraay, C. M. and Jenkins, G. K. (Oxford 1968) 10–15Google Scholar. One of these Gordion coins [no. 1104 = Bellinger pl. 1.27 = Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) 104Google Scholar], though heavily worn and missing its legend, is die-linked with Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) 103Google Scholar, one of the three other WALWE. coins found in an unstratified context during Hogarth's excavations at the Artemision. For the Sardis fragment see Goldstein, S., BASOR ccxxviii (1977) 54–7 and fig. 8Google Scholar, Sardis from prehistoric to Roman times (Cambridge, Mass. 1983) 40 and fig. 60Google Scholar. Virtually all scholars agree that these lion-head coins are Lydian. This is not the place to argue the thesis that virtually all early coins were issued by governments or states. (However, see Part II of my article. ‘The origin of electrum coinage’, AJA [1987].)

⁸ The earliest such reading I have traced is that by Mionnet, T. E., Description de médailles antiques, grecques et romaines ii (Paris 1807) 528 no. 84, Pl. 35 no. 168Google Scholar (this is ignored, e.g., by Six [n.5] 205; Jongkees [n. 3] 251 n. 1; Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) 59Google Scholar). Mionnet is followed by Sestini, D., Descrizione degli stateri antichi (Florence 1817) 51–2Google Scholar and Brandis, J., Das Münz-, Mass- und Gewichtswesen in Vorderasien (Berlin 1866) 177Google Scholar and n. 1, though Mommsen (Geschichte des römischen Münzwesens [Berlin 1860] 20 n. 73Google Scholar) had doubted the attribution and thought the legend possibly Lycian. In the light of Brandis’ publication Head, B. V. (NC² xv [1875] 266)Google Scholar assigned to Kyzikos a second specimen of the WALWE. type (Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) no. 97Google Scholar).

⁹ Hofmann, K. B., NZ xvii (1885) 2–3Google Scholar, cf. xvi (1884) 32; Six (n. 5) 202-210.

¹⁰ See Weidauer's summation of the evidence (p. 60).

¹¹ JHS xlvi (1926) 36-41.

¹² This reading has been accepted, e.g., by Robinson, E. S. G., JHS lxxi (1951) 163Google Scholar; Seltman, C., Greek coins² (London 1955) 25–6Google Scholar; Pedley (n. 6) 74; it is still cited (in the form ‘Walwes-’) as one of two possible readings by Kraay, C. M., Archaic and classical Greek coins (London 1976) 24.Google Scholar

¹³ Jongkees (n. 3) 254.

¹⁴ Gusmani, R., Oriens Ant. xiv (1975) 271Google Scholar, ASNP³ viii.3 (1978) 835, 847. This point had been suggested by Jongkees on the basis of the single example of an archaic s known at the time of his article.

¹⁵ This is followed by Gusmani, , Lydisches Wörterbuch (Heidelberg 1964) 220Google Scholar.

¹⁶ Cf. Gusmani ibid. (n. 15); Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) 61Google Scholar; Kagan, D., AJA lxxxvi (1982) 358Google Scholar; M. J. Price, Studies Grierson (n. 4 above) 2. The reading ‘valvel’ is not mentioned by Kraay (n. 12, cf. n. 17 below).

¹⁷ Aulock, H. von, Syll. numm.gr. (Berlin 1967) no. 8204, Pl. 11.Google Scholar Von Aulock's reading is accepted by Kraay, , Gnomon 1 (1978) 212Google Scholar. Unfortunately it is not discussed by Weidauer, though she knows the publication and reproduces a blurred copy of its photograph.

¹⁸ Lydisches Wôrterbuch, Ergänzungsband ii (Heidelberg 1982) 106Google Scholar. It is curious that Gusmani in fact seems to be ignorant of Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) 95Google Scholar and von Aulock's publication. His interpretation of the final letter apparently rests only on the observation that it seems to differ from the third letter and therefore cannot be. (This argument was discredited by Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) 61.Google Scholar)

¹⁹ Against Buckler see Jongkees (n. 3) 252; against Jongkees see Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) 62Google Scholar with Gusmani, , Kadmos viii (1969) 158–61Google Scholar.

²⁰ Cf, e.g., Seltman (n. 12) 25-6; Robinson (n. 12) 163 and NC⁶ xvi (1956) 4; Pedley (n. 6) 74-5.

²¹ For reff. to these ‘KALI’ coins, see nn. 43 and 45 below. Because the reverse punch on the 1966 coin seemed less worn than the same punch when used on one coin of the WALWE. series, M. Thompson concluded that the 1966 coin was part of an earlier issue. Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) objected to this (p. 106)Google Scholar that the same punch was used on the 1968 coin and shows no greater wear than on the two die-linked WALWE. coins (Weidauer, L., Probleme der frühen Elektronprägung, Typos iv (Fribourg 1975) nos. 97, 99Google Scholar). This suggests that the mint was striking both series at the same period.

²² Cf, e.g., Babelon, E., Traité des monnaies grecques el romaines ii.i (Paris 1907) cols. 44-6Google Scholar; Buckler (n. 11) 36-8; Jongkees (n. 3) 251; Gusmani (n. 15) 220-1, (n. 18) 106 with ref; M. Thompson (n. 45 below) 1-4; Breglia, L., AIIN xviii (1971) 16Google Scholar; Kraay (n. 12) 24, (n. 17) 213; Holloway, R. R., RIN lxxx (1978) 8, 11Google Scholar.

²³ For the identification with mint-masters, see, e.g., Kraay (n. 12) 24-5; Kagan (n. 16) 358; Gusmani (n. 15) 220; Kroll, J. H., ANSMusN 26 (1981) 5Google Scholar; cf. Jongkees (n. 3) 254. For a general (and sceptical) discussion of possible mint-masters’ names on ancient coins, see Furtwängler, A., RSN lxi (1982) 5–30Google Scholar. The monograms of abbreviations of the names of some type of official on a few royal coinages of the Hellenistic Greek world (see Furtwängler 9–12) scarcely constitute a parallel. On these coins the king's name, in full, is always a prominent feature.

²⁴ Sardis from prehistoric to Roman times (Cambridge, Mass. 1983) 78Google Scholar. In 1980 Pászthory, E. could still speak of lion-head coins ‘of the Alyattes type’ (Metallurgy in numismatics i, edd. Metcalf, D. M. and Oddy, W. A. [London 1980] 151)Google Scholar.

²⁵ Steinherr, F., Die Welt des Orients iv (1968) 320–5Google Scholar; H. Otten, ibid. v (1969) 94-5. Nougayrol, J., Le palais royal d'Ugarit iv (Paris 1956Google Scholar = Mission de Ras Shamra ix) 267 had suggested the identification, but without argument.

²⁶ For a succinct discussion of the methods used to decipher Hittite and Luwian see Laroche, E., Dictionnaire de la langue louvite (Paris 1959) 10–15Google Scholar; On the different Anatolian writing systems see Morpurgo-Davies, A. and Hawkins, J. D., ASNP³ viii. 3 (1978) 776–8Google Scholar.

²⁷ See Steinherr, F., Orientalia xx (1951) 108, 112-13Google Scholar and reff.

²⁸ See Laroche (n. 26) s.v. ‘walwi’, and cf. ibid. s.w. ‘-ziti’, ‘Piha-’, ‘Mula-’, ‘takiti-’; with id., Les noms des Hittites (Paris 1966) nos. 1486, 972, 817, 1210Google Scholar.

²⁹ See Friedrich, J., Hethitisches Wörterbuch (Heidelberg 1952) 299Google Scholar; and see Laroche Noms (n. 28) for UR. MAH names with Luwian cuneiform prefixes Piha- (976.1), Ali- (34), Hapati- (284), Lila- (697), Muwa- (839.2). Numbers given in the text for Anatolian names are taken from this work.

³⁰ See Steinherr (n. 25) 322-3 for a list of Hieroglyphic seals that include the logogram LEO (including Piha-LEO and LEO-ziti). This logogram (Meriggi 88 = Laroche, Les hieroglyphes hittites [Paris 1960] no. 97, p. 61Google Scholar) is very clearly a picture of a lion.

³¹ At Noms (n. 28, s.v.) Laroche withdrew his earlier suggestion that the Luwian pronunciation of UR.MAH was ‘-neši’.

³² On the meaning of ^{SAL. MES (=female pl.)} hazgara ( = [cult functionaries]) see Friedrich (n. 29) 68; on the nature and activities of these male and female cult functionaries, see Jakob-Rost, L., Orientalia² xxxv (1966) 420Google Scholar.

³³ To be quite precise, Luwian itself consists of three closely related languages: cuneiform Luwian, Hieroglyphic Luwian, and Lycian: see Laroche, E., BSL lxii (1967) 64Google Scholar and Hawkins, J. D., Morpurgo-Davies, A. and Neumann, G., NAWG vi (1973) 143–97Google Scholar.

³⁴ Indeed, in the Hittite lawcode ‘Luwiya’ is once used interchangeably with ‘Arzawa’: Catalogue des texles hittites (Paris 1971) Law 291 no. 19A/BGoogle Scholar, with Heinhold-Krahmer, S., Arzawa (Heidelberg 1977) 21–2, 318Google Scholar.

³⁵ See Smith, S., JEA 8 (1922) 45–7Google Scholar; Goetze, A., ZAssyr xxxvi (1925) 305–8Google Scholar, Kizzuwatna (New Haven 1940)Google Scholar.

³⁶ See Goetze, A., Journ. Cun. Stud. xiv (1960) 47Google Scholar; Bryce, T. R., AnatSt xxiv (1974) 103–16Google Scholar; Heinhold-Krahmer (n. 34) 3-4, 333-4, who points out that the evidence is still insufficient to determine the exact location of Arzawa.

³⁷ See Cornelius, Fr., Orientalia xxvii (1958) 394–5Google Scholar, Anatolica i (1967) 62; Garstang, J. and Gurney, O. R., The geography of the Hittite empire (London 1959) 84–5Google Scholar; above all, Macqueen, J. G., AnatSt xviii (1968) 169–85Google Scholar, especially 174-5 (accepted by J. Mellaart, ibid. 187).

³⁸ See Heinhold-Krahmer (n. 34) 30, 81-2 and 351-2, with reff. to Cornelius and Garstang and Gurney, and C. Watkins in Proc. Symp. on Trojan War, ed. M. Mellink, forthcoming.

³⁹ See Heubeck, A., ‘Lydisch’, in Altkleinasiatische Sprachen, Handb. d. Orient, i.2.2 (Leiden and Cologne 1969) 419–24Google Scholar; Gusmani (n. 18) 5-8; Oettinger, N.ZVS[= KZ] xcii (1978) 74–92Google Scholar. Oettinger derives Lydian and *Palaic-Luwian (ancestor of Luwian and Palaic) from ‘*Urwestanatolisch.’

⁴⁰ See Pedley (n. 6) 26.

⁴¹ Steinherr (n. 25) 324; Gusmani, loc. cit. (n. 18).

⁴² That this would have been recognized as the meaning of the coin-legend is reinforced by the continued use of one branch of Luwian—Lycian—in western Asia Minor down through the Hellenistic period: see Houwink, Ph. H. J. ten Cate, The Luwian population groups of Lycia and Cilicia Aspera during the Hellenistic period (Leiden 1961)Google Scholar.

There was apparently no Indo-European word for ‘lion’: see Walde, A. and Hofmann, J. B., Lateinisches etymologisches Wörterbuch⁴ (Heidelberg 1965) 785Google Scholar; Windekens, A. J. Van, Orbis xxiv (1975) 211–13.Google Scholar

⁴³ C. M. Kraay, Report of the visitors [of the] Ashmolean Museum 1968, 43-4.

⁴⁴ Cf., e.g., Price (n. 16) 5; Holloway (n. 22) 8, 11; Kagan (n. 16) 358 (‘kalil’: probably the name of a moneyer).

⁴⁵ Thompson, M., ANSMusN xii (1966) 1–4Google Scholar.

⁴⁶ At AIIN xviii (1971) 16Google Scholar L. Breglia refers to the 1966 coin as ‘segnato )Kal[’ (N.B. the capitalization); at ASNP ³ iv. 3 (1974) 659 she refers to it simply as ‘segnato Kat’.

⁴⁷ Gusmani (n. 18) 86.

⁴⁸ Kraay (n. 12) 24.

⁴⁹ Gusmani (n. 14).

⁵⁰ Gusmani, , Kadmos xi (1972) 47–51Google Scholar and (n. 18) 58 s.v. ‘-im’, cf. Die Sprache xvii (1971) 1–6.Google Scholar

⁵ Gusmani (n. 18) 86.

⁵² See Zgusta, L., Kleinasiatische Personennamen (Prague 1964) 662Google Scholar.

⁵³ See, e.g., Balmuth, M., Studies presented to George M. A. Hanfmann, edd. Mitten, D. G. et al. (Cambridge, Mass, and Mainz 1971) 5–6 and reff.Google Scholar

⁵⁴ See Franke, P. R. and Schmitt, R., ΦANEO∑-ΦANO∑ EMI ∑HMA, Chiron iv (1974) 1–4Google Scholar. for the attribution to Ephesos, see Weidauer 68 (now widely accepted). Kraay ACGC 23 had made a case for Halikarnassos.

⁵⁵ See Gusmani, , Kaimos xi (1972) 49, 51Google Scholar.

⁵⁶ On the (possible) identification with lions by Assyrian kings, see Buren, E. D. Van, Anal. Orient. xviii (Rome 1939) 6Google Scholar; Koldewey, R., Die Königsburgen von Babylon (1931-2, repr. Osnabrück 1969) i pp. 7, 9, 20-1, ii pp. 5, 9Google Scholar (N.B.: Nebuchadnezzar stamped clay bricks used for his buildings with lion-seals); Porada, E., BullMFA xlviii (1950) 2–8Google Scholar; and cf. Frankfort, H., The art and architecture of the ancient Orient (Harmondsworth 1958) 104 fig. 41Google Scholar. (Assurnasirpal's tunic). For a similar motif in Achaemenid Persia, see Ettinghausen, R., Oriens xvii (1964) 161–4Google Scholar. The appelation ‘lion’ is common in the NT: see i Macc. 314-7 forjudah as lion. For further reff., see Sardis from prehistoric to Roman times (n. 24) 184 and n. 6 above.

⁵⁷ See Ramage, A., BASOR cxci (1968) 11–2Google Scholar, cic (1970) 16-22; Sardis from prehistoric to Roman times (n. 24) 34-7.

⁵⁸ On this case ending, see Gusmani (n. 15) 43. It should be noted that Gusmani does not cite this as a possible interpretation of ‘valvel’.