¹ This catalogue owes much to the assistance given to me by many members of the staff of the Archaeological Museum in Rhodes and of the Department of Greek and Roman Antiquities of the British Museum, where the presence of a card index to the Fikellura graves, compiled by Professor Corbett, also considerably facilitated my work.

To avoid much repetition, I use very abbreviated forms of reference to objects in these two museums throughout the articles, e.g.:

11234 (vi 246) = Rhodes inv. 11234 (Clara Rhodos vi–vii 246).

BM.333 (F 130) = British Museum 64.10–7.333 (Fikellura grave 130).

Unless otherwise stated all marks are graffiti and under the foot of the vase, all dipinti are red, and all the marks are reproduced at the scale of 1:1 or 1:2 in the illustrations.

Other abbreviations:

ASA = Maiuri, A., ‘Jalisos’, ASA vi–vii (1923–1924) 83 ff.Google Scholar

Biliotti lot… = Lot … in the sale of material excavated by A. Biliotti, Sotheby, 3–5 December 1885.

BMC = British Museum Catalogue of Vases (1893–1925).

Guarducci = Guarducci, M., Epigrafia Greca i (Rome, 1967).Google Scholar

Hackl = Hackl, R., ‘Merkantile Inschriften auf Attischen vasen’, Münchener Archäologische Studien (Munich, 1909), 5 ff.Google Scholar; catalogue numbers unless otherwise noted.

Lindos = Blinkenberg, Chr., Lindos i, Les Petits Objets (Berlin, 1931).Google Scholar

LSAG… = Jeffery, L. H., The Local Scripts of Archaic Greece (Oxford, 1961), 356–9Google Scholar, no. …

NC = Payne, H., Necrocorinthia (Oxford, 1931)Google Scholar, catalogue nos.

TCam = Segre, M. and Pugliese Caratelli, I., ‘Tituli Camirenses’, ASA N.s. xi–xii (1949–1951) 141 ff.Google Scholar

² Jacobsthal, , Gott. gelehrte Am. cxcv (1933) 1 ff.Google Scholar I note here also that Jacopi ignores graffiti and dipinti in his CVA volumes for Tarquinia.

³ The Director of the Musée Cantonal, Lausanne, M. R. Wiesendanger, kindly informs me that none of the vases from Salzmann's excavations, now in Lausanne, bears any mark.

⁴ I append some notes on these pieces:

13809 (vi 90, LSAG 9), the komast vase, is regrettably no longer in the museum, and so it is impossible to check the reading, as suggested by Guarducci, Rend. Linc. (1970) 51 ff. I find at least the beginning of Jeffery's interpretation more persuasive; where she reads a zeta, for the tau in Cl.Rh., I would merely comment that it is not difficult to overlook a horizontal line cut on a reserved surface; in general, I have found Jacopi's readings of the more substantial inscriptions accurate.

12894 (iv 173, LSAG 20) is also discussed by Guarducci, loc. cit.; there can be no doubt about the reading in Cl.Rh. (save the koppa) and so Guarducci's emendations can scarcely stand; therefore, whatever our misgivings, we must assume this to be have been a gift from Akrathetos to another man's wife, not his own.

13757 (vi 56, LSAG 36). In the second graffito there is a further letter after the but it is ill preserved, as is the third letter. The final letter seems to be , although a stroke before it may be intentional, giving a nu; it would be tempting to see a theta in the third letter (certainly the vertical line seems to run through the circle), giving The over-all state of preservation of the whole admittedly makes this a hazardous conjecture.

6581 (iii III) is worth noting. I have nothing to add to the publication, save to stress the unusual doubled consonant, and to suggest that the vase may well have been used as a measure at Ialysos. The graffito has no obviously Rhodian characteristics, but it would be a little surprising to find a jug already marked in this way being exported. In any case we may wonder how and why an Athenian measuring-jug reached Rhodes; its measurements, ht. 0·13, diam. 0·073, are comparable with those of pieces from the Agora—Agora x 61–2.

⁵ But see on pyxides, below, p. 164.

⁶ The large aryballoi are Crotone CR 512, with pi, and Syracuse 7496, from Megara Hyblaea, with phi. The one smaller vase is the lower part of a quatrefoil aryballos from Naukratis, Cambridge, Museum of Classical Archaeology, NA 188, with a faint nu. Dr. E. T. Hall of the Research Laboratory for Archaeology in Oxford kindly attempted a spectrographic analysis of this mark; the state of preservation was too poor for any positive result, but Dr. Hall reports that he is reasonably sure that no lead or mercury is present, which suggests that the pigment is an iron oxide, which cannot be isolated in view of the amount of iron already present in the body of the vase.

⁷ I know of eleven at present; published are 7, another oenochoe, Tocra i 972, with a difficult little graffito, a hydria, Copenhagen ABc 905 (CVA pl. 222, 5), which is published as Etruscan, but is now talen by Shefton as Laconian, with some interesting numerals cut on the shoulder, and four kraters:

Villa Giulia 47995, MA xlii 534, with red AD.

Villa Giulia (foot only), NSc (1937) 417, 29, with graffito, phi.

Limoges 81.00 (CVA pl. 2, 6), with red Χ.

BM o.c. 225 with a partly alphabetic graffito.

⁸ For this problem see BSA lxvii (1973) 187.

⁹ Berlin 1147, krater by the Memnon Painter from Cerveteri (NC 1170). Louvre E618, krater, Middle Corin thian. Palermo N1 1674, floral pyxis, Late Corinthian, from Selinus (?). Oxford 1928.11, lekythos in the manner of the painter of the Slim Sphinxes (worn, perhaps phi).

¹⁰ Some of the larger pieces from Tocra and Tell Sukas may fall into this category, e.g. Tocra i 589, 601–2, and 606, Sukas ii 247 and 251. It is not easy to be dogmatic about the place of manufacture of many of the marked vases from Naukratis; I take Naukratis i 429–33, 450, and 358 (all bowls) to be certainly Rhodian.

¹¹ Two other marks are known to me on Clazomenian vases: Vienna KH loan, amphora with red Σν (BSA xlvii (1952) 134 D3) and the Chanenko askos in Kiev, with a red numerical mark (AA (1929) 235–7).

¹² I know of three other marked Fikellura vases:

BM 1949.5–16.7, neck-amphora (Cook, ibid. N3) from Tell Defenneh, with a ligature of eta and rho in red outlined by incision.

BM 1910.2–22.48, fragment of a neck-amphora from Naukratis, with graffito on wall.

Bucharest, oenochoe (Cook U8) from Istria, red delta.

¹³ Syracuse 12648, from Syracuse, ΚΑ on neck, FIG. 2. Syracuse 24670, from Gela (MA xvii 279–80), FIG. 2. My reading differs from Orsi's; I cannot see a kappa in the first mark, and I am not sure if the rightmost of the upper horizontals was intentional; both marks seem nonalphabetic, pace Orsi, and Agostiniani, , St. Etr. xli (1973) 392Google Scholar, 5. On the shoulder.

Syracuse 24691, from Gela (MA xvii, 374 perhaps, but the description only partly fits), FIG. 2; dipinto ligature of lambda and eta?

Syracuse, from Gela (MA xvii 62), graffito on shoulder ΚΙΜΣ.

Syracuse, from Gela (MA xvii 281), graffiti on shoulder and base of handle.

Taranto 4892, from Taranto (ASA xxi–xxii (1959–60) 126, note), FIG. 2.

Taranto 20739, from Taranto (ASA ibid.), FIG. 2; both these last on shoulder.

¹⁴ The ‘standard’ lambda-eta ligature on Leagran vases must refer to a vase-name or similar designation, if only because of the company it keeps, see PdP xxvii (1972) 418–19. However, we may note that a dedication incised on a sherd of somewhat earlier date than the Leagran period found at Gravisca was made by a I know of no simple lambda-eta ligature on a large vase earlier than the later sixth century, but unligatured lambda eta (or heta upsilon?) occurs dipinto on Hannover 1967.11, a type B amphora, probably not Attic, of c. 540, and possibly Taranto 4320, Panathenaic-type amphora of the group of Vatican 347, although the eta is not well preserved.

¹⁵ For bibliography of the type see CVA München vi 49–50 and Gela ii IID 4–5. Inscribed vases are:

Munich 460, 463 (FIG. 2, the CVA reading is deficient), 467, 468.

Tarquinia RC 1145, with Τν on the neck; RC 1860 (FIG. 2); RC 3440, FIG. 2, on the shoulder.

Once Rome, Società Hercle, from Vulci, St. Etr. xxxiv (1966) 318–19. Louvre S3943, FIG. 2.

Bonn, Hackl p. 17, dipinto, called glaze by Hackl, red by Beazley (archive notes in Oxford).

Of the same fabric is the hydria, Villa Giulia M417; the mark is not quite as given by Mingazzini (as 418 on his pl. 1), since the top bar of the epsilon joins the first line of the second sign.

¹⁶ The shape varies from an apparent broken-bar alpha to two joining xs, but it is probable that many of the marks have the same significance. Examples: Chiot, , Naukratis i 67Google Scholar, BM 88.6–1.767 and 772 (both from Naukratis); Rhodian, , Tocra i 684, 685Google Scholar, Istria ii no. 69; other Greek, East, Naukratis i 368 (handle)Google Scholar, 369 (plain vase, dipinto), 408 (on body of storage amphora); Chalcidian, neck-amphorae, Rumpf 57, 58, 59, and 67 and New York 46.11.1 (Bothmer, , BMM (1946–1947) 132–3)Google Scholar; Attic, a kylix foot from Gravisca (also two non-Attic feet, one of a large closed vase), the foot of a small closed vase from the Campana collection in Florence, and, later, Boston 98.882, type C amphora by the Flying-Angel Painter, as well as 85.

¹⁷ See Hackl group XVI. I know of 54 examples, 23 on vases of the Leagros Group in the broadest sense. The ‘tick’ alone appears on two further Leagran vases, BM B239 and Villa Giulia 50619 (M497), as well as the doubleen Leiden PC55, and a contemporary lekythos presumably from Sicily, Gela, Navarra inv. 1 (CVA iii pl. 1, 1–2; 2, 1 and 3; the graffito, ZPE 17 (1975) 155).

¹⁸ Red marks are also found on some tomb-slabs, III 190, tomb CLXXXII and VIII 47, Annuachia tomb I.

¹⁹ For Timosthenes see ADelt xvii (1961–2) β 270; we should note one vase bearing a different name, Polykles, and no trace of a body, save for a glazed krater marked with full Laconian deliberation

²⁰ Although I comment on it below. Nor do I use the evidence of the partly Rhodian colonies of Gela and Akragas, whose red script need not have been imported from the mother city of Lindos.

²¹ There is an unfortunate clash of opinion over the reading of this letter; Bernand, and Masson, , REG lxx (1957) 10–13Google Scholar reject the horizontal strokes, but they are reinstated by Guarducci 330 with note 1. The result is that one cannot feel confident about an important piece of evidence, and although I argue as if the hour-glass form is assured, it should be borne in mind that this part of my case may be overstated.

²² I forbear giving here a list of its uses, alphabetic or otherwise, in archaic inscriptions; for a recent treatment, Guarducci, , ASA N.S. xiv–xvi (1952–1954) 167 ff.Google Scholar It may appear rarely, but it appears in a large number of scattered areas. Alphabetic uses to add to those firmly attested for Sikyon and Knidos are in a graffito on an SOS amphora from Cheli, Porto, BCH xc (1966) 788Google Scholar and in a dipinto under the foot of a mid sixth-century Attic neck-amphora, BM B260 (the mark published in the old catalogue, no. 508).

²³ See Guarducci 82–3 for a statement of this problem.

²⁴ The use of an omega in it is also worrying if the piece is of Kamiran manufacture of the mid sixth century; the form of the genitive singular used could be Ionic or Rhodian. Yet it is difficult to suggest any other place of production for such artefacts; see Bissing, von, Sitz. Bay. Ak. Wiss. (1941) II 7, 26.Google Scholar

²⁵ See Cahn, Knidos 190. I do not know how best to interpret the change in the legend on the coins of Ialysos, sometime in the early fifth century, from to Some non-Dorian influence was clearly at work around 500, but it is not purely Ionic influence, otherwise eta would have been used, but nor can we immediately connect it with any Kamiran and Lindian penchant for using epsilon for ē. Could it be that was not used at all as a vowel at Ialysos for much of the sixth century (between the attested uses at Abu Simbel and on 144)? Such a consideration could alter the direction of my argument concerning but we would still be left with different alphabetic usages at Ialysos and Kamiros. As far as the coinage is concerned, the failure of the Ionian revolt may have caused the change back to Dorian ways.

²⁶ Χ does appear on 173, but we do not know the date of the piece; it could well be of a period when the Kamiran alphabet had become fully Ionicized.

²⁷ A Lydian type Bamphora in Orvieto, , St. Etr. xxx (1962) 97, 457Google Scholar, may be close, though I have not seen the piece. ‘Miltos’ is used decoratively on a wide range of vases of the later fifth century and appears earlier on smaller shapes.

²⁸ The red (?) swastika which represents the sum of marks on large vases from Xanthos is earlier, see Metzger, , Xanthos iv 103.Google Scholar There are also some earlier dipinto ΔΙ marks, e.g. Louvre F48 and F60 (both retrograde) and E825, C. 565 (orthograde, unpublished).

²⁹ The publication of the material from the Agora should confirm this statement. I have examined a number of black-figure amphorae in the National Museum in Athens and found no marks (confirming the opinion of Riezler apud Hackl 91–2). Decorated vases of any size from Central Greece bearing marks which could fall into the ‘commercial’ category seem confined to: New York 51.11.3, type B amphora by Lydos, with graffiti on the lip (CVA iii pl. i), Tübingen 717.061, oenochoe by the Painter of Vatican G49, with graffito, from Attica, Athens 1690, Nolan amphora by the Charmides Painter from Corinth, with graffito ligature of mu and epsilon (Beazley archive notes), and once Oxford, Miss., pelike by Myson from Athens with graffito CVA Robinson ii 23.

³⁰ The types of Attic vases sent to Rhodes do not match up closely with those sent to Etruria, although the discrepancy is not great; most of the large ‘Etruscan’ groups are represented on Rhodes—Tyrrhenian Group (ABV 99, 55), Lydan circle (33 and ABV 126, 32), Swing Painter (ABV 305, 9 and 306, 48), Antimenean Group (ABV 281, 12 and 288, 17), and Leagros Group (University College, Dublin 331 = Johnston, A Catalogue of Greek Vases in Public Collections in Ireland 984).

³¹ ‘Long’ dipinti are found on two horse-heads, Tarquinia 633 (Hackl p. 62) and Copenhagen 14933 (Auktion 34, 119), and on Florence 70995 (Hackl p. 62), Stockholm, private collection (Auktion 34, 121), both from the Lydan circle, and Orvieto, no details known (Hackl 32). Horsehead amphorae have been studied recently by Birchall, , JHS xcii (1972) 48 ff.Google Scholar and Picozzi, , Studi Miscellanei xviii 5 ff.Google Scholar For the horse-head as a proto-Panathenaic amphora, Thimme, , Jb. Staat. Kunstsamm. in Baden-Württemberg viii (1971) 251–2Google Scholar, Benton, , BSA lxvii (1972) 15Google Scholar, and Boardman, Athenian Black Figure Vases 18.

³² AJA xlii (1938) 495 ff.

³³ We may note in particular a group of horse-heads with a dipinto mark pi rho, Munich 1362 (Hackl 15, CVA i 8) together with Munich 1361 and Erlangen M930, whose marks are noted as probably similar to that of 1362 by Hackl ibid. Similar marks appear on two other type B amphorae, with ordinary figure decoration, of the period 550–540 B.C. (an indication of the late date of the horse-heads), Munich 1381 by the Towry White Painter and 1400 from the circle of Lydos (Hackl 16–17, CVA i 15 and 23).

³⁴ For possible sixth-century vase prices see Greece and Rome 1974, 147–8. If we omit the Leagran vases cited there, there only remains the disputed vase-inscription on the ex-Hearst amphora, New York 56.171.13, CVA iii 8 (with bibliography; more recent discussions have added little).

³⁵ The most striking examples are perhaps Florence 92167 (Hackl 74), Villa Giulia 15537, and Bonn 42 (the latter pair lid only) by the Affecter (see H. Mommsen's forthcoming monograph), Berlin 2159 by the Andokides Painter (Knauer, Die Berliner Andokides-Vase 22) and some lids with the Leagran ‘tailed delta’ mark, Hackl type LXIV, which must belong to Leagran neck-amphorae so marked, though none is at present so paired (only one is published, Cab. Med. 252, CVA ii 56, in an erroneous pairing; others in the Louvre and British Museum). These marks are all post-firing, unlike those on the pyxides discussed below; the regularity of the pairings is a little disconcerting since what was needed was a different mark for each vase plus lid, not the frequent repetition of the same mark.

³⁶ Excluding the simple red cross or wheel which appears commonly on Apulian pelikae and oenochoae, e.g. BM o.c. 1713, etc., and Louvre G552, the only red dipinti on vases of the later fifth century and after, which have come to my attention are: Louvre G353, column-krater by the Orestes Painter with only traces preserved; Compiègne 1090, kylix by the Wedding Painter, with lines and dots; Cab. Med. 852, askos by the Curtius Painter (Beazley archive notes), with pi; Vatican 17915, neck-amphora by the Painter of the Louvre Centauromachy, with traces on the navel; and Syracuse, pelike from Akrai, c. 425 (NSc. (1915) 210) with ligature of (h)eta and rho.

³⁷ The earliest broken-bar alpha otherwise known is in a graffito of 350–325 B.C. from the Agora (Agora xii 1317, with bibliography).

³⁸ See note 16.

³⁹ e.g., in ligature, Leningrad St. 1528, column-krater by the Berlin Painter (Peredolskaya, Krasnofigurie Atticheskie vasi 39), not in ligature, and accompanied by further marks, Louvre G178, hydria in imitation of the Berlin Painter, from Vulci, and G232, pelike by the Syleus Painter, from Chiusi?, and New York 07.286.78, type B amphora by the Eucharides Painter, from Akragas?

⁴⁰ Triple lines on Syracuse 22785 (or 58?), column-krater by the Florence Painter, and 26564, column-krater, 460–450; double lines on 22886, bell-krater by the Eupolis Painter; all three vases from Kamarina, the first and third MA xiv, 793–4 and 850–1.

⁴¹ I am indebted to Dr. H. Mommsen for supplying me with details of 86; there are a number of faint and illpreserved strokes after the deltas; they could be intentional. The stroke of the alpha which Beazley omitted nearly coincides with a wheel mark, but seems assured. For the price interpretation see Amyx, , Hesperia xxvii (1958) 299–300.Google Scholar The Bowdoin vase, Herbert, Ancient Art in Bowdoin College, no. 196 and Johnston, , Greece and Rome (1974) 144Google Scholar, and also 151, 9.

⁴² Al Mina, Beazley, , AJA lxi (1957) 8Google Scholar; Odessos, , Toncheva, , BIAB xxvii (1964) 111–13.Google Scholar Compare the price inscription in syllabic script on New York 22.139.11, bellkrater by the Cassel Painter (Richter and Hall 224).

⁴³ There is only one addition I wish to make to Jacopi's remarks: the much-fragmented inscription 14611 could well belong to the growing number of monumentally ncised or painted dedications on prize Panathenaic amphorae; see Metaxa, and Frel, , AAA v (1972) 246–7.Google Scholar There are further scraps in Olympia and I am tempted to include Lindos 2820. The size and careful cutting of all these graffiti is far removed from everyday scrawl.

⁴⁴ I am indebted to Dr. Philippaki for enabling me to study many of the graffiti published in Graef, and Langlotz, , Die Antiken Vasen von der Akropolis zu Athen ii 114 ff.Google Scholar, which I refer to here by their individual numbers, not by the catalogue entries. Of the so-called mercantile graffiti, 382 ff., a number are incorrectly recorded, while the use of available fount rather than drawings makes it potentially dangerous to rely on the reported readings of a large number. I give some examples which are more misleading than the rest:

390. The second letter is a phi; the two rhos have extremely small loops and long stems.

404. Theta only.

413. The shape of the second mark only vaguely approximates to that printed and its relationship with the pi is far less close.

417. Both letters are read much more easily as two adjoining deltas, perhaps numerical.

447. There is a further horizontal at the base of the sign.

449. The isolated letter has no cross-bar.

451. The final letter has no cross-bar and is not the final letter, being followed by a large alpha-shaped sign at an acute angle to the rest.

457. The second sign is a ligature of upsilon and kappa.

461. More like a koppa with a rhomboid head.

464. Scarred to the left (as printed) and the two parts are joined.

468. The three parts are all merged into one irregularly cut sign.

I add the following corrections to some of the other graffiti:

316. The first preserved letter is certainly a theta.

334. An alpha is omitted between the iota and pi in the first line; the letters of the second line are widely spaced and there are traces of a third letter at the right edge.

338. Published upside-down; the letter on the right is an alpha.

341. I read … ]ΔΕΜΕ[ …

355. There is only one lambda in the second lambda is an independent straight line which reappears on the opposite side of the foot (an apparent example of the mark on 72, but completely isolated from the rest by date and type of vase—bf type A cup, c. 520).

372. The alpha and lambda are incised over a mark approximating to

377. The last three letters are given a separate line and written ςΟΚ

⁴⁵ The arrow delta and the dotted form deserve a full study. I am not convinced of the exclusively Samian origin of the latter (Barron, , JHS lxxxiv (1964) 45 n. 60Google Scholar; Bothmer, , BMM (1968–1969) 432–3)Google Scholar, nor that the two forms are necessarily connected, despite appearing at approximately the same time. The arrow delta is used as a numeral in many areas, though its origin is not clear—I suspect that it may not be alphabetic, i.e. acrophonic, at all; certainly there is no evidence that points to the use of any numeral system other than the ‘Milesian’ in Asia Minor in the sixth and early fifth centuries.

⁴⁶ For the practice see Agora xii 178, with lists.

⁴⁷ Not only are such marks under the glaze on the underside of the lid (and consequently easily overlooked), but the lines of the graffiti have a different character from those of most marks, having slightly raised sides; this is of special interest as it tends to show that most graffiti were applied after firing. Pre-firing marks known to me are: Oxford 1964. 325, mentioned above, Oxford 334 (CVA i pl. 4, 6) with perhaps a tally mark, Vienna 3718 with mu, Vienna 1971 with crossed theta (scarcely alphabetic in the late fifth century) and alpha, BM TB1054 with beta, and several pieces from the Agora, not all with both lid and bowl preserved (P5711 with kappa iota and P24795 with lambda, Agora xii 1315 and 1310 respectively; P22824 with delta and P23549 with pi).

⁴⁸ This may be a crumb of evidence for the survival rate of type D pyxides, but more collecting of evidence is needed. 124 and 138, from F144, have similar marks even if they are not easily interpreted.

⁴⁹ See note 39, and add examples of alpha-rho in ligature on the stamnoi Castle Ashby 25 and Munich 2406 (Hackl 358–9), not in ligature on the volute-krater BM E468 (Hackl 361).

⁵⁰ Knigge, , AA (1972) 602.Google Scholar See also the case of Timosthenes' empty (?) tomb, above, n. 19.

⁵¹ e.g. Corbett, , JHS lxxx (1960) 60Google Scholar; Benton loc. cit. in note 31; Robertson, , Gnomon xxxix (1967) 821.Google Scholar

⁵² It is possible that this vase is to be identified with Canberra 65.03, of about 425 B.C., donated by the Museum of Classical Archaeology, Cambridge, but with no recorded earlier history (information from Miss A. Moffatt and Professor R. M. Cook). The mark on the Canberra vase is small.

⁵³ Hackl 363–4, 369–76.

⁵⁴ Two numerical marks should be so regarded: SCE ii, Marion tomb 60 (33), a glazed bowl, and Vente Dromt 28/4/1887 lot 147, another bowl from Marion with an obviously numerical mark, but one difficult to interpret.

⁵⁵ To judge from the archaeological evidence (if not the literary), the exploits of the Euboeans in the West began long before those of the Rhodians (Ridgway, ‘The First Western Greeks’, in Greeks, Celts and Romans 10 and 20). LSAG 1 is later than the start of Rhodian commercial expansion, a, evidenced by the wave-line aryballoi and proto-bird-bowls such as ‘Nestor's’ cup, but in turn the first inscribed sherds from Pithekoussai are earlier than any such exports. This suggests that the alphabet was ‘passively’ received in Rhodes, as in Crete, in the years between 775 and 725.