Most of this article was written agood while ago. My thanks are due to Mrs. I. Marconi Bovio, Miss Ritcher, Dr. L. D. Caskey, Mr. E. T. Leeds, Mr. F. N. Pryce, Prof. W. Unverzagt and Prof. Robert Zahn for permission to figure lamps in Palermo, New York, Boston, London and Berlin. Prof. Buschor told me of a lamp from his excavations in Samos and generously invited me to publish it. Dr. Jacobsthal kindly read my manuscript, and I have profited by his criticisms. I owe a special debt to Miss Emilie Haspels, who not only took the excellent photographs reproduced in pl. VI and figs. 2, 3, 9, 10 and 14, but also gave me careful notes on the lamps in Sicilian museums: I have been able to check her observations since, and need not say that I found them everywhere accurate. The print used for fig. 17 was given me by Humfry Payne. Another Acropolis lamp (fig. 18) was photographed for me by Mr. H. Wagner. Fig. 1 is due to the kindness of Prof. Ashmole. Pl. V and figs. 15–16 are from photographs by my wife, whom I consult on lamps as on all other archaeological matters.

¹ On the early history of the lamp:—

Walters, Cat. of Lamps in the British Museum;

Burrows, and Ure, in JHS 31, pp. 72–99CrossRefGoogle Scholar;

Pfuhl, in Jb 27, pp. 52–7Google Scholar.

Waldhauer, Die antiken Lampen der Eremitage;

Köster, in Amtl. Berichte 31, pp. 313–14Google Scholar;

Broneer, , Corinth, iv. 2: Terracotta LampsGoogle Scholar;

Messerschmidt, in Anz. 1933, p. 327Google Scholar;

Evans, Palace of Minos (index);

Robins, The Story of the Lamp;

Robins, , ‘The Lamps of Ancient Egypt,’ JEA 25, pp. 184–87Google Scholar.

The chief use of the so-called ‘kothon’ is supposed to have been as a lamp or ‘night-light’ (Burrows and Ure, loc. cit.; Pfuhl, loc. cit.; Ure, in Eph. arch. 1937, pp. 258–62Google Scholar). The Greek ‘kothon,’ like the lamp, begins in the seventh century, and would be an alternative solution of the lighting problem.

² Pfuhl, in Jb. 27, p. 55Google Scholar; repeated by Messerschmidt, in Anz. 1933, p. 327Google Scholar.

On this argument the following things would all be ‘new’ in Homeric times: balances (Θ 69, Χ 209), bonds (Ο 20), brooches (Ε 425), chains (Θ 19), chairs (Θ 436), cups (Δ 3), floors (Δ 2), lashes (Θ 44), sandals (ω 341, δ 97), shackles (Ν 36), thrones (Θ 442), wands (π 172, ω 2), and yokes (Ε 731).

³ These are to be published by Mr. Richard Howland, who kindly showed me the series and overcame my scepticism about the existence of seventh-century clay lamps, which I had often read of, but never seen.

⁴ I use the term in the same limited sense as most writers, for example, Payne, and Jenkins in his Daedalica: writers on Sicilian antiquities are apt to give it a wider significance. There will naturally be hesitation sometimes whether to call a particular piece latest daedalic, sub-daedalic or post-daedalic; earliest daedalic, proto-daedalic or pre-daedalic.

⁵ Part of the blame is due to the word Etagenperücke, which from its hideousness has a beastly fascination for archaeologists.

⁶ Poulsen, , Orient, p. 148, fig. 173Google Scholar; Dohan, in Metr. Mus. St. 3, p. 224, fig. 33Google Scholar.

⁷ BCH 1937, pl. 26, 3 (Karouzou).

⁸ A single cord, showing on the forehead below the line of the hair, but passing under the long back-hair and concealed by it, may be seen in Eastern Greek statuettes from the earlier part of the sixth century in London, B 438 (Pryce, Cat. of Sculpture, i. pl. 39Google Scholar) and B 439 (id. p. 185), and Leipsic (Rumpf, in Ant. Plastik, pp. 219–20Google Scholar).

Another wear appears in the limestone statue from Eleutherna (Mon. Ant. 6, p. 187Google Scholar; Jenkins, , Daedalica, pl. 9, 1–2Google Scholar): here a fillet passing round the head is exposed everywhere except over the forehead; where it is concealed by the ends of the front hair; so also in a clay head from Camiros in Berlin (Tc. 7994; Knoblauch, Studien zur archaisch-griechischen Tonbildnerei, pl. 1, 2).

⁹ For the rosettes see note 8.

¹⁰ Jenkins, (Daedalica, p. 21)Google Scholar speaks of ‘a double row of curls on the Mycenae head,’ and warns us that ‘we should be wide of the mark in assuming that any form of head-band or fillet was represented’. The photographs (Svoronos, Das Athener Nationalmuseum, pl. 178; Poulsen, , Orient, p. 151, fig. 178Google Scholar; Jenkins, pl. 6, 7; Corolla Curtius, pl. 7, pl. 9, 2 and pl. 10) might seem to show two rows of curls, but in the original the difference between curls above, and headband below, is distinctly marked.

In the Cretan head-vase Berlin 307 (AM 22, pl. 6; Pfuhl M.u.Z., fig. 56; Neugebauer, Führer: Vasen, pl. 10, 1; Jenkins, pl. 6, 6) and the crude Rhodian head-vase in London (Maximova, Vases plastiques, pl. 30, 112; Jenkins, pl. 6, 3) a diadem like that of the metope is worn, but it covers the ends of the hair in front.

In the metope the heart, and the divisions between the petals, would be brought out in colour. In the vases the circles with centres are probably an abbreviation for flowers: if not, cf. the diadems Olympia, 4, pl. 18, 309 and pl. 19, 310.

Many rosetted diadems have been found, of gold or pale gold: in the more elaborate ones the rosettes are made separately and fastened to the foundation; in the less elaborate they are stamped. The rosettes fastened: Rhodes, from Camiros (Cl. Rh. 6–7, p. 55Google Scholar: found with late seventh-century vases); Reggio, from Cirò (Orsi, , Templum Apollinis Alaei, p. 87Google Scholar: fifth-fourth century); Berlin (AM 50, pl. 10: Hellenistic). Some fastened, some stamped: London 1160, from Camiros (Marshall, Cat. of Jewellery, pl. 13). Stamped: London?, from Cyprus (Murray, Exc. in Cyprus, pl. 8, 1: thirteenth century B.C.); Rhodes 13762, from Camiros (Cl. Rh. 6–7, p. 65, fig. 69Google Scholar: found with late seventh-century vases); Rhodes, from Ialysos (Cl. Rh. 8, p. 157Google Scholar; late sixth century); London 1154, from Camiros (Marshall, pl. 12); London 1157, from Camiros (Marshall, pl. 13); London 1217, from Aegina (ibid.); from Lousoi, , Jh 4, p. 185Google Scholar. Rosettes for attachment to archaic diadems, some of them very elaborate: London 1220–32 (Marshall, pl. 14); New York, from Rhodes (Alexander, , Jewelry, p. 46, fig. 99)Google Scholar.

The gold bands with stamped rosettes recently found at Delphi (BCH 1939 pl. 30, 1) are not actual diadems, they are representations, since they formed part of chryselephantine statuettes; but owing to the groups of horizontal lines between the roundels, they preserve the look of a ribbon or bandelette of stuff to which rosettes of metal have been fixed.

¹¹ Hauser, in Jh 9, pp. 75 ff.Google Scholar; Jacobsthal, in AM 57, pp. 67–73Google Scholar. I do not take account of such frontlets when worn by non-Greek women, and omit the marble heads from Priene and the Mausoleum in case, as has been suggested, they should represent foreigners.

¹² See Hauser, loc. cit.

¹³ See note 10.

¹⁴ Already quoted by Gàbrici, in Mon. Ant. 32, p. 370, note 1Google Scholar.

¹⁵ See below, pp. 47–48. Two semicircular stone lamps from Cornwall, attributed to the end of the neolithic age, and a modern Eskimo one, are figured by Robins, The Story of the Lamp, pls. 7, a, 3, 5, and 8.

¹⁶ Gàbrici dates the second megaron 580–50 (Mon. Ant. 35, p. 250Google Scholar), but the lamp itself not later than the first years of the sixth century (Mon. Ant. 32, p. 163Google Scholar).

¹⁷ I ought to have made this observation myself, but I owe it to Jacobsthal.

¹⁸ Compare the use of horizontal half-reels in bronze vessels: these often take swivel-handles, but sometimes they are imperforate, ornaments only (Furtwängler, , Olympia, 4, p. 135Google Scholar: Buschor, in FR iii. p. 266Google Scholar).

¹⁹ Ferri, Divinità ignote; cf. Soc. Magna Graecia, 1931, pl. 15, 3, from Agrigento. See also p. 46.

²⁰ Payne, , NC, pp. 22–4Google Scholar. About 650 according to Diodorus. Recent discussion of the foundation dates of the Western colonies, Byvanck, in Mnemosyne, 4, pp. 189–206Google Scholar. See also Matz, in Gnomon 13, pp. 406–7Google Scholar.

²¹ Omnes autem candido tantum marmore usi sunt e Paro insula, quem lapidem coepere lychniten appellare, quoniam ad lucernas in cuniculis caederetur, ut auctor est Varro. Plin. 36, 14.

.

Epigram 3rd C.B.C., Guérard, and Jouguet, , Un Livre d'Ecolier, p. 20, 5Google Scholar; Page, D. L., Greek Literary Papyri, i, p. 450Google Scholar.

. Kallixeinos of Rhodes in Athenaeus 5, 205 f.

. Clement of Alexandria, Protr. p. 41Google Scholar.

λυχνίας. Plato Comicus in Pollux, 7, 100 (the context shows that marble or the like is referred to and not a precious stone).

Monumentum regis Mausoli lapidibus lychnicis. Hyginus, fab. 223.

. Hesychius. Probably refers to marble.

²² Decoration of bronze tripods (Benton, in BSA 35, pp. 95–6Google Scholar); of clay pyxis-lids, CV Athens, III. H. d. pl. 1, 9.

²³ This may be partly due to influence from the Orient, where heads were a favourite motive in decoration from very early times.

²⁴ In Payne and Young, p. 1. Compare Langlotz in Schrader, p. 9.

²⁵ On pairs of ram's heads facing see Jacobsthal, in JRS 29, p. 99Google Scholar. For the birds confronted but turning their heads away he compares the Attic hydria, from Analatos, in Athens (Jb. 2, pl. 3), the Protocorinthian aryballos in Boston, (AJA 1900, pl. 6Google Scholar; Payne, Protokor. pl. 11), the East Greek ‘Daphni’ situla in Rhodes (Cl. Rh. 3 p. 195Google Scholar; CV ii. Dm. pl. 1, 5), and a mirror, from Locri, in Reggio (N. Sc. 1913, suppl., p. 40Google Scholar).

²⁶ On sirens with arms see Haspels, Att. Bf. Lekythoi, p. 158Google Scholar, note 2. For the lyre-like pair of tendrils under the griffins' forelegs compare, as Jacobsthal reminds me, the Tleson cup in Castle Ashby, ABS pl. 5, 1.

²⁷ A group of blue fayence objects, something like these, have also been taken for lamps, but this is not certain: (1–4) Alexandria 18930, from Alexandria (Breccia, Necr. di Sciatbi p. 81Google Scholar) and three others from the same site; (5–6) Cairo 18012 (von Bissing, , Fayencegefässe, p. 98, 1Google Scholar) and 18011; (7) once in the Macgregor collection, so no doubt from Egypt (Wallis, , Macgregor Coll. p. 85, no. 183Google Scholar: quote by Breccia); (8) Tunis, Musée Lavigerie, from Carthage (Delattre, , Nécr. punique voisine de Sainte-Monique, 2^e semestre des fouilles, 1898, p. 14, fig. 27Google Scholar); (9) Cagliari?, from the Punic cemetery of Predio Ibba near Avendrace, S. in Sardinia (Mon. Lincei, 21, p. 132Google Scholar, fig. 41, 2, and p. 154, fig. 59). The date is given by the occurrence in the Chatby cemetery of the late fourth and early third centuries B.C.

²⁸ 72198 (Ruesch p. 368 no. 1622): Mus. Borb. 15, pl. 22. From the Borgia collection at Velletri. It seems doubtful whether the lion-paw base figured with the lamp belongs (Burrows, and Ure, in JHS 31, p. 94Google Scholar): the inventory number is different (72324), and it is missing in De Angelis' reproduction (Tarbell, Cat. of Bronzes in the Field Museum, pl. 41, fig. 17). On the other hand the type of base is archaic, and a bronze lamp in Berlin (our fig. 30) has a very similar one. The satyr-head is perhaps to be compared with that of the infundibulum from Capua in Berlin and Copenhagen put together by MissSauer, (Anz. 1937 pp. 285–308Google Scholar: Riis, in From the Collections of Ny Carlsberg, 2, pp. 153–5Google Scholar).

²⁹ The bronze oinochoe Merlin and Drappier, p. 56 (Mus. Alaoui, suppl. i. pl. 60, 68; see also Neugebauer, in RM 38–9, p. 348Google Scholar) is earlier, and may have been in the house for some time; so may some of the scarabs.

³⁰ At least, while the vases found with the fragment are of late type, the clay pyxis accompanying the complete lamp recalls Corinthian work of the fifth century, and Attic dishes like Merlin, pl. 5, 48, appear in Camiran graves with vases of the fifth century (Clara Rhodos, 4, p. 116Google Scholar, lekythos by the Aischines painter; ibid. p. 166, late fifth-century amphoriskos) or even of the late sixth (ibid. p. 96, bf. oinochoe).

³¹ So I wrote, but see now that Paribeni, in Atene e Roma 5 (1902), p. 45Google Scholar, alludes to a golden lamp from Pompeii in the Museum of Naples.

³² Od. 7, 81.

³³ Paus. 1, 26, 6. Jacobsthal, , Orn. gr. Vasen, pp. 99 ffGoogle Scholar. It is likely enough that there had been an ever-burning lamp on the Acropolis from very early times, and that the sumptuous lamp of Kallimachos replaced a simpler one.

Dr. Pfeiffer refers me to Euphorion in Berl. Klassiker Texte 5, i. p. 58, 3Google Scholar and the passages from Nonnus quoted there by Wilamowitz.

³⁴ Inv. 10787, from Hagios Sostes: Neugebauer, Bronzegerät des Altertums, fig. 10; Neuburger, , Technical Arts of the Ancients, p. 238, fig. 308Google Scholar. For the general shape cf. the clay lamp from Olynthos, AJA 1939, p. 75, fig. 40 above.

³⁵ See Hanfmann, , Altetruskische Plastik, i. pp. 38–9Google Scholar.