Introduction
IV. Metal Lamps

Very few lamps of precious metal have been preserved from antiquity. Many were undoubtedly melted down in later eras. But we can assume that such items were seen as rarities even then, judging from their limited mentions by ancient authors. Most famous is the golden lamp burning continually in the Temple of Athena Polias in Athens, a work by the sculptor Callimachus (Pausanias 1.26.6; De Caro 1998, pp. 240–41; Stupperich 2013). Literary fiction connects a golden lamp, for instance, with Athena (Odyssey 19.34), and a golden boat lamp with an Isis procession (Apuleius Metamorphoses 11.10), and there is mention of gilded palatial lamps (Statius Thebais 1.521).

The reality is equally meager: an exceptional double-nozzled first-century A.D. lamp from Pompeii (De Caro 1998); a gilded bronze lamp from Maserà, near Domodossola (Piedmont) (Notizie degli scavi 1894); and a now-lost crystal-and-gold lamp from the tomb of the wife of Honorius, Empress Maria (died A.D. 407), which was found in Old St. Peter’s Basilica in the Vatican, are tantalizing echoes. (For older refs., see Daremberg and Saglio 1877–1919, 3:2, pp. 1321–22; and RE 13.2, col. 1569.)

Among extant examples, the gold lamp cat. 612 is therefore a true rarity, as it is also for its exquisite workmanship and unusual combination of gold and glass. The small size and lack of burn marks make its actual use for lighting unlikely; it would have made a valued present, or a votive, a role one imagines for most precious-metal lamps. The impression of scarcity produced by the remains must, however, be weighed against Pliny’s description of the widespread and extraordinary use of both gold and silver from the Late Republican period on (e.g., Naturalis Historia 33.42–63 passim).

Although silver lamps may have been more common than gold ones, they are likewise rarely recorded in ancient literature, perhaps less worthy of commentary than gold. The Greek Anthology, for instance, mentions one (5.5); and one is dedicated to divinities in Ostia (De Caro 1998, p. 241). But archaeological records inform us of several once-known but now-lost such lamps (e.g., Bosio 1632–34, 448; and the Esquiline treasure: a small one-nozzled lamp and a candelabrum, see Shelton 1981, pp. 23 and 94).

The actual remains are not many: a figured silver lamp in the British Museum (Bailey BM IV, Q 3567, p. 14, pl. 9) along with two silver lampstands (Bailey BM IV, Q 3917, p. 103, pls. 132–33, and Q 3924, p. 105, pl. 136); a silver-incrusted bronze lamp from a provincial tomb in Tongeren (Belgium) (RE 13.2, col. 1569); two miniature lamps from a young girl’s grave found in the area north of Rome and now in the Antikenmuseum in Berlin (Platz 1978, p. 187, nos. 262–63, with ill.); a first-century lamp of Loeschcke type XIX from Hungary (Valenza Mele 1981, p. 49 and n. 47); a candelabrum from the Kaiseraugst treasure (Baratte 1984); and a silver-plated lamp lid (Bailey BM IV, Q 3614, p. 24).

If we assume votive use for most precious-metal lamps, it is perhaps not surprising that there are almost no lamps in the numerous large Late Antique silver treasures; seen against the overwhelming amount of silver household goods recovered, this lack points to the limited popularity and suitability of silver for lamps. (An excellent listing of locations and content of such treasures of the fourth and early fifth centuries A.D. is in Guggisberg 2003, pp. 333–46; equally lacking in lamps is the extensive Stadler 2006.)

The rarity and lack of context makes it difficult to date the unusual Getty lamps. The single Getty silver lamp, cat. 613, like the gold lamp cat. 612, is small, delicate, and damaged. No parallel has been found for either one.

Metal—bronze in particular—came to play a central role in Roman economy and social culture, especially from the Late Republican period on. But in spite of the large amount of extant ancient bronze lamps, their date is often uncertain, even when their context is known. The reason for this is the obvious longevity of the individual products. Although less costly than the items of precious metal discussed above, they were certainly not the property of every household. Their value is indicated, for instance, by a fortuitous find in Pompeii: a fairly simple one-nozzled bronze lamp found in an intact strongbox in a modest private house (see Conticello de’ Spagnolis 1987). We also know of richer houses where up to six bronze lamps have been recovered.

The formal uniqueness among bronze lamps is due primarily to their method of production, mostly cire perdue, which created baffling varieties, at times making a close typology difficult. With rare exceptional examples from the Bronze Age (see, e.g., Bailey BM IV, p. 5), the general time span of bronze lamps reaches from the ninth or eighth century B.C. until the Arab invasion of the Mediterranean in the seventh century A.D., after which their occurrence wanes.

The longevity goes both for individual items and for types, many of which continue into the Christian era, thus adding to the problem of dating (see some startling examples of documented survival given in Bailey BM IV, p. vii).

The relation between clay lamps and bronze lamps has long been recognized. An approximate calculation assumes roughly a 1:200 numerical ratio in the Greek period, with metal lamps increasing considerably in Roman times (see Forbes 1966, p. 155). No precise numerical comparison of clay or bronze lamps from, for instance, Pompeii or Herculaneum has been done so far.

More significant than the numbers is the parallel or reciprocal formal relationship between the two media. The imitation or inspiration of style is often close, especially in the transition from the first century B.C. to the first century A.D. (Valenza Mele 1981, p. 42) and constitutes a growing field of systematic inquiry, which lies beyond the scope of the present work. (See, e.g., Xanthopoulou 2005, for the Late Roman period, with further refs.). General opinion is strongly in favor of bronze lamps being the models for clay lamps, with obvious adjustments for both market demands and technical necessities (see, e.g., Pettinau 1990, pp. 88–89). The admiration for bronze lamps is seen in a number of ways: for instance, the dark glaze of Ephesus clay lamps was undoubtedly intended to imitate bronze (see cats. 44–50); occasionally metal antecedents survive in clay as decoration, without any function (see Introduction to Firmalampen, before cat. 441, or cat. 147 and cat. 148 for hinges without any lid). The suggestion that such features can help us restore lost metal lamps is well taken (Perlzweig 1961, pp. 4–5). The chronological implications of these relationships are explored especially by Valenza Mele 1981 passim.¹

In spite of the chronological problems indicated, it is possible to give a rough developmental outline of the bronze items, although generally not with the same categorical refinements into types as for the vastly more numerous clay lamps. The material from the Vesuvian region lends itself best to classifications. (For some attempts at developmental overviews, see, e.g., Menzel 1969, p. 106; and Bailey BM IV, pp. 5–7 and 12–13, with introductions to subsequent chapters. For more narrowly focused treatments, see refs. in typological sections below.) The fundamental, classic Roman lamp typologies were established by Loeschcke 1919 on the basis of the restricted dates of the Roman camp at Vindonissa during the first and early second centuries A.D. Loeschcke’s typologies may generally be more suitable to clay lamps than to bronzes (see Leibundgut 1977, pp. 60–64, pls. 18–20); however, because these classifications are used by some scholars also for bronzes, they will be indicated here, when deemed relevant.

Just as the dates of bronze lamps are marked by uncertainty, so often is the place of production. It can be assumed that Early Imperial lamps were mostly produced in Italy, as were the lamps from the second to fourth century A.D. The Late Roman products, on the other hand, seem to derive primarily from the eastern Mediterranean (see Bailey BM IV, p. vii). The types of the first and second centuries spread rapidly to other parts of the empire, as confirmed by general parallels found and by the geographical places of manufacture or origin indicated for the Getty lamps (primarily Asia Minor).

The Getty bronze lamps vary widely in chronology, from the eighth century B.C. to at least the sixth century A.D., but their distribution is not even: the majority belongs to Roman times, or, even more narrowly, to the Early Imperial period.

Cat. 614 is the oldest bronze lamp in the collection and the only Archaic item. Deriving from Sardinia, it is in the shape of a boat. Boat models, often used as lamps, are found both in clay and bronze (Göttlicher 1978 with large bronze repertory, nos. 374–437, pp. 70–76); their functions are much debated. Some are verifiable lamps, other models are assumed to be votive offerings or to have liturgical or funerary use (Göttlicher 1978, esp. pp. 4–11; Thimme 1983, pp. 43–44). If indeed cat. 614 was a lamp, the wick would have hung over the stern. It belongs to a tradition common on Sardinia, starting in the eighth century B.C. or perhaps earlier. These objects often include various animal protomes on the railings and at the prow (for a repertory of animals, see, e.g., Lilliu 1966, p. 27; Depalmas 2005); most often a handle either spans the midship or (later) has the form of a mast; both carry a top loop for suspension; the transverse handle is broken off on the Getty example. (For the export pattern of such lamps, see Lilliu 1971 and Depalmas 2012; for the metal composition of Sardinian bronzes, see Balmuth and Tykot 2002).

Classical-period bronze lamps are rare, and the Getty has none. With the Hellenistic era, on the other hand, their frequency increases, see cat. 615. The characteristics of clay lamps of this period are very evident also in bronze lamps: bulbous bodies and extended narrow nozzles (cf. clay lamps above, esp. cats. 28–32). The latter feature requires a counterweight to balance the lamp, especially metal lamps, and the handles are therefore frequently elaborately developed with backward-leaning attached shields, often in vegetal form. (In addition to the parallels quoted under cat. 615, see also Loeschcke 1919, p. 472, fig. 44.) A lid is often added, hinged or attached with a chain, plain or with figure decoration, a feature that survived for centuries. Especially in Roman Imperial times bronze lamps come into their own; their numbers increase along with proliferation of types, which are preserved particularly from Campania, and in the later phases predominantly from the eastern Mediterranean.

Because inscriptions on bronze lamps are very rare, the prime interest of the small modest lamp cat. 616 is the inscription casually incised on the base: POI retrograde, evidently a secondary act, inscribed at some time after the lamp was cast, presumably part of a name. The triangular nozzle of this lamp ultimately derives from Hellenistic lamps, especially from Asia Minor (see cats. 28–36). Goethert 1997 holds that this lamp type was concentrated in the area of the Rhine, Austria, and eastern France (p. 187). That fits with the reported findspot of cat. 615—the lower Rhine Valley.

For dating bronze lamps, the main chronological peg for the first century A.D. is, of course, the finds from the area destroyed by Vesuvius in A.D. 79, which provides a terminus ante quem. But another earlier important discovery with a reasonably focused date is the Mahdia shipwreck (see Barr-Sharrar 1994). Several large and well-preserved bronze lamps from the wreck are dated to the early first century B.C. Their particular characteristics are the double nozzles with sizeable wide curved volutes ending in a knob or merely a point (see Barr-Sharrar 1994, esp. pp. 643–51). The find has, in fact, given name to this particular shape—Mahdia type. These lamps along with other luxury items were probably on board the ship from Athens or another Greek location, destined for Italy; perhaps their Greek parentage can be seen in large double-nozzled clay lamps such as Broneer type XXI of the late first century B.C. (Broneer 1930, pls. 8 and 9).

Mahdia-type lamps are regarded as the starting point for what would develop into the most prevailing feature of Roman Imperial lamps, both in clay and bronze: the short nozzle-volute lamps with double or single knobs (corresponding to Loeschcke types I–V). Cats. 617 and 618 are first-century A.D. echoes of Mahdia-type lamps, although of modest size and plain decor. The double-nozzled cat. 617 has a large crescent over its ring handle, an iconographic staple since the Late Hellenistic period; it occurs frequently in the Vesuvian area (see Valenza Mele 1981, nos. 69, 75, 79–80, and 82; for the very few Mahdia-type lamps in the Museo Archeologico Nazionale, Naples, see ibid., pp. 29–32, nos. 43–46; and Conticello de’ Spagnolis and De Carolis 1988, no. 13, pp. 41–43). Curiously, after a considerable chronological hiatus Mahdia-type lamps reappear in the third and fourth centuries, often with elaborate added details (see Bailey BM IV, Q 3771–Q 3777, pp. 62–63, pls. 72–74; Xanthopoulou 2010, pp. 3–5). The Getty Museum has no such late example.

Dating cats. 617 and 618 is difficult in the absence of any close parallels. They rather appear to be simplified derivations/evolutions of original Mahdia lamps imported to Italy. Barr-Sharrar 1994 seems to confirm such an interpretation (pp. 648–651). This is particularly true for the rare cat. 618, the single-nozzled lamp, a form also uncommon in clay. The one single-nozzled lamp from the Mahdia wreck is, in fact, quite unlike cat. 618 (see illustration in Barr-Sharrar 1994, p. 650). On the original Mahdia lamps the handle (when present) consists of a large ring, connected with a leaf ornament. Cats. 617 and 618 maintain the ring, but alter—or omit—the decor. For theories on the possible origin of the Mahdia type, see Valenza Mele 1981, pp. 29–30, who suggests a relation to long-nozzled volute-lamps from Asia Minor.

More common than the large volutes discussed above are other first-century A.D. types, as seen on cats. 619 and 620. Instead, there are small protrusions at the inner side of the nozzle only, and the body is considerably lower than before. Both lamps balance their elongated nozzles with a proportionately substantial handle and shield. (Cat. 620 was a suspension lamp.) Both cats. 619 and 620 belong to Loeschcke type XIX, a very numerous and widespread lamp form from the mid-first century all through the second century A.D. (Valenza Mele 1981, pp. 49–50).

Cat. 621 is more unusual: a curious combination of features, each separately familiar. The flat circular discus, enclosed by a low flat rim created by the raised side-walls, refers to Firmalampen, as does the channel from discus to nozzle with a small vent to carry oil overflow; and the curved handle ending in an animal head is a feature common from the first century on. The sum total of cat. 621 is original and striking; highlighted by the finely carved leaf collar, it is a lamp at the edge of a large class of bronze lamps influenced by the Firmalampen, popular during the first and second centuries (for clay Firmalampen, see cats. 442–48 and 449–50).

The curved neck handle ending in a variety of animal heads (horse, bird, deer, griffin), or, occasionally, a human head, was common at Pompeii and elsewhere and is found combined also with numerous examples of the elegant pear-shaped lamps from the first century A.D. (the Getty Museum has none of these, but see, e.g., Comstock and Vermeule 1971, pp. 348–49, no. 498 [from Egypt]; Valenza Mele 1981, nos. 261–80; Conticello de’ Spagnolis and De Carolis 1988, nos. 83–106; and Chrzanovski 2003, p. 91, nos. 109–10 [from Switzerland]. In spite of their obvious fragility, some clay lamps have this type of handle: see cat. 472, and Rosenthal and Sivan 1978, pp. 90–91, nos. 368–72).

Another first-century A.D. type lamp is preserved with its stand: cat. 622. It represents a tendency much more common in metal than clay lamps: suspension facility. In this case, the chain is attached to three loops—for maximum stability, the most frequent arrangement. Possibly the persistent tendency to hang or attach bronze lamps to stands derives from a desire to avoid the heat the metal might generate. This lamp shape still shows the need to balance the long nozzle with a large ring handle and leaf. The volutes are abandoned for small knob protrusions, here both at the outer and inner ends of the nozzle. A domed lid is still attached to its hinges. The production appears limited and is dated from the end of the first century B.C. to the early first century A.D. (Valenza Mele 1981, pp. 42–43, and nos. 72 or 79, although with different handles).

Two further sizeable suspension lamps, cats. 623 and 624, both have double-nozzles on opposite sides of a central circular discus, but otherwise they display some internal typological differences: cat. 623 (belonging to Loeschcke type XIX) has semivolutes and a well-defined circular discus, marked by a raised rim, leaving no room for a shoulder outside. It is a numerous category, with both one and two nozzles and various internal variations of detail (closest parallels: Valenza Mele 1981, pp. 77–78, nos. 192–95; Conticello de’ Spagnolis and De Carolis 1988, esp. pp. 70–71, nos. 54 and 56). It is dated from the mid-first to well into the second century A.D. (Valenza Mele 1981, p. 50). Cat. 624, on the other hand, is well known in bronze, but rarer in clay; there are no volutes, but the whole circumference of the lamp, including the nozzles, is surrounded by a slightly raised flat ledge; the interior is also flat; it is classified as either Loeschcke type XX or type XXI, two closely overlapping groups (our lamp would date to shortly before A.D. 79: Valenza Mele 1981, discussion pp. 118–20, p. 123, no. 296). Both lamps have extant suspension chains (cat. 624’s is broken) attached to loops in the form of gracious swan’s heads and necks. Such bird’s heads—variously identified as swans, geese, or ducks—are occasionally found instead of a plain loop (e.g., Bailey BM IV, Q 3654, p. 35, pl. 40; Valenza Mele 1981, pp. 123–24, nos. 296–97, Loeschcke type XXI; also seen on her group nos. 128–30 and 133–34, pp. 61–63, Loeschcke type XIX). In addition, cat. 623 retains an uninscribed metal plate (tabula ansata) as an intermediary link to the top hook (possibly intended for the owner’s name or a votive inscription). It could theoretically also have existed on cat. 624. The bottom treatment of cat. 623 is notable for the multiplicity and depth of the internal rills; when such base treatments are found on clay lamps, they clearly signal metal models (see, e.g., Perlzweig 1961, nos. 1242 and 1259, pl. 25; among the clay lamps above, see esp. cats. 155–56, clay versions of Mahdia-type two-nozzle lamps, and cat. 472, imitating a bronze lamp with tall curved neck).

The last double-nozzled lamp, cat. 625, is of special interest due to the application of elephant heads on the sides of the body. Applied decorations between the nozzles occur occasionally on multinozzled lamps (seen already on a large three-nozzled lamp from the Mahdia shipwreck: Barr-Sharrar 1994, pp. 640–41 [comic head]; several recorded in Valenza Mele 1981, e.g., p. 31, no. 43 [comic head], pp. 61–62, nos. 128–29 [bull protomes], p. 62, no. 130 [lion’s head], p. 124, nos. 297–98 [bull protomes], and p. 124, no. 299 [female mask]). The elephant motif as such is found sparingly in classical art, but more often from the time of Alexander the Great on. It is rare in connection with lamps: a few use the elephant head as the lamp body, with the trunk as functional nozzle (see Comstock and Vermeule 1971, p. 349, no. 490; and Bailey BM IV, pp. 15–16, Q 3574, pl. 14, third to second century B.C., with further refs. of early occurrences). A most spectacular example comes from the graves at Vani (first century B.C.): three elephant trunks constituting the lamp nozzles with human busts between (see Kakarava and Kvirkvelia 2008, p. 76); for a lampstand using elephant trunks as feet, see Bailey BM IV, Q 3871, p. 92, pls. 104–5 (first century A.D.). On the Getty lamp the applied motif is strictly ornamental.

Figured (plastic) lamps were popular from the Hellenistic period on; they are found widely in the Mediterranean area, possibly with a preponderance in Egypt, which emerges as a major production place, along with Italy. Relatively common in clay (see cats. 586–609), rarer in bronze, their major popularity is considered to fall in the first and second centuries A.D. Subjects cover a broad iconographic range, both anthropomorphic and zoomorphic; clay and bronze lamps follow very similar trends of taste. The extensive figured-lamp collection in Cologne, published in Möhring 1989, gives an excellent overview of the variations and occurrences, although covering only clay lamps. Möhring, along with others, stresses the head lamp as a favorite with wide representation: especially for ethnic varieties (Africans being the most prominent), mythological figures (esp. from the Bacchic circle), and animals such as bulls (see Bailey BM IV, Q 3574–Q 3585, pls. 14–16, illustrating trends in taste, both humorous and at times slightly grotesque).

The Getty collection contains three figured bronze lamps, two from the Early Imperial period: cats. 626–27 (cat. 630 belongs in Late Antiquity). Cat. 626 is a fine head of a comic theater mask. The open mouth serves as the filling-hole, surrounded by a carefully striated beard; expressive features are deeply modeled, beneath a hairdo covered with a wreath of leaves and berries, commonly connected with comedy heads and the Bacchic sphere in general. (Cf. the elaborate garland encircling the large Mahdia-wreck lamp mentioned above.) The elegant curved handle, now detached, has similar vegetal decor. A head lamp in the Naples Museum is iconographically close, but functionally a variant: in the large mouth is a smaller wick-hole, with the filling-hole in a large flower covering the substantial ring handle; a wreath of vines surrounds the head; there are suspension loops (Valenza Mele 1981, pp. 155–56, no. 367, the only preserved comic theater head lamp from the Naples Museum). Two closely related head lamps found in Germany, however, confirm the popularity of the form (Menzel 1966, p. 83, no. 200, pl. 63; Menzel 1986, p. 104, no. 242, pl. 113; both identified by Goethert 1997, pp. 187–88, as Silenus heads).

Cat. 627 belongs to a more unusual category, where the whole lamp takes on a vegetal form, possibly to connect with the Roman taste for decor inspired by nature. It brings to mind a well-known passage in Pliny on a most elaborate example of such a fixture: lamps suspended in imitation of apples on a tree, of Hellenistic Greek background, transferred to the Temple of Apollo on the Palatine Hill in Rome (Naturalis Historia 34.8.14). In cat. 627, the lamp body is a pinecone (torch[?]) at the end of a striated stem. Although unusual, some parallels are known: Valenza Mele 1981, p. 152, quotes examples both from Pompeii, Aquileia, and Ephesus. Two related lamps consisting of acanthus buds for the body come from the Veneto region (now in Vienna, see Beschi 1962, pp. 102–4), thus making Italy the likely origin of the type.

The three lamps from Late Antiquity represent very different trends. Common denominators of the period include bulbous body forms and large circular wick-holes in the shape of sunken cups. Also frequent is a cavity in the bottom for fitting the lamp to a spiked support. In spite of its small size, cat. 628 is typical in those respects. Expressly Christian symbols, such as crosses, Chi-Rho monograms, or laurel wreaths on the handle shield are common accompaniments, but none of the Getty's bronze lamps carries such decors. Instead, cat. 628 has a strongly curved, backward-leaning handle ending in the expressive head of a deer(?) (cf. cat. 621, of the first century A.D.). It is often pointed out that the lamps of this late period revive many characteristics of the Late Hellenistic period. Both the heavy backward-leaning handle and the prominent wick-hole of cat. 628 fit that tendency, although the long Hellenistic nozzle is not present here, as on many late lamps (e.g., Hayes 1984, pp. 139–42, nos. 216–19; Xanthopoulou 2010, pp. 6–7).

Cat. 629 does not have informative parallels, but its general shape and detailed incised decor place it in the Byzantine tradition; see, for instance, Demirel Gökalp 2005, who for her globular lamp I.4, fig. 9, mentions two related glass lamps in Istanbul (p. 70).

The entire cat. 630 lamp can be said to be conceived as a Christian symbol. The popularity of the image of the peacock has given rise to a great amount of such lamps with only minor variations (see Lother 1929). It has even been suggested that the flame and smoke from the wick-hole might have given the appearance of the striking tail of a peacock. Egypt seems to have been an especially active production center and possibly the origin of the Getty lamp. Like many of its kind, its base was made to fit onto the spike of a stand.

The Getty Museum has one open lead lamp, cat. 631, reportedly found in Cologne, Germany. The northern Roman provinces—Germany, Switzerland, and Britain—were supplied with lamps primarily through the military camps, which have left a number of simple metal lamps of lesser quality, such as iron or lead, while bronze lamps are fairly limited north of the Alps, where tallow would have been the prime available fuel. These lamps are open, have mostly a flat bottom with sides often low and straight, and are shaped as a vague oval or figure eight. The triangular form of cat. 631 is, however, coarser than most examples published (see Loeschcke 1919, pl. 22; Leibundgut 1977, p. 302, no. 1043, pl. 21 [lead, from Vindonissa]; Goethert 1997, p. 189, inv. no. 27.244 [lead, from Trier]). Open clay lamps of parallel simple forms are also found in numbers from these areas (Loeschcke 1919, pl. 20). Handles, if extant, vary between a vertical form for hanging (which probably was the case on cat. 631) and a horizontal variant (see Bailey BM IV, Q 3754–Q 3764, pp. 57–58, pls. 65–68). Britain has provided the best parallels for cat. 631, along with a fairly well-documented date range, mostly from late first through the second century A.D.