© President and Fellows of Harvard College
Identification and Creation
Object Number
Serpentine Fibula
Other Titles
Former Title: Half Moon Fibula
Work Type
fibula, pin
mid 8th century BCE
Creation Place: Ancient & Byzantine World, Europe
Iron Age
Persistent Link
Physical Descriptions
Copper alloy
8.7 x 3.6 x 0.4 cm (3 7/16 x 1 7/16 x 3/16 in.)
Technical Details

Technical Observations: The serpentine fibulae were fabricated using a combination of wire-making techniques. Most of the wires were cast into rods and cold worked with a hammer to lengthen them. The wire was hammered to shape two springs and a catchplate. The shape of the wire among the fibulae was formed either round, square, or hexagonal in cross-section. The variation in the shape of a single length of wire resulted from the metal being hammered to create different elements. This can be seen in a spring detail of 1987.135.35, where a square-shaped wire becomes round at the springs. Also visible on 1987.135.35 are square hammer marks adjacent to the spring that give more evidence of hand shaping. The diameter of the serpentine wires has an even taper that gradually becomes thinner towards the tip of the pin.

Three of the serpentine fibulae were tested for alloy composition using EMP analysis (1). The results for all three fibulae (1987.135.40, 1987.135.48.A-B, and 1987.135.50.A-B) revealed a composition of low-tin bronze, with the tin concentration ranging from 4 to 9% (2). All of the samples contained traces of antimony and arsenic, which is a common combination found in tetrahedrite-type copper ores. Two different beam sizes were used for monitoring inclusions in the metal. All of the samples, when monitored with the larger beam raster, showed a consistent increase in the lead and sulfur content, which is a result of copper sulfide or lead globule inclusions.

All nineteen serpentine fibulae were x-radiographed, and the exposures showed a dark line running along the center axis of the wire on at least ten of them. The line can be seen in a detail of the x-radiograph of 1987.135.47. The center of the wire is the thickest section and should be the most radio-opaque (white in the x-radiograph); therefore the dark line could indicate one of three possible situations: that the fibula has a hollow center, that there is a small groove along the outside surface of the wire, or that a shrinkage occurred during casting, which is called coring. There is a visible groove in the wire adjacent to the catchplate of 1987.135.37. However, other fibulae showing a line in the x-radiograph do not have visible grooves in the surface of the wire. This raises the question of whether the wire could have also been fabricated using a strip-folding technique. This technique was common for gold wire making in antiquity; it involves folding a flat ribbon of metal into a C-shape and joining the sides to create a wire that has a seam down one side (3).

In order to gain more information about the wire and its fabrication method, a transverse and a longitudinal section of wire were sampled from the pin break of 1987.135.48.A-B. These samples were polished for metallographic analysis (4). The transverse section of wire showed a fissure across the center of the wire. Under polarized light, the fissure is black in coloration, and SEM-EDX analysis showed that it contains copper and tin. Large pits of corrosion have disrupted the circumference of the wire, making the morphology of the fissure difficult to read. It is not possible to determine from this sample if the fissure extends out to one edge, which would indicate a strip-folding technique of manufacture (5). The fissure in this case is most likely coring, which could also be an explanation for a dark line seen in the x-radiographs.

The longitudinally mounted cross-section section of 1987.135.48.A-B shows elongated inclusions that run parallel to the axis of the wire. The drawn-out inclusions resulted from cold working the metal to lengthen the wire. The inclusions turn black under crossed polars, and analysis using SEM-EDX indicates that they are sulfide inclusions, since they contain only sulfur and copper (no tin).


1. See the ancient bronze online resource or one of the three object records for details on the EMP analysis.

2. Other alloy references of Iron Age fibulae include a brooch from Tarquinia having Sn, 12.5%; Pb, 0.3%; Cu, 86.6%; see U. Anemüller, E. Formigli, and W. R. Teegen, “Technologische und experimentelle Studien und einem früheisenzeitlichen Fibelfragment aus Tarquinia,” Gedenkschrift für Jürgen Driehaus (Mainz, 1990) 31-47, esp. 40. For Italian and Sardinian bronzes that are tin-bronze and have similar alloys and impurities as the Harvard fibulae (Sn, 4.2%-10%; As, 0.005-0.4%; Sb, 0.03-0.35%; Co, 0.003-0.05%; Ag, 0.4-0.6%; Fe, 0.04-0.4%; Ni, 0.1-0.3%), see P. Craddock, “The Metallurgy of Italic and Sardinian Bronzes,” in Italian Iron Age Artefacts in the British Museum, ed. J. Swaddling (London, 1986) 143-52. For Sicilian bronzes with sulfur and iron inclusions, indicating sulfide ores, see C. Giardino, Il Mediterraneo Occidentale fra XIV ed VIII secolo a.C.: Cerchie minerarie e metallurgiche = The West Mediterranean between the 14th and 8th centuries B.C.: Mining and Metallurgical Spheres, BAR Int. Ser. 612 (Oxford, 1995) 320.

3. For gold wire making techniques, see A. Oddy, “The Production of Gold Wire in Antiquity: Hand-Making Methods Before the Introduction of the Draw-Plate,” Gold Bulletin 10.3 (1977): 79-87; and id., “Does Strip-Drawn Wire Exist from Antiquity?” MASCA Journal 4.4 (1987): 175-77. For a study of strip-drawn wire from Narce that has a groove running parallel to the axis of the wire, see G. Warden, “Strip Drawn Wire from Iron Age Narce,” Studi Etruschi 46 (1978): 265-67. See also E. Formigli, “Modi di fabbricazione di filo metallico nell’oreficeria etrusca,” Studi Etruschi 47 (1979) 281-92, who makes the point that a variety of techniques could have been used in wire making. Formigli also claims that it is possible wire drawing could have been done during the Iron Age using drawing tools that have not been recognized as such.

4. See the ancient bronze online resource or 1987.135.48.A-B for details on the metallographic analysis.

5. See photomicrographs of transverse cross sections taken from folded wire in H. Hoffmann and A. E. Raubitschek, Early Cretan Armorers (Mainz, 1972) fig. 2; and Oddy 1987 (supra 3) pl. 3.a. A transverse cross-section of folded wire shows a hollow center with a visible seam on one side where the ribbon edges have joined.

Julie Wolfe

Acquisition and Rights
Credit Line
Harvard Art Museums/Arthur M. Sackler Museum, Gift of Dr. and Mrs. Jerry Nagler
Accession Year
Object Number
Asian and Mediterranean Art
The Harvard Art Museums encourage the use of images found on this website for personal, noncommercial use, including educational and scholarly purposes. To request a higher resolution file of this image, please submit an online request.

Published Catalogue Text: Ancient Mediterranean and Near Eastern Bronzes at the Harvard Art Museums
The wire is round in section and becomes slightly rectangular at the springs. The springs have a fine dimple mark that may be from hammering. The surface has green corrosion and thick burial accretions overall.

The serpentine fibula has an elegant form that consists of a single length of wire formed into a crescent-shaped fastener for clothing. One end of the wire has been sharpened into a point to puncture the cloth, while the opposite end has been shaped into a catchplate to hold the tip of the pin. Two single springs incorporated into the wire apply tension between the catchplate and pin. The section between the springs, called the bow, curves in harmony with the arch of the pin. There are nineteen serpentine fibulae in the Harvard Art Museums’ collection, and these fibulae have minor stylistic variations in form and decoration. They have no known provenience; however, serpentine fibulae were common during the Iron Age, and they are found distributed throughout Italy, Sicily, Sardinia, and France (1).

The serpentine type of fibula is considered to have descended from the Late Bronze Age violin-bow type, which has a straight pin and a single spring. Replacing the violin-bow, the serpentine fibulae came into general use during the Early Iron Age, and the type continued to be used throughout the next half century. During this time, communities created their own types of fibulae, and discrete variations in the form developed. Trade in Italy during the eighth century BCE contributed to the spread of provincial styles; as a result, the traceability of their origins has been complicated. Serpentine fibulae were common in Sicily during Greek colonization, hence the type has also been termed “Sicilian” (2). Changes in the form have been used to date serpentine fibulae, the earliest of which have a flattened spiral catchplate. As the Iron Age progressed, the catchplate developed into a simple, folded channel, which became longer over time (3). 1987.135.53 and 1987.135.40 have the longest catchplates in the collection and probably date to the Late Iron Age, while the remaining fibulae with shorter catchplates are most likely from the Early Iron Age. Some of the fibulae in the Harvard collection have incised decorations, including 1987.135.38, which bears a herringbone pattern partially obscured by corrosion (4). Inscribed, consecutive parallel line patterns, similar to the ones on 1987.135.51, have been found on serpentine fibulae in central Italy (5).


1. Compare serpentine fibulae in J. Sundwall, Die älteren italischen Fibeln (Berlin, 1943) DII-ßb (Apulia, Bologna, and Tarquinia); C. Giardino, Il Mediterraneo Occidentale fra XIV ed VIII secolo a.C.: Cerchie minerarie e metallurgiche = The West Mediterranean between the 14th and 8th centuries B.C.: Mining and Metallurgical Spheres, BAR Int. Ser. 612 (Oxford, 1995) 243, 247, 291, and 330 (Sicily, Sardinia, and France); J. de la Genière, “Torano Castello (Cosenza): Scavi nella necropoli (1965) e saggi in contrada Cozzo la Torre (1967),” Notizie degli scavi di antichità 31 (1977): 389-422, esp. 391, 393, 400, 405, 408, 412, and 414 (Torano Castello); O. C. Colburn, “Torre del Mordillo (Cosenza): Scavi negli anni 1963, 1966 e 1967,” Notizie degli scavi di antichità 31 (1977): 423-526, esp. 519 (Torre del Mordillo); R. M. A. Procelli, “Calascibetta (Enna): La necropoli di Cozzo S. Giuseppe in Contrada Realmese,” Notizie degli scavi di antichità 36 (1982): 425-632, esp. 486, 539, and 553 (Calascibetta, Sicily); G. C. Pescatori, “Cairano (Avellino): Tombe dell’età del Ferro,” Notizie degli scavi di antichità 25 (1971): 481-537, esp. 485, fig. 4 (Cairano); P. Righetti, “Veio (Isola Farnese): Ricerche sul terreno prima degli scavi della necropoli in località ‘Quattro Fontanili,’” Notizie degli scavi di antichità 30 (1976): 185-220, esp. 198, fig. 5 (Veio); and F. Lo Shiavo, “Francavilla Marittima, Necropoli di Macchiabate: Le fibule di bronzo,” Atti e memorie della Società Magna Grecia, 2.18-20 (1977-79): 93-109, esp. 95, no. 5, fig. 37 (Francavilla Marittima).

2. This type of fibula is also termed “bent-bow” by R. R. Holloway, The Archaeology of Early Rome and Latium (London, 1994) 39, fig. 3.4.

3. Holloway 1994 (supra 2) 40; Giardino 1995 (supra 1); N. Hartmann, “Society and Technology in the Villanovan Iron Industry,” in The Bronze Age-Iron Age Transition in Europe 1, eds. M. L. Stig Sorensen and R. Thomas, BAR Int. Ser. 483 (Oxford, 1989) 93-99.

4. Compare 1987.135.38 with fibulae in H. Müller-Karpe, Beiträge zu italienischen und griechischen Bronzefunden, Prähistorische Bronzefunde 20.1 (Munich, 1974) pl. 10.A.6 (Torre Galli, grave 149); with alternating ribbed and herringbone patterns in H. Henken, Tarquinia, Villanovans and Early Etruscans (Cambridge, MA, 1968) fig. 24.b (Selciatello cemetery at Tarquinia); Giardino 1995 (supra 1) 243 (Pantilica, Sicily); de la Genière 1977 (supra 1) 391 (Torano Castello, Calabria); and Sundwall 1943 (supra 1) 150, DII-ßb (Sicily, Cuma, Torre Galli, Torre Mordillo, and Vetulonia).

5. Compare 1987.135.36.A-B and 1987.135.51 with fibulae at Quattro Fontanili published in Italy Before the Romans: The Iron Age, Orientalizing and Etruscan Periods, eds. D. Ridgeway and F. R. Ridgeway (London, 1979) fig. 2; Hartmann 1989 (supra 3) fig. 11.e; and A. Guidi, La necropoli veiente dei Quattro Fontanili nel quadro della fase recente della prima età del ferro italiana, Biblioteca di “Studi etruschi” 26 (Florence, 1993) fig. 20.5. See also A. Pasqui, “Scavi della necropoli di Torre Mordillo nel comune di Spezzano Albanese,” Notizie degli scavi di antichità (1888): 462-80, esp. 465, fig. 3, pl. 19; and A. M. Bietti Sestieri, The Iron Age Community of Osteria dell’Osa (Cambridge, 1992) 97.

Julie Wolfe

Publication History

Julie Wolfe, "Analysis of Iron Age Bronze Fibulae from Southern Italy in the Collection of the Harvard University Art Museums" (thesis (certificate in conservation), Straus Center for Conservation and Technical Studies, June 1998), Unpublished, p. 1-14 passim.

Séan Hemingway and Julie Wolfe, "Art and Technology: The Study of Ancient Bronzes at the Harvard University Art Museums into the 21st Century", Proceedings of the XVth International Congress of Classical Archaeology, Amsterdam, July 12-17, 1998, ed. Ronald F. Docter and Charlotte Moormann, Allard Pierson Series (Amsterdam, 1999), 196-99, p. 197-98, pl. 17.b.

Subjects and Contexts

Ancient Bronzes

Related Works

This record has been reviewed by the curatorial staff but may be incomplete. Our records are frequently revised and enhanced. For more information please contact the Division of Asian and Mediterranean Art at am_asianmediterranean@harvard.edu