Egyptian Red Gold

ARCHEOMATERIALS 4: 133-152 (1990) Egyptian Red Gold JAMES H. FRANTZ Conservator in Charge, Department of Objects Conservation, The Metropolitan Museum of Art, New York, New York 10028 DEBORAH SCHORSCH Assistant Conservator, Department of Objects Conservation, The Metropolitan Museum of Art, New York, New York 10028 ABSTRACT Red surface colorations frequently found on ancient Egyptian gold objects have been examined by X-ray diffraction, energy-dispersive X-ray spectroscopy, and other methods. The results indicate that these colorations are most often associated with tarnish films in which the predominant species is a silver gold sulfide, AgAuS, occurring on a substrate silver gold alloy. Another silver gold sulfide, Ag3AuS2, may also be associated with these alloys and especially with those of higher silver content. These results have been corroborated by the synthesis of visually similar tarnishes on gold silver alloy coupons that have yielded analytical results identical to those obtained for the archaeological material. INTRODUCTION The surface of ancient Egyptian goldwork frequently has been found to have a distinctive red or sometimes purple coloration apart from any intrinsic color of the alloy itself (Figs. 1, 2). In the early years of this century, Egyptologists and other interested scholars attributed this phenomenon to a variety of causes, including the presence of organic residues, copper corrosion products, colloidal modifications of the gold itself, and tarnish films (Berthelot 1901, 1902; Williams 1924: nos. 23, 25, 31, 33; Lucas 1926: 90). While the color variations observed on gold surfaces are sometimes described in these accounts, the methods of analysis either have been limited to microchemical tests or have not been mentioned. The discovery of the tomb of Tutankhamun in 1922 aroused interest in the topic, and new suggestions for possible causes of the colorations were made. This burial contained a profusion of gold objects, notable both for their quantity and for the quality of their workmanship (Carter 1923—33). Many of these objects were found with red surface films. In his appendix to the second volume of Carter's The Tomb of Tut-Ankh-Amen, Lucas wrote (1927: 172—173): One very noticeable feature of the gold was the varied color it presented, sometimes in patches and sometimes over the whole surface. The colours comprised bright yellow, dull yellow, grey © Archeomaterials, Vol. 4, No. 2 (Summer 1990) and red of various shades, including reddish brown, light brick colour, blood colour, dull purple (plum colour), and a very remarkable rose colour, all except the last named manifestly being fortuitous…1 In 1934, R. W. Wood reported on the examination of a group of small, red gold sequins, which he claimed had been found in the tomb arranged in alternating array with gold bars of a yellow, metallic color (Wood 1934: 62—65).2 Wood determined by emission spectroscopy that the red gold sequins 1 Lucas offers explanations and textual parallels for the fortuitous variations: "The bright yellow gold is evidently quite pure and doubtless corresponds to the `fine gold' referred to in the ancient records. The dull and tarnished yellow gold contains small proportions of other metals, such as silver and copper, which on the surface have undergone chemical changes and thus caused tarnishing. The gold that has become gray contains a large proportion of silver that on the surface has become converted into chloride, which has then darkened in the manner usual with this compound; it also contains a little copper and a trace of iron. Such gold corresponds to the naturally-occurring alloy of gold and silver—termed electron by the Greeks and electrum by the Romans—that was largely employed by the ancient Egyptians.... The reddish-brown colour gives tests for iron, silver and copper, and is evidently due to iron and copper in the gold that have oxidized. In some instances a red or purple colour proved to be a staining of the gold by organic matter, since it was not soluble in either acids or in organic solvents, but could readily be removed by heating" (1927: 173–174). What we consider to be a more probable explanation for this behavior is offered in this article. 2 Carter does not mention these sequins in his publication of the tomb, nor does Wood indicate his source for the claim that they were found in a specific pattern with the yellow gold bars. As noted above, Lucas had earlier postulated that some occurrences of red gold in the tomb were intentional and that the color in these cases might be due to the presence of iron. He provides only a general reference, however: "That this colour is intentional is shown by its regular and systematic distribution on certain objects, or on certain parts of objects" (1927: 174). A color photo of four of the red sequins described by Wood appeared recently in Ogden (1982: pl. 2). Fig. 1. Mummy of Ukhhotpe, detail with gilded wooden mask (MMA 12.182.132C) J. FRANTZ AND D. SCHORSCH Fig. 2. Coffin of Nephthys, detail of gilded face, ca. half life-size (MMA 11.150.15B) 135 136 EGYPTIAN RED GOLD contained iron, whereas the yellow gold bars did not. By adding less than one percent iron to small amounts of molten gold, Wood produced sequins that he claimed resembled in their surface appearance the red gold sequins from the tomb and on this basis he suggested that the color resulted from the intentional addition of iron salts to gold that had then been heated (1934: 63-64).3 Lucas included these findings in his second edition of 1934, and since then the theory has gained considerable acceptance in discussions of ancient Egyptian gold technology (Lucas 1932: 108; 1934: 266– 267; Plenderleith and Werner 1971: 215— 216). We will have more to say concerning these findings and hypotheses after first discussing the results of our own investigation of red gold surfaces found in the Egyptian collection of the Metropolitan Museum of Art. There are a considerable number of gold objects in the collection, dating from all periods of Egyptian history, that display a variety of red surfaces. Many of these are found on foil or leaf fragments that offer extensive opportunities for a variety of analytical methods. In this study, typical examples of these surfaces were first examined by X-ray diffraction (XRD) and energy dispersive X-ray spectroscopy (EDS). Other techniques, including optical and scanning electron microscopy, were also used to assist in characterizing some of the surfaces encountered. On the basis of the results obtained from these analyses, an effort was made to synthesize red gold surfaces that would compare favorably in appearance and chemical nature with those found on the archaeological objects. EXPERIMENTAL (I) A complete listing and description of the archaeological material examined in this study are given in the Appendix. Figures 1– 4 illustrate some of the colorations observed on this material, which range from red to purple to nearly black in certain in3 Experiments with similar results were conducted at the British Museum Research Laboratory using gold-silver-copper alloys and iron pyrites. Pyrites were chosen as the source of iron because the researchers believed that this material may have been confused with gold by early metal workers (Plenderleith and Werner 1971: 215-216). stances. Because these colorations were generally confined to a thin layer well bonded to the substrate metal, it was often difficult to remove samples mechanically without including substantial amounts of the alloy itself. In such cases, it was sometimes possible to examine directly surfaces of leaf or foil. All the surfaces examined were first tested for solubility of the coloration in a series of organic solvents that included water, acetone, benzene, hexanes, methylene chloride, methanol, and dimethylformamide. Because the surface colorations appeared to be unaffected in virtually all cases by these solvents, while at the same time adventitious dirt and adsorbed hydrocarbons were detected, it was decided at the outset to clean all the surfaces judiciously with acetone and hexanes before further methods of analysis were applied. X-ray Diffraction (XRD) Fragments of leaf or foil measuring approximately 2 cm. by 1 cm. were cleaned with acetone and hexanes and mounted on glass petrographic slides with double-faced tape. The slides were then placed in the sample holder of a Philips X-ray diffractometer, and diffraction patterns were obtained over a 2-theta range from 2° to approximately 80° using copper radiation and a focusing monochromator. In a few cases the coloration occurred as an accretion of sufficient thickness such that small samples could be removed for analysis in a DebyeScherrer powder camera. Energy Dispersive X-ray Spectroscopy (EDS) The surfaces and cross sections of the samples examined with X-ray diffraction were analyzed for their elemental compositions using an Amray 1600T scanning electron microscope coupled to a Kevex Delta IV energy-dispersive X-ray spectrometer. The surfaces were scanned over an area of approximately 0.004 cm2. Each sample was analyzed in the same spot at operating voltages of 30, 20, and 10 kV, in order to establish the depth profile of elemental compositions. These compositions were obtained with MAGIC IV ZAF cor- J. FRANTZ AND D. SCHORSCH 137 Fig. 3. Gold cylinder beads from Hatshepsut Temple Foundation Deposits, avg. length ca. 1.2 cm. (MMA 27.3.444) Fig. 4. Gold leaf fragments from LNP 600-B, length of ear fragment (bottom) ca. 6.0 cm. rections. Gaussian deconvolution was used where correction for the overlap of sulfur K and gold M peaks was necessary. The cross sections were analyzed at 30 kV and quantified without Gaussian deconvolution. Optical and Electron Microscopy Six samples were prepared in cross section for metallographic examination. The samples were mounted in epoxy, cut using a diamond wafering-blade saw, and pol- ished with a graded sequence of diamond abrasives. The sections were examined and photographed. They were then carbon coated and analyzed for elemental composition using the energy-dispersive spectrometer as described above. Other Methods Several surfaces were also examined by infrared spectroscopy. These samples were placed on the specular reflectance aperture of a Perkin-Elmer 283B infrared spectro- 138 EGYPTIAN RED GOLD photometer in order to obtain the spectra of any surface films present. A variety of spot tests were conducted on surface films that were so thin as to preclude the removal of a sample and on objects that could not be mounted in the diffractometer because of their shape or size. EXPERIMENTAL (II) Gold-silver alloys of selected compositions between 7 and 20 weight percent silver were prepared by heating weighed portions of 0.999 silver and 0.9999 gold in a porcelain crucible with an acetylene torch. Each of the resulting alloys was allowed to solidify in air and was remelted two more times to promote mixing. After the third melting the metal was quenched in water and rolled into sheet approximately 1 mm. thick with a rolling mill; annealing was carried out as necessary to avoid embrittlement during the rolling process. Coupons of the prepared alloys measuring approximately 2 cm. x 2.5 cm. x 0.1 cm. were heated in covered alumina crucibles with sublimed sulfur in a Fisher Isotemp furnace at selected times and temperatures. At the end of a heating period, the crucibles were removed from the furnace and allowed to cool to room temperature. After cooling, any free sulfur that remained as a crust on the coupons was removed either mechanically or with carbon disulfide. The surface films created on these samples were examined using X-ray diffraction and, for some of them, energy-dispersive spectroscopy. Cross sections of five samples were also prepared as described above to allow for visual examination and elemental analysis. In a separate experiment, coupons of the prepared alloys were placed in a sealed dessicator together with a beaker of 20 percent ammonium sulfide solution and allowed to stand undisturbed for two months. At the end of this period, the coupons were examined for the presence of any tarnish films. RESULTS EDS analysis of the surfaces of archaeological samples with red, purple, and sometimes black films indicated that the principal elemental components are, in most cases, gold, silver, and sulfur (Table 1). As can be seen by comparison of the results obtained at different applied voltages, the proportion of sulfur detected consistently diminishes with increasing depth of penetration through the surface film toward the metal substrate. To a lesser extent, the proportion of silver also exhibits this behavior. In all cases copper is present. A substantial amount of chlorine was detected on the surfaces of the Hatshepsut cylinder bead (MMA 27.3.444) and sheet gold from the tomb of Senebtisi at Lisht (LNP 763, Box 1). The red surfaces of the Lisht fragment (LNP 600-B) and of the Nephthys sheet gold sample (MMA 11.150.15B) yielded essentially identical diffractometer patterns that could be resolved into two phases: the gold alloy substrate and a silver gold sulfide with the formula AgAuS (Table 2). Four separate powder camera patterns obtained from surface scrapings removed from LNP 600-B, Nephthys, Ukhhotpe (MMA 12.182.132C), and a red gold Hatshepsut cylinder bead gave comparable results, with clear patterns being recorded in each case for the substrate alloy as well as for the AgAuS phase. Several examples of these patterns are listed in Table 2. The black surfaces of four silver gold cowrie shells (MMA 30.8.382—385) yielded diffraction patterns that suggested the mineral uytenbogaardtite (Ag3AuS2), rather than acanthite (Ag2S), as the predominant phase. Fragments of a Twelfth-Dynasty gold bracelet (MMA 30.8.386—393, 394A) also bore black surfaces and evidenced a mixture of Ag3AuS2 and AgAuS in film patterns obtained from scrapings. The gold leaf from the tomb of Senebtisi at Lisht (LNP 763) consists of many fragments that had previously been separated into two boxes on the basis of physical appearance. The fragments from one set (Box 2) have surfaces that are similar in color to most other samples of Egyptian red gold leaf; these pieces generated typical AgAuS diffraction patterns (Table 2). The fragments of the second set (Box 1) appear darker and grayer in color, and the metal itself was found to contain more silver (see Table 1). Diffractometer patterns obtained 139 J. FRANTZ AND D. SCHORSCH TABLE 1. A. Elemental analysis of surfaces of archaeological samples Sample: Nephthys Ukhhotpe* Hatshepsut Lisht 28.3 26.9 7.8 10.3 26.6 (Cl) 99.9% 54.5 42.2 nd 3.3 Senebtisi (Box 1) Senebtisi (Box 2) Atomic percent at 30 kV Au Ag S Cu Other Total 68.5 25.8 2.9 2.9 100.1% 48.7 45.6 4.6 1.1 100.0% 100.0% 18.7 68.8 nd 5.1 7.5 (Cl) 100.0% 85.4 14.6 nd nd 100.0% Atomic percent at 20 kV Au Ag S Cu Other Total 62.9 29.0 8.3 nd 100.2% 46.2 44.3 9.5 nd 100.0% 20.4 28.1 11.0 10.9 29.5 (Cl) 99.9% 38.1 40.9 18.1 2.8 13.0 32.7 12.8 10.4 31.1 (Cl) 100.0% 41.9 39.1 19.0 nd 99.9% 18.7 67.2 nd 5.4 8.7 (Cl) 100.0% 79.3 17.9 2.8 nd 100.0% Atomic percent at 10 kV Au Ag S Cu Other Total 40.2 39.4 20.6 nd 100.2% 28.8 47.1 24.1 nd 100.0% 100.0% 14.4 70.5 2.1 nd 13.0 (Cl) 100.0% 42.2 32.8 25.0 nd 100.0% B. Elemental analysis of cross-sections of archaeological samples Sample: Nephthys Ukhhotpe* Hatshepsut Atomic (and weight percent)1 [with standard deviation] Au 65.3 [1.1] 47.2 [1.6] 35.8 [0.1] (78.6) [0.7] (62.8) [0.9] (51.5) [0.1] Ag 28.8 [0.9] 48.5 [0.9] 58.2 [0.2] (19.1) [0.7] (35.3) [0.6] (45.8) [0.1] Cu 6.1 [0.1] 4.3 [1.6] 6.0 [0.2] (2.4) [0.1] (1.9) [0.6] (2.8) [0.1] Total 100.2% 100.0% 100.0% (100.1%) (100.0%) (100.1%) Lisht Senebtisi (Box 1) 54.4 [0.9] (69.0) [0.9] 43.3 [0.9] (30.1) [0.8] 2.3 [0.3] (0.9) [0.1] 100.0% (100.0%) 18.6 [0.4] (30.2) [0.5] 73.2 [0.2] (65.4) [0.4] 8.3 [0.2] (4.4) [0.1] 100.1% (100.0%) Senebtisi (Box 2) 76.2 [1.8] (85.8) [1.1] 22.1 [1.4] (13.6) [1.0] 1.7 [0.6] (0.6) [0.2] 100.0% (100.0%) Notes 1. Based on average of three scans, except * (Ukhhotpe), which is based on the average of six scans. nd = known to be present, not detectable under present analytical conditions. from surfaces of these pieces also showed the presence of AgAuS, as well as another unidentified phase with three strong lines at 8.97 Å, 4.50 Å, and 2.99 Å. Surfaces and samples that were characterized as silver gold sulfides by EDS and X-ray diffraction also gave strong positive reactions for sulfide ions in spot tests with sodium azide-iodine reagent (Feigl, Anger, and Oesper 1972: 437-438). Viewed in cross section, the thickness and color of the red material on these objects varied considerably, both from sample to sample and within the individual sections themselves. Under bright-field illumination the sulfide appeared dark gray, while with crossed po- lars the observed color ranged from orange to bright red to dark red brown (Fig. 5). In a number of cases, the red sulfide occurred below the surface of the gold, either as the filling of what appeared to be pits in the surface of the metal or as stringers running parallel to the surface. In addition to the many red gold surfaces found to contain silver gold sulfides as the principal species, a few instances were observed in which other phenomena appeared to be responsible for the coloration. These cases include the red surface films found on an assortment of Eighteenth-Dynasty gold rosettes that belonged to the funerary treasure of three foreign queens of Thut- 140 EGYPTIAN RED GOLD TABLE 2. X-ray diffraction data for published synthetic and geological material (d-values in Angstroms listed together with relative intensities) AgAuSa synthetic JCPDS 19-1146 7.16 15 AuAg (S, Se)b petrovskaite JCPDS 38-396 7.25 Ag3AuS2c synthetic JCPDS 33-587 10 3.16 10 3.04 10 3.02 10 100 2.77 45 2.63 50 2.54 2.45 10 20 2.55 2.46 10 20 2.39 35 2.39 40 2.12 10 10 30 15 2.25 2.12 2.04 10 3.46 20 3.57 3.44 3.38 6 35 20 3.25 30 3.09 60 3.08 60 2.94 30 2.81 60 2.84 70 2.73 100 2.61 90 2.66 2.61 2.58 45 100 70 2.46 2.44 2.42 2.38 70 80 60 75 100 2.63 2.30 2.25 2.23 3.96 30 3.16 2.77 80 60 10 20 3.87 Au gold JCPDS 4-784 30 6.98 4.38 3.99 3.96 Ag2S synthetic JCPDS 14-72 2.37 30 2.32 2.30 20 10 2.24 50 2.15 2.12 50 80 2.36 100 2.04 52 40 20 10 2.01 1.99 1.93 10 50 40 20 50 40 10 1.87 10 1.87 10 1.79 1.74 30 10 1.80 1.75 30 20 1.89 1.81 1.78 1.74 1.68 20 1.70 10 1.59 1.54 10 20 1.54 20 1.51 10 1.47 40 1.47 30 2.21 45 2.09 2.08 2.07 2.05 16 45 16 16 2.00 1.94 1.92 1.90 1.87 1.82 1.80 1.73 1.72 1.69 1.61 1.58 1.55 1.54 1.51 1.48 1.47 1.46 16 4 4 14 16 4 4 12 20 6 4 10 8 8 12 10 10 14 141 J. FRANTZ AND D. SCHORSCH TABLE 2. Continued. X-ray diffraction data for selected archaeological material. G = goniometer scans; F = Debye-Sherrer camera films. Lisht (G) Nephthys (G) 7.10 21 7.09 40 3.95 13 4.14 3.95 3.73 11 25 15 3.37 13 3.15 3.01 5 16 2.77 100 2.62 2.54 40 13 3.15 3.02 2.99 2.78 2.73 2.62 2.55 2.44 17 17 25 100 40 62 21 36 offscale 2.24 offscale 32 2.25 2.24 2.20 2.14 2.11 offscale 1.87 8 1.79 1.68 13 11 offscale 42 42 23 17 23 1.95 offscale 19 1.78 1.69 1.64 28 13 15 1.54 1.49 19 13 Hatshepsut (F) 7.19 6.80 10 4 3.96 8 3.43 8 3.27 40 3.01 Senebtisi (G) Box 2 Ukhotpe (F) 11.0 50 7.15 60 7.14 6.70 4.21 3.95 35 8 6 17 3.36 9 3.35 10 8 3.15 3.01 13 17 3.15 20 3.02 20 2.77 100 2.77 100 2.63 2.56 10 4 2.63 2.54 40 13 2.63 70 2.54 10 2.38 2.34 2.25 10 35 10 offscale 2.39 30 2.34 80 2.24 2.10 2.03 1.95 1.87 4 20 50 2 1.78 1.67 4 20 1.60 20 1.46 8 44 offscale 1.94 11 1.83 1.79 1.68 1.64 1.58 1.54 1.47 9 13 9 12 11 13 35 offscale 3.95 40 2.76 100 2.24 60 2.12 2.03 1.94 1.87 20 60 10 10 1.79 1.68 1.63 1.59 1.54 30 20 20 10 10 1.47 50 1.44 50 142 EGYPTIAN RED GOLD TABLE 2. Continued AgAuSa synthetic JCPDS 19-1146 AuAg (S, Se)b petrovskaite JCPDS 38-396 1.45 Ag3AuS2c synthetic JCPDS 33-587 20 1.41 1.39 1.23 1.18 Ag2S synthetic JCPDS 14-72 20 30 20 Au gold JCPDS 4-784 1.44 32 1.23 1.18** 36 12 30 1.38 6* a Tavernier 1966: 326. b Nesterenko et al. 1984: 606. c Graf 1968: 499. * Plus five lines to 1.305 A. ** Plus four lines to .8325 A. mosis III. EDS analysis of the red material from one of these rosettes (MMA 26.8.117) showed iron to be the predominant element, while the infrared spectrum was closely comparable to that for lepidocrocite, FeO(OH). Typically, the red films on these rosettes are concentrated along tide lines or accretionary fronts and distinguish themselves in this manner from the surface films identified as silver gold sulfides. Our attempts to create red tarnish films by heating gold silver alloys with sulfur, while somewhat primitive in design, produced adherent red films on samples containing between 7.5 and 15 weight percent silver at several temperatures above the melting point of sulfur (112.8°C). The synthetic tarnishes displayed a variety of hues similar to the reds, purples, and blacks exhibited by the archaeological samples (Figs. 6, 7) found to have silver gold sulfides on their surfaces and yielded X-ray diffraction patterns that similarly could be resolved into the two principal phases consisting of substrate alloy and AgAuS (Tables 2, 3). Like the archaeological samples, the synthetic coupons of higher silver content also yielded a few diffraction lines that are possibly associated with the species, Ag3AuS2 (Tables 2, 3). In surface EDS scans of the synthetic samples, sulfur was detected at consistently lower levels than in the archaeological material (Table 4), and the cross sections of coupons bearing synthetic sulfide films showed these to be considerably thinner than sulfide films found on ancient Egyptian gold. Our efforts to produce AgAuS by exposing metal coupons to vapors of ammonium sulfide in a sealed de- sicccator at room temperature yielded barely discernible tarnishes after an exposure of two months. DISCUSSION Samples of AgAuS and Ag3AuS2 have previously been synthesized under controlled conditions in thermodynamic studies of the AgAuS system (Barton and Toulmin 1964; Barton 1980). These two substances are the only silver gold sulfides known to exist in this system. Both are monoclinic below 310°C, and both convert to simple cubic structures above this temperature. Ag3AuS2 is found in nature as the gray-black mineral uytenbogaardtite (Barton et al. 1978),4 while a selenium-containing silver gold sulfide with the formula, AgAu(S,Se), has been discovered as a mineral in Kazakhstan and has been named petrovskaite (Nesterenko et al. 1984). Although no simple gold sulfides are known, the silver sulfide acanthite (Ag2S), which is monoclinic and gray in color, is common in geological and archaeological contexts. Until recently, X-ray diffraction data published by Tavernier (1966) were used by the Joint Committee on Powder Diffraction Standards (JCPDS) as the reference listing for AgAuS (JCPDS File No. 19-1146). This listing, however, has been deleted by the JCPDS and has been replaced with a listing based on the mineral, petrovskaite (JCPDS File No. 38-396). Analyses of this mineral have shown it to be close to the stoichiometric formula for 4 In an archaeological context, this sulfide has also been reported to occur as niello inlays in two medieval gold objects (Oddy, Bimson, and LaNiece 1983: 29-35). 143 J. FRANTZ AND D. SCHORSCH TABLE 2. Continued Lisht (G) Nephthys (G) Hatshepsut (F) Senebtisi (G) Box 2 Ukhhotpe (F) 1.39 AgAuS with a slight deviation resulting from the presence of between 1 and 2 weight percent selenium (Nesterenko et al. 1984: 604). In our analyses of archaeological red gold surfaces and substrates no selenium has been detected. Although the diffraction data for petrovskaite obtained by Nesterenko and colleagues (1984: 606) are similar to those published by Tavernier for synthetic AgAuS, there are some differences. In the high-angle region, the data that we obtained from both archaeological and synthetic samples agree more closely with those published by Tavernier, while in the low-angle region the addition by Nesterenko and colleagues of several lines of moderate intensity now accounts for previously unassigned peaks that consistently appeared in our work. A peak at 2.036 Å (I/I° = 10) reported by Nesterenko and colleagues was generally masked in our work by the (200) peak of the substrate alloy occurring at or near 2.039 Å (I/I° = 52). However, other peaks at this intensity were sometimes indicated to be below background. In nearly all the X-ray diffraction scans on both archaeological and synthetic red gold samples carried out, both with diffractometer and powder camera methods, the most intense peaks were those of the alloy substrate. While it is possible that the gold peak at 2.35 Å (I/I° = 100) is masking the peak at 2.30 Å (I/I° = 10) reported for synthetic AgAuS, it should also be noted that this peak is omitted by Nesterenko and colleagues in their listing for petrovskaite. As noted above, the compound Ag3AuS2 is gray in its natural and synthetic forms. The variation in color of both ancient and modern red gold may result from differences in both the actual thickness of the AgAuS layer, as well as in the amount of 20 the darker Ag3AuS2 present. Although petrovskaite is said to range in color from dark gray to black, Nesterenko and colleagues described its streak as dark gray sometimes with a slightly red-brown hue, and the mounted and polished mineral itself, when viewed in reflected light, as black with dark red splashes (1984: 602-603). These authors have also suggested a hypogenic origin for petrovskaite based on its discovery at a depth of 60 to 65 m. in association with hypogenic native sulfur and secondary copper and silver sulfides (1984: 607). Prior to our own study, there do not appear to have been any reported natural occurrences of AgAuS formed under what we presume to have been normal atmospheric conditions. When samples of both archaeological and synthetic silver gold sulfides were heated in air at 500°C, the coloration of the tarnish film quickly disappeared. Subsequent EDS analysis of these surfaces showed only the presence of the alloy substrate. We suspect that the discovery by Lucas that certain red and purple films could be removed by heating most often represented the oxidation in Fig. 5. Cross section of gold leaf fragment from LNP 600B, thickness ca. 4 microns 144 EGYPTIAN RED GOLD TABLE 3. X-ray diffraction data for selected synthetic samples (d-values in Angstroms listed together with relative intensities) 1 7.16 5 88 3.96 22 3.16 3.02 2.78 2.64 14 20 100 72 2.46 22 12 14 21 7.05 31 7.16 64 7.08 100 6.59 13 5 16 9 7 16 6.70 3.93 6.64 4.24 3.95 3.95 3.56 18 6 3.34 10 3.16 3.02 2.77 2.64 2.54 9 29 100 36 9 3.15 3.00 2.77 2.62 2.53 2.44 2.37 4 4 79 40 4 5 45 2.76 2.62 2.53 100 26 6 6.98 6.58 21 9 3.92 15 3.33 7 2.99 2.75 2.61 2.53 2.43 9 100 32 8 3 2.36 2.34 9 11 2.22 29 2.03 1.93 1.87 1.78 1.67 1.63 46 3 5 23 10 2 1.38 offscale 10 offscale offscale 2.25 2.23 39 1.94 1.87 1.78 1.68 4 6 10 15 offscale 2.24 2.22 2.17 2.13 2.11 20 33 9 16 14 offscale offscale 1.79 16 offscale 1.78 31 1.61 1.58 9 7 offscale 16 1.63 1.58 4 4 offscale 5 1.39 1.39 indicate that their formation is not dependent on the presence of copper. For the archaeological material, the apparent shift in d-spacings to values lower than those given in the literature for AgAuS or AgAu(Se,S) may reflect solid substitution by the small amount of copper contained in these alloys. There are unfortunately very few published analyses of ancient Egyptian gold, and there is much uncertainty as to when the Egyptians and other peoples of the ancient world began to refine gold.7 It has been suggested that in the earliest periods natural alloys of gold and silver, varying over a broad compositional range, were used (Aitchison 1960, II: 166-168; Stos-Fertner and Gale 1978). Our analyses of the archaeological gold substrates with red sulfide tarnishes indicate that the alloys conA number of analyses, but none from objects dating to later than the Second Intermediate Period, appear in StosFertner and Gale (1978). Lucas summarizes various earlier analyses of gold and electrum (1962: 490-491), but these data, as suggested by Stos-Fertner and Gale, may suffer from the primitive means by which they were acquired. 7 See above, Note 1. Scott 1983: 194-203. The topic is discussed with regard to pre-Columbian gold, which generally contains large percentages of silver and copper. The formation of AgAuS was not reported; a bibliography of vapor-phase tarnish studies is included. 5 5 8 14 16 air of sulfide tarnishes and not the combustion of organic material.5 The failure to produce conspicuous tarnish films on gold silver alloys by exposure to ammonium sulfide vapors at ambient temperatures may have resulted from the relatively short time over which the experiment was carried out. The exposure is being continued for a longer duration. Other researchers have determined that gold silver alloys containing less than 46.5 weight percent gold are susceptible to the formation of silver sulfide tarnish in contact with ammonium sulfide and that this ability to form tarnish is much greater in ternary (Ag-AuCu) alloys.6 Although all the archaeological AgAuS samples examined were found to contain small amounts of copper, our results in producing red tarnish films 6 2.29 2.25 2.24 1.79 offscale 1.39 offscale J. FRANTZ AND D. SCHORSCH 145 Fig. 6. Synthetic red gold coupons (left column from top: nos. 14, 1, 21; right column from top: nos. 6, 5, 12), length of 5 ca. 2.5 cm. Fig. 7. Synthetic red gold coupon no. 5 (bottom) and gold leaf fragment from LNP 600-B, length of no. 5 ca. 2.5 cm. no. tain between 15 and 68 percent silver, while our attempts to synthesize red sulfide tarnish films on gold silver samples at elevated temperatures were repeatedly unsuccessful using alloys containing less than 7.5 weight percent silver. These results should be interpreted with the caution appropriate to analyses of archaeological alloys. Although we have attempted to obtain reproducible geometries through the use of polished cross sections, the vitiating effects of such factors as small sample area, local enrichment or depletion of components, and inhomogeneities in the original object should be kept in mind. In the case of the Hatshepsut foundation deposit material, it will never be known if the red gold cylinder beads came from the same deposit as the untarnished gold ones, sharing the same "micro-burial" conditions, but the difference in their compositions can be documented.8 We attribute the occasional detection of sizeable amounts of chlorine on the archaeological materials examined to the not unexpected presence of silver chloride. Although photolytically altered deposits of silver chloride that exhibit a purple coloration are frequently found, neither the diffraction data nor the EDS results showed a correlation between the 8 The surface of a representative non red gold Hatshepsut cylinder bead was analyzed at 10, 20, and 30 kV, with the following results (expressed in atomic percent): 30 kV 20 kV 10 kV Au 83.6 86.1 84.4 Ag 13.6 11.5 15.6 Cu 2.9 2.1 nd No sulfur was detected. These values may be compared with those for the red gold bead reported in Table 1, where the silver content was found to be 26.9, 28.1, and 32.7 atomic percent at 30, 20, and 10 kV. Far more copper was found in the red gold bead but, as noted above, the absence or presence of small amounts of copper seems unrelated to the formation of a silver gold sulfide film. A substantial amount of chlorine was detected on the surface of the red gold bead, but no silver chloride was found in X-ray diffraction analysis. 146 EGYPTIAN RED GOLD 1 Sample: Nominal compositiona weight percent Ag 9.35 5 12 14 21 9.00 15.00 7.50 20.00 Cross-sectionsb [with standard deviation] w/o Ag 8.7 [0.3] 8.7 [0.2] 14.9 [0.5] 14.9 [0.3] at/o Ag urface scans Atomic percent at 30 kV [with standard deviation] Au Ag S 82.9c [1.0] 16.4c [0.1] 0.7c [1.0] Total Atomic percent at 20 kV Au Ag S 100.0 Total Atomic percent at 10 kV Au Ag S 100.0 Total 100.0 75.8d[0.7] 17.2d[1.3] 7.0d[1.7] 54.6d[2.0] 26.6d[1.3] 18.8d[2.0] 65.0 23.1 11.9 100.0 56.8 27.0 16.2 100.0 39.6 34.4 26.1 100.1 14.6 [0.2] 23.8 [0.3] 7.1[0.4] 12.2 [0.6] 72.0e 24.7e 3.3e 80.7 19.3 0.0 100.0 65.1e 26.6e 8.3e 100.0 44.7e 34.2e 21.2e 100.1 100.0 69.9 20.4 9.7 100.0 46.9 31.8 21.4 100.1 20.6 [0.2] 32.2 [0.3] [ 54.6 30.0 15.5 100.1 48.1 31.5 20.5 100.1 33.5 35.9 30.6 100.0 Notes a b c d e Based on preparation of synthetic samples. Based on average of three scans of cross-section at 30 kV. Based on average of two scans. Based on average of three scans. Recalculated to eliminate minor amount of chlorine detected. presence of chlorides and the observed red colorations. Implicit in this discussion and in the descriptions by Lucas quoted above is the question of whether all the coloristic effects observed on ancient Egyptian gold derive from more than one phenomenological origin. The purple-red coloration typically found to be associated with the presence of AgAuS in the present study appears to be the predominant variety of red gold from ancient Egyptian sources. In the Egyptian Museum in Cairo, there are numerous objects either made of gold or with gold leaf attachments that have this coloration on all or part of their surfaces. The most striking examples are the gilt wood shrines from the tomb of Tutankhamun originally found nested together; inside the smallest of these resided the sarcophagi and mummy of the king. The outer surfaces of these shrines are decorated with overlapping squares of gold leaf. Some of these squares are of a yellow gold color, while others are of the purple-red hues associated in the present study with AgAuS. Typically, the squares of a given color appear contiguous to one another in irregular groups as might be expected if, during application, the supply of gold leaf had been replenished as needed but without regard for differences in silver content between batches.9 In the original metallic state the slight variations in color produced by such differences in silver content may well have been deemed acceptable. In our survey of objects in Cairo, as well as in other institutions including the Metropolitan Museum, by far the most numerous instances of red gold fall into this category with respect to both the irregular distribution and the predominantly dark purple hue of the color. Indeed, it would seem that such instances of red gold are, as 9 This is especially evident on the innermost shrine, No. 1319. As noted above, at least two very different silver gold alloys were used for the gilding of Senebtisi's anthropoid coffin. J. FRANTZ AND D. SCHORSCH Lucas first suggested, the result of slow, fortuitous tarnishing.10 In contrast, there are other, far less numerous instances in which the gold is of a distinctly brighter and more transparent, cherry red color. These occurrences include several fenestrated plaques or belt buckles from the tomb of Tutankhamun which are exhibited in Cairo in a case together with the well-known gold mask. The sequins investigated by Wood may also belong to this visually described group.11 Typically, this luminous red occurs on solid gold objects rather than on foil or leaf overlays and, generally, over the entire surface of the pieces in question. In our study, we have not encountered red gold surfaces of this particular variety that were available for analysis. The only instances in which we have found iron to play a role in the coloration have been those where opaque crusts of lepidocrocite have occurred as what are clearly adventitious accretions. A repetition of the experiments carried out by Wood, however, indicates that it is quite simple to color small pieces of gold with iron so that the resulting surface closely resembles the transparent red seen on the fenestrated plaques and sequins mentioned above.12 Wood's replicated sequins can still be seen in the Egyptian Museum in Cairo. Their surface appearance is also very close to that of the fenestrated plaques, and a partial closure on this topic would be provided by the characterization of the surfaces on these unusual objects. Unfortunately, the manner in which Wood's sequins are displayed confuses the issue, as they are shown not in proximity to the fenestrated plaques, but in a case some distance away that also contains an assortment of purple-red gold 10 Other examples from Tutankhamun' s tomb together with their exhibition numbers include: Horus falcon (No. 418; standing Horus (No. 412); pharaoh with flail (Nos. 994, 995); walking sticks with gilded handles (Nos. 129, 130, 1052, 1056, 1608, 1609, 1610); standing gilt wood pharaohs (Nos. 468, 474, 906, 1086, 1087); standing gilt Isis (No. 425); standing gilt Nephthys (No. 421). 11 A reasonably accurate color reproduction of a pair of Tutankhamun's earrings also having this transparent cherry red surface is found in Aldred 1971: pl. 122. 12 Using an acetylene torch, a sample of pure gold was heated with approximately 1 weight percent hematite (Fe203) in an open porcelain crucible until the gold was melted. Upon cooling, the gold drop displayed a thin, adherent pink film on its surface. 147 beads, most of which visually resemble the AgAuS variety and appear conspicuously different from Wood's sequins, as well as the ancient examples they emulate. To judge by Wood's description of the patterned array in which the red gold sequins from Tutankhamun's tomb were found, they appear to be the rare case where an aesthetic ground can be offered for their color, regardless of its origin. In our examination we have found no objects coated with an even and unbroken red film, and no cases in which the color seemed explicitly to be associated with the design of an object. A single but often repeated ancient source is quoted as textual evidence for the deliberate manufacture of red gold. Among the so-called Amarna Letters is one received by the Eighteenth-Dynasty king, Amenhotep III, describing "gold through which blood shines" among the gifts sent to him from the king of Mitanni (northwestern Mesopotamia) (Knudtzon 1908—15: XXII [p. 156-157, line 20], XXV p. 190— 191, line 27). This allusion has been taken to mean that the art of making red gold was learned from a western Asiatic culture.13 Red surface films on gold antiquities from other cultures are not uncommon, although in most cases there have been no theories about intentional technological processes.14 We have often observed such coloration on untreated objects, and it seems probable that numerous other examples once existed that have been lost through cleaning.15 One 13 Ogden writes: "Mesopotamian texts do refer to the addition of hematite (iron oxide) to gold and colouring seems to be the only likely reason for doing so." He gives no references to such texts (1982: 19). 14 An early reference to gold tarnish (though not specifically red) appears in Strabo's Geography (16.2.42) in which he refers to Lake Sirbonis. ("Strabo seems obviously to be confusing the Asphaltities Lacus [the Dead Sea] with Lake Sirbonis, which latter `broke through to the Mediterranean Sea'. " [Jones, tr. 1917-32, VII: 293, fn. 3].) "It is full of asphalt. The asphalt is blown to the surface at irregular intervals. . . . With the asphalt there rises also much soot, which though smoky, is imperceptible to the eye; and it tarnishes copper and silver and anything that glistens, even gold; and when their vessels are becoming tarnished the people who live around the lake know that the asphalt is beginning to rise" (Jones, tr. 1917-32, VII: 293). The Akan (Ghana) practice of boiling gold objects in ochre, occasionally described as a coloring practice, is in fact carried out with ochre and salt and appears to be for the purpose of surface enrichment prior to burnishing (Garrard 1980: 122-123). 15 An early work on the conservation of antiquities contains a reference to red surfaces, presumed to be iron corrosion products, and instructions on their removal (Rathgen 191524, II: 44). 148 EGYPTIAN RED GOLD TABLE 5. Summary of results Name Object MMA acc. no. Date Site Cause of coloration Method(s) of identification Nephthys Coffin 11.150.15B Ukhhotpe Funerary Mask 12.182.132C Vulture Pectoral 26.8.104 Rosette Headdress 26.8.117 12th dyn. Meir AgAuS XRD, EDS 12th dyn. Meir AgAuS XRD, EDS 18th dyn. Thebes AgAuS Wet chemistry 18th dyn. Thebes FeO(OH) XRD, IR Hatshepsut Tubular beads 27.3.444 Cowrie Cowrie beads 30.8.383–385 18th dyn. Thebes AgAuS XRD, EDS Middle Kingdom site unknown Middle Kingdom site unknown 12th dyn. Lisht Ag3AuS2(?) XRD, EDS Ag3AuS2(?) XRD, EDS AgAuS XRD, EDS 12th dyn. Lisht AgAuS XRD, EDS Bracelet 30.8.386–393 30.8.394A Lisht Gold leaf LNP-600B Senebtisi Gold leaf LNP-763 famous find of red gold that has been partially cleaned is from the Royal Cemetery in Ur excavated by Sir Leonard Woolley in 1927.16 Two other conspicuous examples of non-Egyptian red gold are seen among the trappings from a warrior's burial of the sixth century B.C., believed to be from northern Greece and now in the Walters Art Gallery in Baltimore (Garside 1980: cat. no. 235), and the recently excavated late Hallstatt gold treasure from Eberdingen-Hochdorf and now in the Wurttembergisches Landesmuseum in Stuttgart (Biel 1978-79). In the case of the Ur gold, the red surface coloration is not much noted or explained in the literature. The red surface on the mask in the Walters Art Gallery has not 16 Plenderleith 1934: 284-310. The bulk of the material from the Royal Cemetery at Ur was divided between the British Museum, the University of Pennsylvania Museum in Philadelphia, and the National Museum in Baghdad. A number of pieces of jewelry and ornaments are in the Metropolian Museum of Art (MMA 33.35.1–50). The jewelry now in the British Museum was cleaned; comparative material in Philadelphia as well as a few items in the Metropolitan Museum are still bright red. been analyzed, but is generally assumed to result from the presence of iron.17 The treasure from Hochdorf was analyzed by A. Hartmann of the Wurttembergisches Landesmuseum, and the red surface was found to be an iron oxide.18 CONCLUSIONS Our study indicates that the majority of red gold surfaces from ancient Egypt derive from the formation of the species AgAuS on gold silver alloy substrates (Table 5). The different hues observed on surfaces exhibiting this sulfide appear to be associated 17 The pieces are believed to have belonged to a "secondary burial," a type known at an analogous site in Trebenischte (Yugoslavia), in which the fragments of a partially cremated body and its funerary equipment were arranged in a fulllength grave. The red coloration is presumed to have been produced according to the mechanism described by Wood, though accidentally, and is taken as evidence for the cremation (Hill 1964–65). The piece was recently examined and, because of the thinness of the film, was difficult to sample; powder diffraction analysis was inconclusive but iron does not seem to be the cause of the coloration. 18 A. Hartmann, personal communication, 10 August 1984; the method of analysis was not indicated. J. FRANTZ AND D. SCHORSCH with variations in thickness of the film as well as with relative amounts of other phases present, including the compound, Ag3AuS2. Since our study was begun, other researchers have identified similar occurrences of silver gold sulfides on ancient Egyptian gold surfaces (Hatchfield and Newman 1990). At elevated temperatures, using molten sulfur, we have been able to produce adherent red, purple, and black films of AgAuS and Ag3AuS2 on alloys ranging in composition from about 8 weight percent to about 20 weight percent silver. We have not investigated the conditions under which these films are formed at ambient temperatures other than through the primitive, shortterm exposure tests with ammonium sulfide described above. We plan to continue this research on a rigorous basis with exposure of alloys to controlled amounts of hydrogen sulfide, water vapor, oxygen, and other agents. In the interim, we conjecture that the burial environments for the archaeological material examined undoubtedly contained both biological and geological sources of sulfur that resulted in the atmospheric tarnishing of the alloys.19 In19 Lucas notes the high sulfur content of embalming materials and other substances in his chapters on cosmetics and mummification. He suggests decomposed orpiment (As2S3) as the source of sulfur responsible for silver sulfide tarnishes observed on electrum and silver among the Tutankhamun objects (1927: 174). A few sulfur objects have been excavated in Egypt, and Pliny describes a niello process, necessitating the use of sulfur, undertaken by Egyptians of his time; however, little evidence exists for this practice in pharaonic times, and not before the early Eighteenth Dynasty (Lucas 1962: 250-252, 269). Sulfur-containing materials may also have been used as supports for raised or embossed gold work. 149 deed, it appears likely that the conditions of environment and alloy composition necessary for the formation of silver gold sulfide tarnishes are not limited to ancient Egypt, and that a survey of red surface films on gold objects from other cultures would also yield examples of these phenomena.20 We have found instances where other causes of surface coloration may be invoked—notably, the cases of lepidocrocite accretions and photolytically altered silver chloride deposits. These occurrences, however, appear to be in the minority and to be visually distinct in coloration and morphology. The question of the transparent, cherry red surfaces found on the few artifacts discussed—and apparently the subject of Wood's earlier study—remains open. Whether this phenomenon represents the occurrence as a thin film of accretionary lepidocrocite or related species, or whether it derives from the inclusion of iron-bearing salts—either deliberately added or naturally occurring—in the gold alloy itself, are topics for further research. In retrospect, our conclusions confirm many of those arrived at by Lucas more than sixty years ago, albeit with some added specificity for the causative agents. In agreement with those earlier suggestions, all the colorations—with the possible exception of the transparent reds mentioned above—seem decidedly fortuitous in origin. 20 A sample of a red surface accretion removed from a gold strip excavated at the Royal Cemetery at Ur (MMA 33.35.30) was identified by X-ray diffraction as AgAuS (MMA Film 1008). Acknowledgments We would like to thank the following colleagues for their various contributions: Mark Wypyski, for the EDS analysis; Karin Willis, for photography; Alexander Shedrinsky, for translation; Richard E. Stone and Mark D. Barton for their useful commentary. We also thank the Egyptian Department of the Metropolitan Museum of Art, and especially Christine Lilyquist, Dorothea Arnold, and Marsha Hill, for their cooperation and comments; and Terry Drayman Weisser of the Walters Art Gallery, Maude de Schauensee of the University Museum, and Nasry Iskander of the Egyptian Museum in Cairo for facilitating the examination of gold objects in their institutions. This research was supported in part by the L. W. Frohlich Charitable Trust. APPENDIX Description of Archaeological Gold Samples Nephthys (Fig. 2). MMA 11.150.15B (Hayes 1953-59, I: 311–312); anthropoid coffin of Nephthys; 12th Dynasty; Meir, B 3, pit 3; Khashaba excavations, 1910. This anthropoid coffin belongs to one of five burials in the Metropolitan Museum of Art excavated at the Middle Egyptian site of Meir. Three of these burials include ob- 150 EGYPTIAN RED GOLD jects partially gilded with red gold leaf.21 The coffin is painted cartonnage; from the neckline of the dress to the hairline is gilded. Bright reddish coloration is concentrated on the face proper; the underside of the chin and the top part of the neck are very dark red, while the ears are the bright yellow color of metallic gold. Prior to a recent consolidation, several loose fragments were available for sampling. The undersides of the samples were untarnished. The rectangular outer coffin of Nephthys (11.150.15A) is also partially gilded with red gold leaf. There is no mention of the red coloration on either coffin in the published excavation report (Kamal 1911: 11— 13) or in early museum records. Ukhhotpe (Fig. 1). MMA 12.182.132C (Dorman, Harper, and Pittman 1987: 35— 37); funerary mask of Ukhhotpe; 12th Dynasty; Meir; Khashaba excavations, 1910. This mask belongs to another of the three "red gold" burials from Meir. The mask is painted wood, and the entire face, including the separately made ears, is dark red. A small fragment of loose gold leaf was removed from an inconspicuous area on the proper left side of the face for EDS analysis of the metal and the red surface. A small surface scraping was taken for X-ray diffraction analysis from the same area. There is no mention of the red coloration in the published excavation report (Kamal 1912) or in early museum records. Vulture. MMA 26.8.104 (Winlock 1948: 43– 44, pl. XXV); vulture pectoral; 18th Dynasty; Thebes. This pectoral is one of three from the socalled "Three Princesses Treasure," a group of objects in the Metropolitan Museum that were discovered in 1916 and that are thought to have come from a royal tomb in the Wadi Gabbanat el-Qirud (Lilyquist, forthcoming). According to the inscriptions on some of the objects, the tomb belonged to three foreign queens of Tuthmosis III In addition to the funerary mask of Ukhhotpe, which was sampled for this study and described below, the canopic box of Hapyankhtifi (MMA 12.183.14) is partially decorated with red gold. The burial of Khnumhotpe is believed to have been excavated at Meir as well; the face of the funerary mask (MMA 12.182.131) is a striking example of red gold. 21 (ca. 1479—1425 B.C.) The burials contained hundreds of gold objects, the majority of which were purchased by the Metropolitan Museum. The red coloration on the vultures and many other of the gold pieces was not noted when the treasure was published in 1948. Rosette. MMA 26.8.117 (Winlock 1948: 13– 17, pls. I I I and V); inlaid rosette; 18th Dynasty, reign of Tuthmosis III; Thebes, Wadi Gabbanat el-Qirud. This rosette belongs to the so-called Three Princesses Treasure, and must have come from a tomb for three foreign queens of Tuthmosis III in the Wadi el-Qirud that was discovered in 1916. It is one of a large number of similar carnelian and glass inlaid gold rosettes that are believed to have belonged to elaborate headdresses. The rosettes in the Metropolitan Museum were acquired over a period of many decades. Many are covered with splotchy red stains ranging from a bright resinous red to an opaque dark brownish red that were not noted when the rosettes entered the museum or in the 1948 publication of the "Three Princesses Treasure." Hatshepsut (Fig. 3). MMA 27.3.444 (Hayes 1953–59, II: 88); tubular beads; 18th Dynasty, reign of Hatshepsut; Thebes, Deir el-Bahri, Temple of Hatshepsut, deposits 7–9; Metropolitan Museum excavations, 1926–1927. These beads are among the hundreds of gold and silver beads from three of the fourteen foundation deposits of the temple of the Eighteenth-Dynasty queen, Hatshepsut. The specific deposit in which a given bead was found is not recorded. Within the group, four variations can be distinguished: untarnished gold beads, gold beads with a red tarnish, silver beads with a black tarnish, and silver beads with a massive silver chloride crust. There is no mention of the surface colorations in early records. Cowrie. MMA 30.8.382–385 (Winlock 1934: 40, n. 13); cowrie shell beads; Middle Kingdom; provenance unknown. These beads are believed to have been part of a girdle. They were donated to the Metropolitan Museum in 1915 by Theo- J. FRANTZ AND D. SCHORSCH 151 dore M. Davis. There are no references in early records about the surface appearance of the beads; it was, however, noted that they were cleaned and lacquered in 1933. At the present time the cowries are covered with a thick film of black corrosion. Surface analysis of a recently cleaned spot indicates that the alloy is very rich in silver. excavated in the early twentieth century. One side of each gold fragment is covered with a continuous silvery red tarnish; the other side was presumably attached to the object, and each piece has reddish coloration around its edges, in many cases extending into the interior.22 ments of a bracelet; Middle Kingdom; provenance unknown. These fragments were donated to the Metropolitan Museum in 1915 by Theodore M. Davis. There is no mention of their dark gray coloration in early records. gold leaf fragments; 12th Dynasty; Lisht, North Cemetery, shaft 763, burial of Senebtisi; Metropolitan Museum excavations, 1905—1906. The gold leaf fragments found in the burial of Senebtisi at Lisht covered a wooden anthropoid coffin that did not survive burial. In the course of preparations for the installation of the present Lisht galleries, these fragments were separated for display purposes on the basis of their physical appearance. One portion (Box 2) exhibits a surface typical of Egyptian red gold, while the other portion (Box 1) is darker and grayer in color. The metal itself from Box 1 appears more silvery than that from Box 2. Senebtisi. LNP 763 (Mace and Winlock Bracelet. MMA 30.8.386—393, 394A; frag- 1916: 16, 38, pls. X and XI, frontispiece); Lisht (Fig. 4). LNP 600B; gold leaf fragments; 12th Dynasty; Lisht, North Cemetery, burial 600B; Metropolitan Museum excavations, 1907—1908, 1908—1909. The gold leaf found in this deposit at Lisht probably belonged to a plaster mask that did not survive burial or was destroyed by looters. The red surface coloration was noted by the excavators, who believed it to be a glue used to attach the gold (unpublished excavation notes, Egyptian Department, MMA). There are no records and no visible evidence to suggest that these gold scraps have been cleaned or their surfaces otherwise tampered with since they were 22 Based on observations of the two pieces among the scrap that clearly ornamented the ears of the mask, for which inner and outer surfaces could be conclusively recognized, it appears that what originally were the outer surfaces of the gold are more heavily tarnished. REFERENCES Aitchison, L. 1960 A History of Metals. 2 vols. MacDonald and Evans Ltd., London. Aldred, C. 1971 Jewels of the Pharaohs. Praeger, London. Barton, M. D. 1980 The Ag-Au-S System. Economic Geology 7 5 : 313-316. Barton, M. D., C. Kieft, E. A. J. Burke, and I. S. Oen 1978 Uytenbogaardtite, a New Silver-Gold Sulfide. Canadian Mineralogist 26: 651-657. Barton, P. B., Jr., and P. Toulmin, III 1964 The Electrum-Tarnish Method for the Determination of the Fugacity of Sulfur in Laboratory Sulfide Systems. Geochimica et Cosmochimica Acta 28: 619-640. Berthelot, M. 1901 Sur l'or égyptien. Annales du service des Antiquités de l'Égypt 2 : 157-163. 1902 Nouvelles recherches sur les alliages d'or et d'argent et diverses autres materières prove- nant des tombeaux égyptiens. Annales de chemie et de physique, ser. 7, no. 25: 5965. Biel, J. 1978-79 Das frühkeltische Fürstengrab von Eberdingen-Hochdorf, Landkreis Ludwigsburg. Denkmalpflege in Baden-Württemberg 7 - 8 : 168-175. Carter, H. 1923-33 The Tomb of Tut-Ankh-Amen. 3 vols. Copper Square Publishers, Inc., New York. Dorman, P., P. O. Harper, and H. Pittman 1987 Egypt and the Ancient Near East. Metro politan Museum Series -I. Metropolitan Museum of Art, New York. Feigl, F., V. Anger, and R. E. Oesper 1972 Spot Tests in Inorganic Analysis. 6th ed. Elsevier, Amsterdam. Garrard, T. F. 1980 Akan Weights and the Gold Trade. Longman, London and New York. 152 EGYPTIAN RED GOLD Garside, A. (editor) 1980 Jewelry, Ancient and Modern. Viking Press, New York. Graf, R. B. 1968 The System Ag3AuS2-Ag2S. The American Mineralogist 53: 496—500. Hatchfield, P. and R. Newman 1990 Ancient Egyptian Gilding Methods. In Gilded Wood Conservation: Proceedings of the 1988 Conservation Symposium (forthcom ing). Hayes, W. C. 1953—59 The Scepter of Egypt. 2 vols. Metropolitan Museum of Art, New York. Hill, D. K. 1964—65 Helmet and Mask and a North Greek Burial. Journal of the Walters Art Gallery 27—28: 9—15. Kamal, A. 1911 Rapport sur les fouilles executees dans la zone comprise entre Deirout au nord et Deir-elGanadlah au Sud. Annales du Service des Antiquites de l'Egypte 11: 3—39. 1912 Rapport sur les fouilles executees dans la zone comprise entre Deirout au nord et Deir-elGanadlah au Sud (suite). Annales du Service des Antiquites de l'Egypte 12: 97—127. Knudtzon, J. A. 1908—15 Die el-Amarna Tafeln. 2 vols. J. C. Hinrichs'sche, Leipzig. Lilyquist, C. 1992 The Tomb of Tutmosis III's Three Asiatic Wives. Metropolitan Museum of Art, New York (forthcoming). Lucas, A. 1926 Ancient Egyptian Materials and Industries. 1st ed. Edward Arnold and Co., London. 1927 The Chemistry of the Tomb. In The Tomb of Tut-Ankh-Amen, by H. Carter, Vol. 2, pp. 162—168. Copper Square Publishers, New York. 1932 Antiques, Their Restoration and Preservation. Edward Arnold & Co., London. 2nd revised ed. 1934 Ancient Egyptian Materials and Industries. 2nd ed. Edward Arnold and Co., London. 1962 Ancient Egyptian Materials and Industries. 4th ed. Revised and enlarged by J. R. Harris. Edward Arnold and Co., London. Mace, A. C. and H. Winlock 1916 The Tomb of Senebtisi at Lisht. Metropolitan Museum of Art, New York. Nesterenko, G. B., A. I. Kuznetsova, N. A. Palchik, and Y. G. Lavrentiev 1984 Petrovskaite: A New Selenium Sulfide of Gold and Silver. Vsesoiuznoe Mineralogicheskoe Obshchestvo Zapiski 113: 602—606. [In Russian] Oddy, W. A., M. Bimson, and S. LaNiece 1983 The Composition of Niello Decoration in the Antique and Mediaeval Periods. Studies in Conservation 28 (1983): 29—35. Ogden, J. 1982 Jewellery of the Ancient World. Rizzoli, New York. Plenderleith, H. J. 1934 Metals and Metal Techniques; Metal Objects. In Ur Excavations II: The Royal Cemetery, by C. L. Woolley, pp. 284—310. Oxford University Press, London and Philadelphia. Plenderleith, H. J. and A. E. A. Werner 1971 The Conservation of Antiquities and Works of Art. 2nd ed. Oxford University Press, London. Rathgen, F. 1915—24 Die Konservierung von Altertumsfunden. 2nd ed. 2 vols. Walter de Gruyter, Berlin. Scott, D. A. 1983 The Deterioration of Gold Alloys and Some Aspects of Their Conservation. Studies in Conservation 28: 194—203. Stos-Fertner, Z. and N. H. Gale 1978 Chemical and Lead Isotope Analysis of Ancient Egyptian Gold, Silver and Lead. In Proceedings of the 18th International Symposium on Archaeology and Archaeological Prospection (Archaeo-Physica 10), edited by I. Scollar, pp. 299—314. Rheinland/Rudolf Habelt Verlag, Cologne and Bonn. Strabo 1917—32 The Geography of Strabo. 8 vols. Translated by H. L. Jones. The Loeb Classical Library, Harvard University Press, Cambridge, Massachusetts. Tavernier, B. H. 1966 Über Silber-Gold(I)-Chalkogenide. Zeitschrift für anorganische and allgemeine Chemie 343: 323—328. Williams, C. R. 1924 Catalogue of Egyptian Antiquities, Nos. 1160, Gold and Silver and Related Objects. The New York Historical Society, New York. Winlock, H. E. 1934 The Treasure of el Lahun. Metropolitan Museum of Art, New York. 1948 The Treasure of the Three Egyptian Princesses. Metropolitan Museum of Art, New York. Wood, R. W. 1934 The Purple Gold of Tutankhamun. Journal of Egyptian Archaeology 20: 62—65.