The introduction of the lapidary engraving wheel in Mesopotamia.

SAX, MARGARET ; MEEKS, NIGEL D. ; COLLON, DOMINIQUE 等

One of the most significant advances to have been made in early lapidary technology was the introduction of the bow-driven engraving wheel. The high-speed cutting action, allowed by rotary motion, made possible substantial improvements in the efficiency of working hard stones. However, there is uncertainty concerning the date of this key invention and its adoption in the Near East. Early in this century, Ward (1910: 9) and Frankfort (1939: 5) proposed a date in the 2nd millennium BC while, for the past 25 years or so, a much earlier date in the second half of the 4th millennium BC has generally been assumed (for example, see Nissen 1977: 16; Gwinnett & Gorelick 1979: 25-7; Collon 1986: 13-14). These views have generally been based upon the interpretation of the designs cut into stone cylinder seals.

Cylinder seals (FIGURE 1) developed alongside the cuneiform system of writing, beginning sometime around 3500 BC, and they provide a more or less continuous record of lapidary methods over three millennia. Amuletic and decorative functions were also associated with cylinder seals and they were perforated longitudinally, so that they could be suspended and worn by the owner. The principal methods used to drill these perforations have been identified by Gwinnett & Gorelick (1987). Furthermore, the fine detail preserved in the intaglios of the seals is ideal for study of ancient lapidary technology and a systematic investigation by the present authors has shown that it is possible to identify the methods used to engrave the seals (Sax & Meeks 1995; Sax et al. 1998; see also Sax & Meeks 1994). The various techniques, tools and abrasive materials can usually be recognized from the characteristic morphology or `tool marks' of the engraved features.

[Figure 1 ILLUSTRATION OMITTED]

It emerged in our earlier investigation that simple flaking techniques of engraving prevailed in the late 4th millennium BC and throughout the 3rd millennium BC, with no evidence for the use of wheel-cutting techniques. However, wheel-cutting became the dominant technique in the 1st millennium BC. We have therefore focused the present study upon the changes that occurred during the 2nd millennium, with the aim of establishing the chronology of the introduction of the wheel and the other technological changes with which it was associated.

Methodology

Identification of the materials of the 2500 or so seals in the collections of the British Museum has shown that increasing proportions of harder stones, such as rock crystal, chalcedony, carnelian, agate and other varieties of quartz, were worked with time (Sax in Collon 1982; 1986; in press; Sax 1991; Sax & Middleton 1992). Quartz was the hardest material commonly worked (Mohs' hardness, H=7) and it is likely that the difficulty of working it stimulated innovations in technique which were less advantageous to the working of seals in softer materials. The investigation into engraving methods therefore concentrated upon the 400 or so quartz cylinder seals in the British Museum collections. On stylistic grounds, they range from c. 3100 BC to c. 400 BC (Collon 1982; 1986; & work in progress; Matthews 1990).

Following the approach of Sax & Meeks (1995: 26-7), a chronological survey of the intaglios on the seals was made with a low-power binocular microscope. Seals that were considered to represent key stages in the application of engraving techniques were selected for examination using scanning electron microscopy (SEM). To avoid the need to apply a conductive coating to the seals, detailed impressions were made with a silicone moulding material. As the designs on the seals were worked in intaglio, the engraved features appear on moulded impressions as positive features. Hence, SEM images of the impressions (FIGURES 2-3 & 5) show the intaglios `in reverse' with protrusions on the image representing depressions in the original seal. An advantage of this approach is the ability to examine details of the deeper parts of engraved features.

[Figure 2-3, 5 ILLUSTRATION OMITTED]

As noted above, the present study was focused upon seals dated to the 2nd millennium BC. These comprised: 54 of the Old Babylonian period (c. 2004--1595), 35 of Kassite style (predominantly 14th century), 16 of Mitannian style (c. 1500-1330) and 22 of the Middle Assyrian period (13th century). In addition, because a very high proportion of the 650 Old Babylonian seals in the collections are hematite, a mineral that is softer than quartz but still relatively hard (H=5-6), about 60 of these seals were examined.

Our initial interpretation of working methods was based on observation of seal intaglios and engraved features produced experimentally, using a range of tools and abrasive materials (Sax & Meeks 1995; Sax et al. 1998). The results indicated that the intaglios were engraved by one or more of four principal techniques: micro-flaking (termed micro-chipping in our earlier papers), involving the use of a sharply pointed tool; filing, involving the use of an elongate tool, which was either straight or, occasionally, like a modern riffler, curved, in a backwards and forwards sawing motion; drilling, involving the use of a tool-head rotating about an axis that was essentially perpendicular to the surface being engraved and wheel-cutting, involving the use of a fiat, disc-like tool-head rotating about an axis that was essentially parallel to the surface being engraved. The tools used for these four techniques did not usually possess working surfaces themselves

Engraving methods in the 2nd millennium BC

Tools and abrasives

The engraving experiments indicated that the tools used during the 2nd millennium BC were metal. Indeed, Frankfort (1939: 5) described what appeared to be the stock-in-trade of a lapidary: it contained `copper' gravers, chisels and a drill packed in a small pot together with cylinder seals and beads. The hoard was excavated at Tell Asmar to the east of Baghdad and dated to the Akkadian period, c. 2334-2193 BC. Although the composition of the excavated chisels etc. was not confirmed by analysis, both copper and its alloys provide reasonably rigid tools.

Copper-based tools are rarely useful for engraving hard stones unless they are suitably charged with abrasive. Emery has the advantage of being considerably harder (H=9) than quartz (H=7) and our experiments showed that relatively continuous and parallel grooving, similar to that on the skirt of the goddess in FIGURE 2b, is produced by emery abrasive mixes but that grooving of this type is not produced by softer abrasives such as quartz (Sax et al. 1998: figure 4). The results of the present investigation suggest that emery was employed for working the intaglios of quartz seals throughout the 2nd millennium and are in agreement with those of Gwinnett & Gorelick (1987; 1989) who found evidence for the its use by c. 2000 BC in the drilling of the central perforation of seals.

Techniques

The characteristics of the four techniques that we have been able to identify, micro-flaking, filing, drilling and wheel-cutting, are discussed briefly below and illustrated in FIGURES 2 & 3.

The impression of the king in FIGURE 2a exemplifies micro-flaking. Almost all of the surface of the intaglio is finely pitted with angular cavities. The king was probably microflaked using bronze chisels whose working edges were charged with emery (Sax et al. 1998: 4-5, 14-17, figure 2). The experiments showed that conchoidal cavities are produced perpendicular to the direction of force applied to the chisel, creating cross-sectional cavities, the size of which depends upon the magnitude of the applied force. The minute size of the cavities on this seal suggests that it was engraved by pressure flaking, perhaps with hafted chisels, rather than by indirect percussion using a hammer.

The goddess in FIGURE 2b illustrates not only filing but also the limitations of this technique. Engraved features with a horizontal or diagonal orientation on the seal, such as the layers of the skirt, have surfaces that are grooved longitudinally. The features are straight in plan and, furthermore, their longitudinal profiles of depth are straight. These characteristics are consistent with filing (Sax et al. 1998: 6, 18, figures 3-4). They were replicated by applying a simple hand-held copper rod, c. 70 mm long and charged with emery, in a backwards and forwards sawing motion. The experiments demonstrated that the production of simple filed features was straightforward across curved surfaces, i.e. in horizontal and diagonal orientations on cylinder seals (FIGURE 1), but it was not possible to engrave small features along flat surfaces, i.e. in a vertical orientation on seals, with a straight file. Thus, the vertical features of the goddess in FIGURE 2b, such as the sides of the skirt, have a pitted texture indicating that they were micro-flaked.

In FIGURE 2c, the circular tool marks of the unfinished design cut into a hematite seal indicate that the main elements of the figure were blocked out by drilling. The faint circular grooving present on them, as well as similar features observed on contemporary quartz seals is consistent with the use of copper or bronze drills charged with emery (Sax et al. 1998: 7, 18-19, figure 5).

The seated king in FIGURE 3 exemplifies wheel-cutting. The surfaces of almost all the deeply engraved features are characterized by faint, parallel and more or less continuous longitudinal grooving. The design is curved to linear in plan (FIGURE 3a) and the features are also curvilinear in depth (see oblique view of king's head in FIGURE 3b). Furthermore, the engraved features occur in all orientations on the seal. These four characteristics are typical of wheel-cutting (Sax et al. 1998: 7-8, 19, figure 6).

Chronology of engraving techniques

The main chronological changes in engraving practice which occurred in Mesopotamia and the surrounding areas of the Near East between c. 3100 BC and c. 400 BC are summarized in FIGURE 4. There appears to be no evidence for the use of the engraving wheel prior to the mid 18th century in the Old Babylonian period. Instead, the intaglios of quartz seals were typically worked by micro-flaking; filing and drilling techniques were used for minor parts of the intaglios. However, wheel-cutting was definitely practised in some Babylonian workshops on seals engraved in the Kassite style of the 14th century (it was also used in Assyria during the 13th century). The intervening phase between the mid 18th and 14th centuries is complex. It was a time of technical change during which several new approaches to engraving were developed, including an increase in the use of filing and drilling. In addition, we found tentative evidence for the first use of the engraving wheel.

[Figure 4 ILLUSTRATION OMITTED]

A factor that is likely to have contributed to the concept of the engraving wheel was the increased use of drilling. The distinction between the two rotary tools, drills and wheels, can be blurred and this is occasionally apparent in the intermediate phase between the 18th and 14th centuries. For example, while the unfinished design in FIGURE 2c shows some typical circular features characteristic of the drill, some of the features in the lower part of the figure are elliptical in shape and grooved longitudinally. These elliptical features appear to have been cut using the sides rather than the end of a drill. The `drill' was probably roughly spherical in shape so that, when the seal surface being engraved was held obliquely or parallel to its rotating shaft, the tool-head acted as a type of `wheel'. It would not have been practicable to use a drill which was held vertically by a capstone in this `wheel' mode. Rather, it would have been necessary to construct the tool with the shaft mounted in fixed bearings. It appears that this necessary precursor to the use of the bow-driven wheel was in use by c. 1750-1595 BC. The use of the `drill' in this `wheel' mode would, presumably have led to an appreciation of the principle that cutting with the side or `rim' of a rotary tool can be faster and more efficient than with the centre of the tool. In particular, the linear speed at the rim (sometimes referred to as peripheral speed) is proportional to the diameter of the tool, so that the greater the diameter, the faster is the rim speed for a given of rotational speed of the shaft. A disc-shaped tool upholds this principle and a high-speed cutting action can be produced at the rim of a bow-driven wheel. Furthermore, the convex rim of a thin disc can be used to engrave features in any orientation on a seal.

In the intermediate period between the 18th and 14th centuries there is some evidence that the new technique of wheel-cutting was used along with filing and/or drilling and occasionally micro-flaking. Interpretation of the intaglios in this phase can be difficult. An example of the characteristics observed in this phase is given in FIGURE 5, showing the upper body of a king in an oblique view. It seems likely that the feature between two arrows and the almost vertical arm, immediately to the left, were wheel-cut. These features are characterized by longitudinal grooving. They are curved in depth and have a vertical orientation on the seal. In contrast, the sub-horizontal feature which forms the shoulders was worked by a straight file. It has a straight longitudinal profile of depth and is characterized by longitudinal grooving which is more pronounced than that of the vertical features. Our examination of the intaglio on this seal suggested that, while the larger features were worked with files, many of the smaller features exhibit characteristics which can be explained by wheel-cutting; some of them may have been partly micro-flaked and/or filed. The technical changes that occurred in the late Old Babylonian period, c. 1750-1595 BC, resulted in a cursory style of engraving with very little detail (FIGURES 2b & 5). Buchanan (1970) saw the introduction of this style as an artistic revolution. Our evidence suggests that its adoption proceeded alongside a technical evolution that culminated in the invention of the engraving wheel as it is understood today.

Unfortunately, a lack of detailed chronological evidence following the fall of the First Babylonian dynasty in c. 1595 BC, precludes a precise estimate for the timing of the first use of the lapidary wheel. The period intermediate between the two dynasties in Babylonia is conventionally regarded as a Dark Age. However, Collon (1987: 58) and Matthews (1990: 55-7) have noted a continuity in style between Old Babylonian and Kassite glyptics. The results of the present investigation provide parallel evidence for continuity in the practice of engraving techniques. Wheel-cutting, probably conceived in the late Old Babylonian period, appears to have been progressively developed for the engraving of hard stone seals so that efficient methods of wheel-cutting were established sometime between c. 1400 BC and the end of the reign of Kurigalzu II in 1308 BC.

Discussion and conclusion

The chronological development of engraving practices is compared with the percentage of hard quartz stones in the total cylinder seal assemblage in FIGURE 6. It is apparent that a marked increase in quartz use occurs with or immediately after the transition to a predominantly wheel-based engraving technology. The introduction of rotary cutting seems to have allowed the working of harder stones on a more routine basis. The introduction of emery abrasive by about 2000 BC is likely to have contributed to these changes; just as the new abrasive allowed the replacement of stone drills in the production of the central perforations of seals by copper drills (see above), it would appear from the present study that it led ultimately to a wider adoption of rotary techniques for working hard stones. Thus the introduction of the engraving wheel was not an isolated discovery but part of a complex framework of technological invention involving the adoption of a new abrasive, a range of new working techniques and, presumably, a matching consumer desire for more hard stone seals. Neither was the adoption of the engravers' wheel sudden: it seems to have been a gradual process, lasting perhaps 200 years, following its inception, probably in the late Old Babylonian period, c. 1750-1595 BC, to its establishment as a principal technique of engraving in Babylonian workshops, c. 1400-1308 BC on seals of Kassite style. Our interpretation of the `tool marks' on the seal intaglios corresponds with the views of Ward (1910) and, particularly, Frankfort (1939) who proposed that `small cutting disks of various sizes' were used `from Kassite times onwards' but `a few instances of the First Babylonian Dynasty are known where the rotating disk' was used.

[Figure 6 ILLUSTRATION OMITTED]

Thus, the results of our study revise the view tacitly accepted in the literature since the mid1970s, that the engraving wheel was introduced about 1500 years earlier, in the second half of the 4th millennium BC. We believe that uncertainty over the date of the innovation of the lapidary wheel arose because the traditional micro-flaking technique of engraving was mistaken for wheel-cutting and that this confusion was compounded by the failure to recognise filing as a distinct technique. The earlier, 4th millennium date for the introduction of the bow-driven lapidary wheel would have placed this innovation alongside the first use of the potters' wheel and solid-wheeled vehicles (Childe 1954). Instead, the later date proposed here correlates more closely with subsequent developments in Mesopotamia of these wheel-based technologies. For example, Collon (1987: 15861) has referred to the replacement of solid chariot wheels by spoked wheels while Courty & Roux (1995; 1998) have focussed on the inception of wheel-throwing from methods of wheel-shaping. The developments in lapidary technology, which led to the invention of a bow-driven wheel and were to establish a range of working methods that remained essentially unchanged until the last century, appear to have been part of a much broader period of technological change.

Acknowledgements. Margaret Sax is immensely grateful to the British Museum Society, especially the Townley Group, for their financial support of the present investigation. The project has benefitted from the guidance of many of our colleagues in the British Museum, particularly from the advice of Ian McIntyre, Department of Conservation, on the practicalities of engraving. We would like to thank Tony Milton and Tony Simpson for their photograghic and graphic work respectively and, also, Sheridan Bowman, Ian Freestone and Andrew Middleton for their constructive comments on the presentation of this paper.

References

BUCHANAN, B. 1970. Cylinder seal impressions in the Yale Babylonian collection illustrating a revolution in art circa 1700 BC, Yale University Library Gazette October: 53-65.

CHILDE, V.G. 1954. Rotary motion, in C. Singer, E.J. Holmyard & A.R. Hall (ed.), A history of technology I: 187-215. Oxford: Clarendon Press.

COLLON, D. 1982. Catalogue of the Western Asiatic seals in the British Museum. Cylinder seals II. Akkadian, Post Akkadian, Ur III periods. London: British Museum Publications.

1986. Catalogue of the Western Asiatic seals in the British Museum. Cylinder seals III. Isin-Larsa and Old Babylonian periods. London: British Museum Publications.

1987. First impressions. Cylinder seals in the ancient Near East. London: British Museum Publications.

In press. Catalogue of the Western Asiatic seals in the British Museum. Cylinder seals V. Neo-Assyrian and Neo-Babylonian periods. London: British Museum Press.

COURTY, M. & V. ROUX. 1995. Identification of wheel throwing on the basis of ceramic features and microfabrics, Journal of Archaeological Science 22/1: 17-50.

FRANKFORT, H. 1939. Cylinder seals. London: Macmillan.

GORELICK, L. & A.J. GWlNNETT. 1989. `Collars' in the holes of Near Eastern cylinder seals, Archaeomaterials 3: 39-46.

GWINNETT, A.J. & L, GORELICK. 1979. Ancient lapidary: a study using scanning electron microscopy and functional analysis, Expedition 22: 17-32.

1987. The change from stone drills to copper drills in Mesopotamia, Expedition 29: 15-24.

MATTHEWS, D.M. 1990. Principles of composition in Near Eastern glyptic of the later second millennium BC. Fribourg: Biblical Institute of the University. Series Archaeologica 8.

NISSEN, H.J. 1977. Aspects of the development of early cylinder seals, in M. Gibson & R.D. Biggs (ed.), Seals and sealing in the ancient Near East, Bibliotheca Mesopotamica 6: 15-23. Malibu (CA): Undena Publications.

ROUX, V. & M. COURTY. 1998. Identification of wheel-fashioning methods: technological analysis of 4th-3rd millennium BC oriental ceramics, Journal of Archaeological Science 25/8: 747-63.

SAX, M. 1991. The composition of the materials of first millennium BC cylinder seals from Western Asia, in P. Budd, B. Chapman, C. Jackson, R. Janaway & B. Ottaway (ed.), Archaeological Sciences 1989: 104-14. Oxford: Oxbow.

SAX, M., J. MCNABB & N.D. MEEKS. 1998: Methods of engraving Mesopotamian cylinder seals: experimental confirmation, Archaeometry 40: 1-21.

SAX, M. & N.D. MEEKS. 1994. The introduction of wheel cutting as a technique for engraving cylinder seals: its distinction from filing, Iraq 56: 153-66. 1995. Methods of engraving Mesopotamian quartz cylinder seals, Archaeometry 37: 25-36.

SAX, M. & A.P. MIDDLETON. 1992. A system of nomenclature for quartz and its application to cylinder seals, Archaeometry 34: 11-20.

WARD, W.H. 1910. The seal cylinders of Western Asia. Washington (DC): Carnegie Institution.

MARGARET SAX, NIGEL D. MEEKS & DOMINIQUE COLLON(*)

(*) Sax & Meeks, Department of Scientific Research and Collon, Department of Western Asiatic Antiquities, British Museum, London WC1B 3DG, England.

Received 25 May 1999, accepted 14 July 1999, revised 4 February 2000.