The ground stone components of drills in the ancient Near East : Sockets , flywheels , cobble weights , and drill bits

Three types of drills are known from antiquity: the bow drill, the pump drill and the crank drill. Each type often included ground stone components sockets, weights and flywheels. However, these components are inconspicuous; on their own they are almost never associated with drills. The result is that the drill is nearly invisible in many assemblages, particularly those of the proto-historic and historic periods, from the Chalcolithic through to late antiquity. In this article I focus on the identification of the possible ground stone components of each of these drill types. The means by which these components were attached or applied to the drill shaft is examined and the way that they related to the rotary motion of drills is laid out. I briefly discuss the historical development of each type, referencing more detailed studies, where available. This study should be seen as a prelude to a more comprehensive study that will test hypotheses by means of experiment and catalogue more completely and precisely the ground stone components of drills that have been unidentified or misidentified in archaeological contexts.


Introduction
Drilling is, and always has been, an essential manufacturing activity.In modern-day industrial societies drilling is practiced infrequently in the daily lives of most people, but in antiquity drilling was a routine action both for the making of fire and in artefact fabrication and decoration.Several studies have investigated drilling apparatus in the ancient Near East, particularly in the Egyptian context (e.g., Petrie 1884;1917;Lucas & Harris 2012: 42-44, 66-71, 104-105, 423-426, 499;Hartenberg & Schmidt 1969;Gorelick & Gwinnett 1987;Moorey 1994: 103-110;Stocks 2003;Saraydar 2012).Stocks (2003) is widely considered the most comprehensive.
If we narrow our geographical purview to the Levant, recent work has focused on the use of the drill in Natufian and Neolithic contexts: Goren-Inbar et al. (2012) investigated clay cylinders that may have functioned as spindles for fire making; Grossman & Goren-Inbar (2007) discussed the use of drills in the quarrying of flint nodules; and Wright et al. (2008) analysed the remains of a bead manufacturing workshop.
Notwithstanding the above researches, the 'ground stone components' of drilling are striking in their absence from excavation reports of Levantine sites from virtually all archaeological periods.One of the few exceptions are the small-find catalogues published by Sass & Cinamon (2006) and Sass (2004a) where ground stone artefacts associated with drilling are at least identified in the excavation reports of work at Tel Megiddo and Tel Lachish, respectively.But even here there is no discussion of the manufacture, use, and context of these stone artefacts.
This paper presents a roster and the criteria for identification of the various ground stone artefacts, which can, and often should, be associated with drilling.Since this volume deals with ground stone, I do not address chipped stone or metal drill bits, an artefact type which has been the subject of other studies (e.g., Gwinnett & Gorelick 1987;1993;Moorey 1994: 103-110;Rosen 1997: 68-71;Endo et al. 2009;Morero 2011).I start with the bow drill, which, in heavy-duty work, entails the use of a cap-stone, or drill socket.We then proceed to the pump drill, which involves a flywheel that is often made of stone.Finally, we look at the twist-reverse-twist, or crank drill, whose weights would have usually taken the form of stone cobbles (in 3 rd -2 nd millennium BCE contexts) or wheels (in 2 nd and 1 st millennium BCE contexts), and whose drill bits would have been of ground stone.

The Bow Drill
The bow-drill is well-known from ethnographic observation and from ancient Egyptian depictions (e.g., Petrie 1917: 39: Gorelick & Gwinnett 1987;Stocks 2003: 23-24, 42-46;Saraydar 2012).The bow drill was probably used for fire-making more than anything else, but it was also an essential tool in craftwork, especially for carpentry, lapidary work, and beadmaking.

Bow drill components
The bow drill is comprised of a spindle, stock or shaft (generally of wood), a bow (which induces the rotary motion of the spindle), and a drill-socket, which allows a broad grip so as to apply downward pressure, creating friction and abrasion at the lower working end (Figure 1).When a fire ember is the goal there will be a notched wooden fireboard at bottom, upon which the tapered end of the wooden spindle is rotated to create an ember.In this case there will be no drill bit.A piece of skin or bark will underlie the fireboard to catch the ember and transport it to the kindling material (demonstrations of the fire-making process can be readily accessed on YouTube by searching "bow drill", see for example, National Geographic Channel 2014).For a demonstration by Denys Stocks of the ancient Egyptian stone coring technique using a bow drill, large stone drill socket and a tubular bit of copper, see: Science Channel 2009.
The heavy-duty drilling required for carpentry, bead-making and bone and ivory work demanded a bit, inserted in the working end of the spindle.In prehistoric periods this would have been of chipped stone (flint or chert), and in the Bronze Age, of copper or bronze (Gwinnet & Gorelick 1987: 15-16;Stocks 2003: 142).It has been pointed out that copper drill bits are probably underreported in excavation reports dealing with sites in the Levant, due to their small size and lack of distinguishing characteristics (Yahalom-Mack 2009: 123, table III:6).In the Iron Age, iron might be used, though this may have been an even later development.Here we are concerned only with the drill socket made of ground stone.Egyptian wall reliefs depicting carpentry show drill sockets clearly (Figure 2).Therefore, we should expect to find them in the archaeological record as well.(Davies 1935: pl. 52).Note the drill socket grasped by the workman at the top of the spindle.

Nomenclature
The stone artefact used to exert downward pressure on the spindle is given different names in the literature: drill cap, capstone, handhold, hand piece, bearing block, or drill socket.Here, I adopt the last term: 'drill socket'.The drill socket can be fashioned of various hard materials: wood, bone, or stone.A wood or bone handhold is perfectly acceptable for producing a fire-starting ember.Unfortunately, wooden examples are bound to be virtually absent from the archaeological record.Bone drill-sockets will show up from time to time (e.g., Sass & Cinamon 2006: fig. 18.30:653).Stone sockets are the most likely to survive and it is these with which we are concerned.Wright's (1992) widely used Classification System for ground stone tools from the Prehistoric Levant lacks a type with the designation of a capstone, drill socket or hand-hold.But her "pebble mortar" (Type 15) and her "potlid" platter (Type 115) are likely to be drill sockets.Shallow sockets would suit fire drills, while deep, robust sockets would be more suitable for carpentry and masonry drilling (Figures 2 and 3).The basalt drill sockets in Figure 3 are from the Iron Age I (ca.1200-950 BCE) levels at Tel Dan, in northern Israel.Item 1 is the smaller, simpler item with the shape of a rounded cone.Item 2 is broken.It has a roughly cylindrical transverse section and two axial depressions.Four longitudinal grooves of similar size were made equidistant around the circumference at the midsection, perhaps to improve the grip.Lubrication (oil) may have been applied at the contact between the butt of the spindle and the socket interior.

Ground stone tools features and typological identification
Drill sockets have been identified by Wright et al. (2008: 146, fig. 14a) in a bead-making workshop at the Pre-Pottery Neolithic C site of Jilat 25 in Jordan.A basalt example from a Late Bronze Age I context has been identified by Sass & Cinamon at Megiddo (2006: 395, 2.19, No. 652) and an Iron Age II example was noted by Sass at Lachish (2004b: fig. 28.13:10).In the Bronze Age report for Lachish (Sass 2004a), no drill sockets were identified.This is probably a case of serendipity.P.G.Dorrell may have identified one (called a "handrest") in the Early Bronze Age levels at Jericho (Dorrell 1983: fig. 231:4).

The pump drill
In the Old World the pump drill is first attested to in the Roman period (Mercer 1975: 212).It is not documented in pictorial representations of the ancient Near East.Whether it was utilised in earlier periods must remain an open question.It is known in the ethnography of indigenous peoples of the Americas, in particular with regard to the Iroquois, who were once credited with its invention (Morgan 1851;Parker 1968:60).But it is noted amongst a number of tribal groups in archaeological contexts, perhaps going back to the Archaic period (circa 8000-1000 BCE; e.g., Hranicky 2011: 736).It is also documented as a tool of premodern northeastern Asian cultures (Hough 1912: 195).

Pump drill components
Pump drills (Figure 4) create rotational friction using a motor principle that differs from that of the bow-drill.A pump drill involves a wooden crossbar perforated at its midpoint, where the drill spindle is inserted into the perforation.A single length of cord is attached to the ends of the wooden crossbar, while also being anchored to the spindle by threading the cord through a perforation at the spindle's top, leaving enough slack to allow the cord to become wrapped tightly around the spindle when spun.Thus the crossbar is positioned tautly toward the top of the spindle.Rotational friction is created, alternately clockwise and counterclockwise, by pumping the crossbar down.
The momentum of the spin rewinds the cord and draws the crossbar back to the upper position.But to acquire sufficient momentum for the rewinding and for the required friction at the spindle point, a flywheel is necessary, positioned under the crossbar, closer to the spindle bit.This flywheel can be of wood or bone, but when heavy-duty drilling is the goal a stone flywheel is used.Like the bow-drill, the point will be either that of the wooden spindle (for making an ember) or a drill bit (of stone or metal), for drilling holes in hard materials.

Ground stone features and typological identification
A stone flywheel will usually be disc-shaped (this is not essential, but will ensure an even momentum).The centre of the wheel is perforated.The diameter of the perforation is a function of the size and weight of the flywheel and the diameter of the spindle (Figure 5).In the archaeology of the ancient Near East, perforated ground stone disks are often thought to be whorls used in the spinning of textiles into yarn.The key to this interpretation lies in the size and weight of the disk, the diameter of its aperture, and in the morphology of the perforation.A spindle whorl cannot be too large and the aperture cannot be too wide.It might be proposed, therefore, that large perforated disks with broad-aperture perforations were not intended for textile manufacture; one should look for other interpretations.The flywheel is an alternative identification.I must reiterate that no hard evidence has been found for use of the pump drill in pre-Roman contexts.But perforated disks may also have belonged to crank drills (see below).Very large perforated disks may have been weights with other purposes, such as digging stick weights (compare with Amiran & Ilan 1992: 74-75), or suspension weights.
As a hypothetical exercise, adopting the above criteria, some examples of perforated ground stone disks which could have served as flywheels are presented below in the section on crank drill weights.Pump drill operation is amply demonstrated on the Internet (see, for example, Winn 2014).

Crank drill or borer components
The crank drill or borer consisted of a wooden shaft with a fork at the base into which a drill bit is inserted (Figures 6 and 7).The fork can be a separate component lashed to the main shaft.At the top of the shaft was a diagonal hand crank, either an integral part of the shaft or attached by lashing.Weights were attached to the upper part of the shaft, at the join with the crank (Stocks 2003: 137-157).The crank drill or borer was chiefly utilised for stonework, with the help of a sand abrasive.In 4 th millennium BCE Egypt the initial rough cavity in stone would have been made by pecking and chiselling and then the crank drill or borer would be put into action.Once copper metallurgy and copper tube corers were introduced in the Naqada II period the twistand-reverse-twist borer with a copper tube bit was employed following the initial shaping (Stocks 2003: 96, 101-105, 133).The resulting cylindrical core was removed with a chisel and the drilling was continued until the desired depth was achieved (Stocks 2003: 122).If a broader cavity was desired, or one that expanded below the tubular neck, a stone bit or borer was then introduced.
In Stocks ' (2003: 128-133) experimental study, three different drill bits seem to have been used: (a) a crescent-shaped bit, most often made of flint or chert, (b) a conical or circular bit with hafting grooves at the top (Figure 8), and (c) a "figure-of-eight" or hourglass form that would be grasped at the waist in the wooden fork (Figure 9).The conical type was used for drilling bores that are generally deeper than they are wide and the figure-of-eight type was used to drill wider cavities, expanding and smoothing those begun by other means (Stocks 2003: 131-132).This is particularly germane for the manufacture of closed stone vessels whose openings are narrower than the deeper part of the cavity.

Ground stone features and typological identification
The ground stone components of the crank drill or borer that we are looking for are (a) the stone weights and (b) the drill bits.Weights could take the form of sand-filled bags, which would be irrelevant to our discussion.But they could also, perhaps preferably, take the form of rounded cobbles of uniform weight and shape (Figure 10).Stone weights will be difficult to identify as such if not found in clear association with the other (mostly wooden) components.The cobbles should be smooth so as not to tear the cords that fasten them to the shaft and crank.Their size and weight must conform to the size of the borer apparatus.Smaller drills or borers for the making of stone vessels, for example, would require spheroid or ovoid cobbles with diameters (or length-width measurements) of 10-15 cm and weights in the range of 1-2 kilograms.
The crank was turned clockwise and then counter clockwise, alternately, to ensure both stability and a cantered coring; hence the term "twist-and-reverse-twist drill or borer" (TRTD, Stocks 2003: 136-144).
Denser stones with high specific gravity would be preferred, such as fine-grained basalt, siliceous limestone or flint (at least in the Levantine context).Stone weights for more massive borers could reach 2-3 kilograms (see, for example, Stocks 2003: figs 5.21-5.22).Cobbles of this sort are perhaps best available as naturally ground river cobbles; but if these are not available, quarried stones can be ground to the appropriate shape.
Another form of drill or borer weight came into vogue in Egypt in the Middle Kingdom and remained popular (together with the double-weighted variety) throughout the New Kingdom.This was a stone ring perforated at its centre through which the drill or borer shaft was passed (Figure 5; Stocks 2003: 132-133, fig. 5.17).It is essentially a flywheel, but placed higher up, just below the crank.It can be a flat ring or domed (the domed form is Gardiner's (1957) hieroglyphic sign U24, see below Figure 12).Drill bits will be manufactured of minerals with a hardness of more than 7 on the Mohs scale.In a Levantine context, fine-grained basalt, siliceous limestone, quartzite, or flint would be suitable.In neighbouring regions various crystalline igneous minerals, such as granite or diorite, might be used, where available, though siliceous limestone may be the preferred material (see the below discussion of Shaykh Sa c id-Wadi Zabayda in Egypt).
The earliest representation of the crank drill or borer appears as a hieroglyph in a 3 rd Dynasty tomb at Saqqara (Firth & Quibell 1935-1936: pl. 93).The earliest naturalistic depictions are found in Egyptian wall reliefs of the 5 th dynasty, such as the tomb of Ty (Steindorff 1913: pl. 134) and the 6 th dynasty tomb of Mereruka (Duell 1938: Part I, pls. 30-31) at Saqqara.It appears thereafter in wall reliefs until the 26 th dynasty (Stocks 2003: 132-136;Lucas & Harris 2012: 425).A wooden model dating to the Middle Kingdom was recovered from the north side of the Teti Pyramid complex (No.J. 45319 in Quibell & Hayter 1927: 40, pl. 24).
More recently, a group of drill bits was reported from the "Lost City of the Pyramids" project (Old Kingdom) at Giza (Tavares 2009).And a 1000+ assemblage of drill bits of five different types was reported at Al-Shaykh Sa c id-Wadi Zabayda near Hermopolis in Middle Egypt (Figure 11; Willems et al. 2009).This site has been identified as a calcite alabaster vessel manufacturing workshop, probably dating to the New Kingdom and the Third Intermediate Period, though it is possible that some of the stone bits date to the Old Kingdom.The large majority of the bits are of silicified limestone.The pattern of wear striations was a key factor in their identification as drill-bits.The five types were determined as follows (and see Willems et al. 2009 for detailed descriptions): 1. Figure-of-eight (Petrie's (1917) "hour-glass borer") type; N = approximately 20; here Figure 11A.
2. Boat-shaped; N = 65; here Figure 11B.This is a newly identified form.3. Vertically oblong; N = approximately 170; here Figure 11C-D.This type shows a wide variety of forms ranging from long and thin (Figure 11C) to long and broad (Figure 11D).
4. Discoidal with flat undersides; N = approximately 10; here Figure 11E.These often seem to be "remodelled" old tools.They are also a newly identified drill-bit form.
5. Bits with shallow hollows; N = approximately 45; here Figure 11F.An additional circa 200 crescent-shaped drill bits of flint were recovered, but these are not within the purview of this article.
As noted above, the crank drill or borer appears to be the origin of the hieroglyphs U24 (Old Kingdom) and U25 (New Kingdom) in Gardiner's (1957) list (Figure 12).They are determinatives for crafts.
For Mesopotamia, Moorey (1994: 56-58) has provided the best summary of the few finds related to the crank drill or borer.He notes the "figure-of-eight" stone drill bits found at Neolithic Jarmo (Moholy-Nagy 1983: 294, figs. 132:7, 140:8), protohistoric Ur (drill sockets too: Woolley 1956: pls.13:U.1640,15:U.14925),though Woolley believed such bits to be associated with the bow drill), and Uruk (Eichmann 1987:110).Woolley identified "…stone borers for hollowing stone vases…common in the Uruk and Jamdat Nasr periods; the drillpoint is a circular stone, flat on the top and curved underneath, and from each side a piece is cut out for attachment; the shaft of the drill would be made of two pieces of wood laid together (or one piece forked at one end …[which] gripped the stone drill head" (Woolley 1956: 14, fig. 5).
More recent analyses of masonry involving stone drill bits -stone bowl making in this case -have been presented by Ciarla & Bökönyi (1985) and Bocutti et al. (2015) for the Early Bronze Age levels at the site of Shahr-i Sokhta in eastern Iran, where stone drill bits were identified tentatively (Ciarla & Bökönyi 1985: fig. 8a: 4c).In the Levant, at least one studyof Neolithic "cupmarks" -suggests a very early use of the crank drill or borer (Grossman & Goren-Inbar 2007).Figure 12.Hieroglyphs derived from the crank drill or borer, Gardiner's (1957) signs U24 and U25.

Concluding Remarks
This paper has described the components and utilization of three different types of drill: the bow drill, the pump drill and the crank drill.This is essentially a summary of previous research, without any intention of complete coverage.More specifically, I have emphasised the ground stone components of each of these drill types.A survey of the literature shows that these somewhat unbecoming stone artefacts are most often either ignored or misinterpreted as either other artefact types or, in the case of cobbles, not considered artefacts at all.Since other drill components are organic and therefore only rarely preserved in archaeological contexts, the misidentification of stone drill parts means that drills are largely missing from spatialfunctional site interpretations.If the interpretations presented here are correct, drills of different kinds were commonplace at almost every site.The author hopes that this contribution will encourage the identification of drills in future field work and that drills will be given their rightful place in material culture analysis.Certainly, a more comprehensive study should be carried out, including experimental work with the aim of defining the specific qualities and metrics of drill components made of ground stone, and a more complete, contextualized cataloguing of such items from archaeological contexts.

Figure 2 .
Figure 2. A wall relief from the Tomb of Rekhmire (Theban Tomb 100) showing the drilling operation (Davies 1935: pl.52).Note the drill socket grasped by the workman at the top of the spindle.

Figure 5 .
Figure 5.A ground stone perforated disk which is a possible flywheel for a pump drill or a weight for a crank drill; from Tel Dan, Locus 8189 (Stratum V [Iron Age I], Registration number 20672/1), weight: 1.63 kg.

Figure 8 .
Figure 8. Drill bit of the conical type, from the Lost City of the Pyramids project, Old Kingdom, Giza, Egypt (Tavares 2009:4, courtesy M. Lehner).It is held upside down in the photo.

Figure 9 .
Figure 9. Drill bit of the "figure-of-eight" type, from the Lost City of the Pyramids project, Old Kingdom, Giza, Egypt (Tavares 2009:4, courtesy M. Lehner).

Figure 10 .
Figure 10.Cobbles from Iron Age I Tel Dan: some of these could have functioned as crank drill or borer weights (the two groupings on the right; the small spherical cobbles on the left are too small).
Figure 11.A drill-bit assemblage collected from the site of Al-Shaykh Sa c id-Wadi Zabayda in Middle Egypt, dating to the New Kingdom (1550-1069 BCE) and Third Intermediate Period (1069-653 BCE), (source: Willems et al. 2009: fig.3; drawings by A. van den Broeck).