Technical Transfers Between Chert Knappers: Investigating Gunflint Manufacture in the Eastern Egyptian Desert (Wadi Sannur, Northern Galala, Egypt)

1.1 Overview of Gunflints Research

Gunflint production in Europe emerged in the sixteenth century with the development of flintlock firearms. This technology, based on the use of a chert piece, was primarily employed for igniting gunpowder and firing guns (Austin, 2011; Barandiarán Maestu, 1974; Chelidonio, 1987; Ciarlo et al., 2019; Dolomieu, 1797; Edeine, 1963; Emy, 1978; Emy & de Tinguy, 1964; Kenmotsu, 1990; Kent, 1983; Skertchly, 1879; Witthoft, 1966). The increasing use of “flintlocks” during the seventeenth century contributed to the spread of this technology, gradually reaching many parts of the world under European influence. The distribution of gunflint production followed the adoption of firearms, propelled by both armed conflicts and established trade routes (Austin, 2011; Ciarlo et al., 2019). While this technology had long been well practiced, its industrial character gradually emerged as it became part of the military supply chain controlled by the political powers of Western countries (Emy, 1978). This industrial process (Caliste & Carnino, 2022; Jarrige, 2021) is primarily characterized by massive, standardized production, which involves specialization and segmentation tailored to military or commercial needs. France and England have been considered the major providers through military conflicts and commercial markets for nearly three centuries worldwide (Ballin, 2012; Emy, 1978; Kent, 1983; Whittaker & Levin, 2019; Witthoft, 1966). However, the process of manufacturing gunflints may have spread unevenly across different regions, making certain productions more difficult to identify in the historical record (Whittaker & Levin, 2019). Given this scarcity, evidence of autonomous production, as revealed through collections of military objects or archaeological records, is of particular interest.

More broadly, two major issues have traditionally been observed from historical and archaeological records at different scales: on one hand, viewing the gunflint as an isolated artifact, and on the other hand, considering the whole manufacturing process. The first involves identifying gunflints that have been found in the archaeological context, as chronological markers that highlight their typology, and the historical dynamics related to their context of discovery (Hamilton, 1964; Witthoft, 1966). Beyond offering limited chronological data (Hamilton, 1980; Hamilton & Emery, 1988; Hamilton & Fry, 1975), various inferences can be drawn from raw materials, morphology, and style (Austin, 2011). The type of raw materials used is commonly accepted as evidence of provenance (Werra et al., 2019); however, some materials can be confusing macroscopically (Austin, 2011; Ballin, 2012; Durst, 2009; Horowitz & Watt, 2019), and there may be segmentation in the production chain across different countries (Emy, 1978, p. 119; Witthoft, 1966, p. 41). Morphology and style define an archaeological typology and differ from the original typologies adopted by factories for commercial needs. Gunflint types are primarily interpreted through chronological criteria, indicators of manufacturing techniques, evidence of specific weapons, and/or resulting as different technical processes associated with the country of production (Ballin, 2012, 2014; Buscaglia et al., 2016; Chelidonio, 2013; De Lotbiniere, 1977, 1980, 1984; Kenmotsu, 1990; Luedkte, 1999; Witthoft, 1966). This has been traditionally seen with the so-called French and British styles, which differ mainly in the transition from “D-shape” to “rectangular” forms (Austin, 2011; Ballin, 2012, 2014; Barnes, 1937; De Lotbiniere, 1984; Whittaker, 2001). Although none of these criteria is fully diagnostic (Ballin, 2012), they are useful for understanding an almost universal mode of distribution, often linked to major historical events. In this sense, gunflint production also documents processes of military industrialization driven by centralized policies and linked to major armed conflicts or commercial trade.

The second issue, which is more challenging in the absence of remains related to the production chain, concerns understanding the skills associated with the technical processes and the entire chaîne opératoire (Lemonnier, 1986), from production to use and discard. This has primarily been applied to French and English factories demonstrating industrial production from the end of the nineteenth to the twentieth centuries (Barnes, 1937; Clarke, 1935; De Mortillet, 1908; Dolomieu, 1797; Edeine, 1963; Evans, 1872; Lovett, 1887; Salmon, 1885; Schleicher, 1910, 1927; Skertchly, 1879), focusing in particular on labor organization (namely, segmentation or the gendered division of labor) and the working conditions for craftspeople within an ethnographic or historical perspective (Clarke, 1935; Emy, 1978; Lovett, 1887; Skertchly, 1879). This study of the production chain has also been developed in archaeological contexts such as in Italy, France, Spain, and Scotland (Ballin, 2012; Barandiarán Maestu, 1974; Chelidonio & Woodall, 2017; Weiner, 2017). Furthermore, the study of manufacturing techniques and methods has revealed a technical convergence in the field of flintknapping, particularly with respect to prehistoric times, highlighting the retroactive nature of the lithic technologies (Barnes, 1937; Clarke, 1935; Emy, 1978, p. 9; Skertchly, 1879). This anthropological dimension, prominent in early surveys, has facilitated an understanding of the phenomenon and its various stages.

Here we are investigating these two issues through the archaeological record, focusing on both the gunflints and the manufacturing process. This study is based on the discovery of a gunflint factory in the Egyptian eastern desert (Northern Galala), dated from Mehemed Ali’s reign in the first half of the nineteenth century. This discovery, initially identified by German botanist A. Schweinfurth (1885) at the end of the nineteenth century, appears to be the earliest evidence of gunflint production in this geographical area. According to this record, historical sources and archaeological data from two sites, it appears to have been exploited for the military supply of Mehemed Ali’s armies between 1810 and 1830. These quarries, which include extensive extraction and manufacturing areas, provide an exceptional example for documenting the production process. They are particularly well preserved due to the dry conditions and their unique desert location, which is a significant advantage. It allows us to precisely describe the technical traditions observed in the archaeological context, placing them within the broader historical framework of Egypt during the reign of Mehmed Ali.

1.2 Gunflints and the History of the Egyptian Armies

From the beginning of the nineteenth century, Egypt’s history is marked by several conflicts, including the defeats of the French armies (1801–1803) and the persistent struggles related to the rise of Mehemed Ali between 1803 and 1805 (Al-Sayyid Marsot, 1984; Fahmy, 2009). This rise to power gradually led to the independence of Egypt from the Ottoman Empire and the renegotiation of alliances with Western empires. The political and economic development initiated at the beginning of Mehemed Ali’s reign is notably characterized by the rise of a new army (Douin, 1923; Dunn, 1993; Fahmy, 1997; Farhi, 1972; Guémard, 1836; Gouin, 1847) as a tool for political control (Fahmy, 1997). The state’s mastery over armies appears to align with the broader advancements in military armament observed within the Ottoman Empire from the eighteenth to the nineteenth centuries (Grant, 1999; Krause, 1992). In this military upgrade, firearms were initially imported from Europe before local production began. This influx of military goods was accompanied by the transfer of technical skills to the armament industry. This was facilitated by the intervention of mercenaries from Western European countries, preceding the autonomous development of the new military industry (Douin, 1923; Dunn, 1993; Gouin, 1847; Guémard, 1936). Regarding the gun supply, French muskets were first imported in 1821 and 1824 (Guémard, 1936, p. 118, 131) and powder from England. Gun, powder, and cannon manufactures have been developed in Cairo since the 1820s (Guémard, 1936; Planat, 1830; Rochefort Scott, 1837; Russell, 1835), using the expertise of European mercenaries. The French flintlock model “1777 modifié An IX” was manufactured at Cairo’s Citadel and adopted by the new regular troops (Bowring, 1840, p. 144; Guémard, 1936, p. 141; Planat, 1830, p. 87, 350; Rochfort Scott, 1837, pp. 164–166). This upgrade had to adapt to new political needs, consolidating the economic independence of the military system and leading to the establishment of autonomous factories managed by foreign advisers.

Until recently, there have been few mentions of gunflint production in Egypt (Clarke, 1935). In addition to a site mentioned in the Galala, a manufacturing site has been reported in the village of Abu Rawach (Schweinfurth, 1885). Historical sources indicate that gunflints were used for flintlocks and pistols by Mamelukes and armed troops in Egypt during the first half of the nineteenth century (Bardin, 1851). However, its supply was likely limited by opportunistic self-procurement and manufacturing (Bardin, 1851; Emy, 1978, p. 144). Although little is known about the supply of gunflints to the Ottoman armies, it is likely that Mehmed Ali recognized their strategic importance and took measures to ensure they were adequately supplied. The creation of new troops through conscription increased the demand for ammunition and gunflints, as the supply of guns was adapted to meet this need (Guémard, 1936, p. 152). Here, gunflint production during Mehemed Ali’s reign represents new evidence for the industrialization process involved in the Egyptian army (Fahmy, 1997, 2009). Nevertheless, the evidence of such standardized production raises questions about the political initiative and the introduction of new skills.

3.1 Description of the Sites and Assemblages

The first site WS003 is located on the slope of the Wadi Umm Nikhaybar, a tributary of the Wadi Sannur. The site is composed of an extraction area at the bottom of a wadi (dry valley or channel outside of the rainy season), while the workshop area is located nearby, up on the plateau (Figures 1 and 2). This gunflint factory has been confused with some workshops from the Pharaonic periods (Harrell, 2024) since it is located a few hundred meters away from a miners’ camp currently being excavated and dated to the Old Kingdom (Briois et al., 2021; Briois et al., 2024). The extraction area appears to follow the natural wadi bed, where a quarry face and spoil piles can be observed. These piles consist of regular limestone blocks resulting from the exploitation of chert bands. The manufacturing area is located just above, near a collapsed building interpreted as a fort from the Ramessid period, based on pottery artifacts identified in the vicinity (Harrell, 2012, 2024, pp. 392–394). The lithic remains are observed both inside and around the site, covering an extensive area of approximately 600 m² on average and extending onto the wadi flank (Figure 2). Although some parts of the area and the building have recently deteriorated, several concentric heaps can still be seen within this waste slick, indicating the presence of different workstations. Thirty-two of these workstations can be identified. Besides lithic remains, fragments of modern pottery and long-stemmed tobacco pipes are observed within the manufacturing area or its vicinity. Two lithic assemblages have been considered among the 32 concentrations observed in this area. The first (WS003.9) corresponds to a small lithic accumulation measuring 3 m long by 2 m wide, divided into two parts and located on a natural slope among other similar concentrations. The lithic remains include cores, flakes, fragmented and entire blades, chunks, and various gunflints. The second assemblage (WS003.15) is part of aggregated concentrations, located 80 m away near the Ramessid fort and the main accumulation area. The lithic remains in this assemblage consist only of fragmented blades, chunks, and gunflints, while the other associated concentrations are composed of cores, entire blades, and flakes. An additional assemblage of 34 gunflints was collected through a survey around the Ramessid fort within the manufacturing area and added to the other assemblages to provide a more significant sample.

The second site (WS004), located on the plateau 2 km west of the first, consists of an extensive workshop area that partially covers a mining trench and some lithic remains from a pharaonic quarry. Several loci can be identified: the main excavated area, resulting from the quarrying activity and covering approximately 300 m², is almost entirely filled by a large spoil heap and includes a quarry face between 2 and 3 m deep (Figure 2). This quarry is partially surrounded by a vast lithic accumulation similar to site WS003, which corresponds to the manufacture area. A second excavated area was also observed, along with empty spaces behind the spoil heap that suggest the presence of dwelling structures. Some fragments of modern pottery and evidence of old fires on the ground (combustion marks) support this hypothesis (Figures A3 and A4). A short distance from the main knapping area, separated by a small wadi, there is a second zone consisting of three concentrations of lithic remains arranged in a sub-circular pattern. Some depressions noticed in their centers, along with the circular shape of certain concentrations, suggest the presence of dwelling structures or different workstations (Beauvais, 2024). The lithic concentrations observed in the workshop area consist of debitage waste such as flakes, cores, blades, and chunks. The two assemblages collected in the main sector are part of overlapping clustered accumulations. One assemblage (WS004.7) corresponds to a 1 m-diameter concentration of 153 lithic pieces, including cores, blades, and flakes (Figure A5). The other one (WS004.8) involves a selective collection of 49 blades, which are part of a larger lithic concentration (Figure 3, no 2).

Figure 3

Key features of the site. (1) Pile of discarded lithic cores (photo B. Midant-Reynes). (2) lithic blades forming part of a major concentration (WS004.8), indicative of the artifact density on the site (photo P.-A. Beauvais).

3.2 The Lithic Remains

The two manufacturing areas reflect the same spatial organization, with lithic concentrations that are either clustered or isolated, indicating a segmented manufacturing process. The four assemblages were chosen because they provide varying levels of information regarding the types of lithic remains sorted from their technological characteristics (Table 1). The fact that this production appears to be standardized in terms of the nature of the products and apparent spatial organization allows us to define the chaîne opératoire based on other industrial contexts, even though they are not contemporaneous. We used Clarke’s (1935) description, applying terminology derived from the English mining contexts. The “quartering” stage corresponds to the volume preparation, “flaking” refers to blank production and “knapping” is the final fabrication step. The two assemblages from site WS004 are exclusively devoted to extraction and the first two stages and are analyzed to describe these processes, while the assemblages from site WS003 allow us to describe all three manufacturing stages. Two technical models of gunflint manufacture, mainly based on the observation of the cores and the end products, are identified within the chaîne opératoire: one, present at both sites, produces blades, while the other that produces flakes is only attested in the WS003 site (Table 1).

The first category of lithic remains found at both sites corresponds to the tabular or ovoid nodules, ranging from about 20 cm to 1 m in length, that are extracted from the limestone’s bench and transported to the manufacturing areas. Blocks are parts of the original nodules that are then divided into “quarters” by hard hammer percussion. These quarters, averaging 30 cm in length, have asymmetrical morphologies and display impact marks and large conchoidal fractures. They are often found in piles with blade cores at the WS004 site (Figure 3), near the extraction area, where they appear ready for further exploitation. These block fragments provide an adequate volume of material for the blade and flake reduction processes that follow.

3.2.1 The Cores

Primary observations regarding blade cores within the sites are based on the WS004.7 assemblage, which comprises 5 blade cores, 92 blades, and 52 flakes and fragments. Three cores refitted appear to originate from the same nodule divided in “quarters.” Although a significant portion of the products is missing from this assemblage, these refits reveal characteristics associated with the first two stages, “quartering” and “flaking” (Figure 4). The blade cores observed in both WS004 and WS003 sites display different morphologies, which depend on the original volume or correspond to different steps in the reduction process. The two main morphologies both combine an orthogonal platform with a unidirectional debitage surface. The first morphology, an octagonal prism, results from enveloping or full exploitation of the core volume. The second, flat cores with a natural or knapped ventral face and a dorsal knapping surface, result from semi-enveloping exploitation until the core is exhausted (Figures 4 and 5). The flat platform is usually the negative of the ventral surface’s knapped “quarters.” The refit from this assemblage shows that the initial “quarter” volumes have been knapped using the length, width, or thickness to produce a large number of blades, even if only the best are selected.

Figure 4

Lithic refits from WS004.7. (1) refit showing two quarters divided from the same nodule. (2) blade debitage on the first “quarter”; (3) core preparation on the second quarter. (4) exhausted core after blade debitage (refits P.-A. B. and photos B. M.-D).

Figure 5

Cores from WS003. (1) Blade core exhibiting a flat morphology. (2) Flake core showing centripetal removals (short and elongated flakes) (photo B. M.-R.).

Flake cores that depend on an autonomous reduction process have been found in four lithic concentrations, including WS003.9. The main morphologies correspond to flat or bi-pyramidal discoid cores, exhibiting centripetal flake removals. The blanks selected for reduction are typically block fragments or large flakes, with peripheral preparation leading to a discoid morphology. Debitage from these cores is multidirectional and produced using direct percussion, resulting in both elongated and short flakes (Figure 5). The shorter flakes are thick and removed using an internal motion, where the core is struck closer to the center, while elongated flakes are detached through a marginal motion, where the core is struck closer to the edge. The alternating removal of these two types of flakes ensures the maintenance of the core and the presence of flat surfaces, which are intentionally sought after.

3.2.2 The Blanks

Two categories of produced blades have been observed at both sites: fragmented blades showing segmentation fractures at the WS003 site and entire blades at the WS004 site (Table 2). The entire blades are additionally assigned to different categories, including cortical or partially cortical blades, as well as hinged or plunging blades with partially cortical ends (Figure 6). Based on a sample of 49 blanks from the WS004.8 concentration, the average size is 7.3 cm in length (σ = 1.8), 2.5 cm in width (σ = 0.73), and 0.76 cm in thickness (σ = 0.34). The general morphology corresponds to a straight profile with one or two central ridges and a rectilinear delineation; the distal part is slightly convergent or plunging. The proximal part exhibits a flat platform without preparation and an impact point with a complete circular ring crack, sometimes containing iron residues. The ventral face displays an overhanging platform with a prominent bulb in clear relief, bulbar fissures, and cracks, but without lipping (Figure 7, nos 8 and 9). These characteristics indicate the use of metallic hard hammer percussion. The estimated angle between the platform and the knapping surface averages around 90°. Two technical accidents are associated with the percussion: the first is a Siret fracture, which splits the blade into two parts and is often accompanied by a bulbar scar; the second is the loss of the overhang on the dorsal face of the proximal part (Figure 6). The two accidents result from a mechanical process caused by the percussion tool and the force applied by the craftsman. Blade debitage produces a large number of blades, from which the most regular ones are selected for the next step. The reduction process operates through unidirectional debitage, where each blade is produced without preparation, overlapping the previous ones in an enveloping manner. Based on the WS004.7 assemblage, we interpret the absence of most of the blades as the result of a selection process, whereas the shorter and more irregular blanks (hinged, plunging blades, curved profiles) are discarded and can be refitted later.

Table 2

Technological sorting and frequencies related to different lithic remains in a selection of the assemblages

	WS003.9		WS003.15		WS004.8	WS004.7	Total
Technological category	N	% (N)	N	% (N)	N	N	N
Blade core	1	0.1%	—	–	—	5	6
Flake core	1	0.1%	—	–	—	—	1
Flake	132	18.5%	85	2.8%	—	51	268
Flake fragment	84	11.7%	22	0.7%	—	1	107
Debris	52	7.3%	430	14.1%	—	—	482
Entire blade	1	0.1%	8	0.3%	41	92	142
Bladelet	19	2.7%	25	0.8%	8	4	56
Fragmented blade (proximal)	75	10.5%	567	18.6%	—	—	642
Fragmented blade (mesial)	52	7.3%	234	7.7%	—	—	286
Fragmented blade (distal)	69	9.7%	208	6.8%	—	—	277
Retouch flake	207	29.0%	1355	44.4%	—	—	1,562
Gunflint preform	3	0.4%	97	3.2%	—	—	100
Gunflint	19	2.7%	24	0.8%	—	—	43
Total	715	100%	3055	100%	49	153	3,972

Figure 6

Blades from two different lithic assemblages (WS004.1 and WS004.7). (1 and 5) loss of the overhang on the proximal part of the blade (photo B. M.-R.).

Figure 7

Blade fragments discarded after segmentation. The white circles indicate the location of the impact points. (1–3) Distal parts showing typical impact point locations. (4, 5, 7) Mesial parts, broken and segmented. (6) Mesial part with two impact points. (8 and 9) Proximal parts (photos B. M.-R.).

Flake blanks at the WS003 site are mainly produced during the flake reduction sequence but are also generated during blade core preparation. The proportion of flakes is well documented for WS003.9 where both types of debitage are attested. The presence of blanks is difficult to distinguish among the discarded flakes. However, it is possible to mentally reconstruct the knapping process using unfinished gunflints and flake cores in the vicinity. Flakes of sufficient size and thickness for gunflint production are detached using direct percussion in either an oblique or frontal motion from an angulated platform (estimated angle around 50°–60°), formed by the intersection of the two surfaces (Figure 5). The same platform stigmata and bulb surfaces observed in blades are present on the flakes, and the percussion negatives are highly marked on the cores. The short flakes exhibit a flat dorsal surface – originally the ventral face of the core – and a sub-triangular cross-section. This flat dorsal surface becomes the ventral face of the finished gunflint, while the retouching process shapes the former ventral face of the flake, which retains the bulb of percussion. In this model of flake debitage, the shorter flakes are selected as blanks for gunflints (Barandiarán Maestu, 1974; Chandler, 1917; Weiner, 2017). This technique closely resembles those used to produce gunspalls – or “wedge-shaped (Witthoft, 1966)” – which some authors describe as “flake-based” gunflints (Ballin, 2012). We interpret these flakes as intentional blanks produced for gunflint manufacture. However, craftsmen may have selected a wide range of products – both short and elongated flakes – for the retouching process.

3.3 Gunflints and Their Fabrication Process

The final step, observed only at the WS003 site, involves the segmentation of blanks until the final retouch. The description of the gunflints is based on the WS003.9 and WS003.15 assemblages, as well as a selected set of samples collected from the manufacturing area (Table 1). Two main types are distinguished, based on the nature of the blank and the retouch mode.

The first type of gunflint widely identified in all the lithic concentrations is a rectangular piece made from a mesial blade fragment. The lithic wastes consist of blades and flakes fragments as well as short flakes resulting from the retouch process within the WS003 site. Both assemblages primarily consist of these items, especially fragmented blades – proximal, mesial, and distal parts – that mostly show signs of segmentation (Figure 7). The straight or oblique fractures observed on the blade fragments are associated with a circular crack or an incipient demi-cone located on the dorsal or ventral surfaces, along the central debitage axis of the blank (Figure 7). These stigmata result from the percussion impact of a metallic tool. The circular crack represents the negative stigmata of the percussion, while the incipient demi-cone is the positive stigmata, caused by the counter-blow when the blank is held against an anvil (Barnes, 1937). The absence of these marks on some blade fragments may indicate incidental breakage during blade debitage or indirect breakage during the segmentation process. The frequency and morphology of these marks likely depend on factors such as the force applied by the knapper, the force transmitted back by the steel, and the type of hammer used. The position of the diagnostic marks on the dorsal or ventral faces helps determine how the blank was segmented (Barandiarán Maestu, 1974; Barnes, 1937). When one piece is produced from each blank, only the distal and proximal fragments remain; when two pieces are produced, the short mesial part – marked with double percussion – is discarded (Figure 7). In the WS003.15 and WS003.9 assemblages, blades mainly produce one piece since there is a low presence of short mesial parts showing double impact fractures. The mesial parts observed in the assemblage are discarded because they are either too short or due to irregular profiles or fractures. The wide array of short flakes associated with these fragments, ranging 0.5–2.5 cm, corresponds to the final stage of manufacturing (Figure 8). For the WS003.15 assemblage, the retouch flakes are removed in a direct motion with an anvil percussion technique; they show a surface on the ventral face and counter-blow marks, indicating the use of a steel (Barnes, 1937).

Figure 8

Diagnostic flakes and chips associated with the retouch process (1–6). Unfinished and broken rectangular gunflint pieces (7–10). D-shape gunflint (11) (photos B. M.-R. and P.-A.B.).

The first type of gunflints consists mostly of unfinished and finished pieces, featuring one or two arises, two or three retouched sides, and one or two leading edges with marginal and regular inverse retouch. These discarded pieces represent various stages of retouch, ranging from preform to finalized states. Our focus is on unfinished gunflints displaying initial and interrupted retouch and finished pieces exhibiting defects, irregular shapes, or breakage during the retouch process (Figure 8). On these pieces, the active part is the leading edge, corresponding to the right or left edge of the blade blank. The presence of one or two leading edges depends on the degree of retouch applied but could also define an independent category (Table 3) based on specific functional purposes. These pieces may offer two active edges, potentially extending the lifespan of the object. Based on a sample of 29 rectangular finished gunflints, the average size of these pieces ranges 2.63 cm (σ = 0.33) in length, 2.4 cm in width (σ = 0.31), and between 0.71 in thickness (σ = 0.15). A comparison between the complete blades and finished products reveals overlapping width/thickness dimensions. No distinct clusters were observed regarding the morphometry of these pieces (Figure 9).

Figure 9

Graph showing the dimensions of the blades compared to the dimensions of the gunflints (length/width; width/thickness).

Although the produced gunflints could have been highly standardized, the discarded pieces are considered non-conforming and unsuitable for further use. The gradual variation in dimensions of these final products suggests adaptation for different types of firearms, such as muskets and pistols, particularly for the smaller rectangular gunflints ranging in average, 1.9 cm (σ = 0.16) in length, 1.8 cm in width (σ = 0.16), and between 0.56 in thickness (σ = 0.13). In terms of typology, the rectangular gunflint (N = 54) is traditionally associated with British style, in contrast to the convex “D-shape” characteristic of the French style seen in a limited range of pieces (N = 2). However, this assumption remains unclear and does not lead to definitive conclusions regarding the influence of style and expertise developed at this factory.

The second type is a rectangular gunflint with irregular edges, shaped by invasive retouch and produced from flake blanks (Figure 10). These pieces, found only in certain concentrations, are shaped using unifacial techniques and exhibit scars from a previous generation of removals that were since discarded. For these pieces, the ventral face typically corresponds to the dorsal face of the flake blank, while the blank’s ventral face has been reworked through invasive retouch (as per Barandiarán Maestu, 1974). This can be observed in unfinished gunflints, where the old dorsal face – now the gunflint’s ventral face – displays centripetal removals (Figure 10). However, invasive retouch may also have been applied directly to elongated flakes, as suggested by some examples. The lithic wastes resulting from the retouch process consist of short shaping flakes with low thickness, flat butts, and dorsal surfaces displaying unifacial removals. These small flakes reveal an invasive retouching technique applied either unifacially or bifacially (Barandiarán Maestu, 1974). The retouching process appears to have been performed using direct percussion, as no evidence of counter-blows has been identified. The average dimension of these pieces ranges 2.9 cm in length (σ = 1.6), 2.6 cm in width (σ = 1.3), with an average thickness of 0.8 cm (σ = 0.5). The orientation of these gunflints relative to the active part is problematic, as the edges are equally worked (Barandiarán Maestu, 1974). The frequency of this type of gunflint is considerably smaller than the blade-based first type, although some workstations show evidence for combined production of both models (e.g., WS003.9; Table 3). The shaping technique could, in theory, correspond to strike-a-light flints production, as observed in certain contexts within the Mediterranean basin (Evans, 1887). However, based on the morphology and average dimensions of these pieces – especially when compared to blade-based gunflints (Figure 9) – and considering the lack of comparative morphometric data on strike-a-light flints, we argue that these pieces represent a distinct method of gunflint manufacture involving invasive retouch. This retouch technique is already attested in the Mediterranean basin, notably in Spain (Barandiarán Maestu, 1974). While the frequency of certain gunflint types may vary depending on the area studied, as well as selective discard processes during production, the primary manufacturing focus appears to have been on rectangular, blade-based gunflints.

Figure 10

Gunflints with invasive retouch. (1) Preform showing unifacial shaping on a flake blank. (2 and 3) Shaping flakes discarded during the retouch process (drawing P.-A. B.; photo B. M.-R.).

The coexistence of both blade-based and flake-based gunflints at the site raises questions about the specific tools used for knapping and retouching. During the initial stages of production, the same tools may have been employed for both blade and flake debitage. A heavier splitting hammer was likely used for “quartering” the blocks, while a double-pointed hammer – commonly used in France (Barnes, 1937; Emy, 1978; Weiner, 2017) – may have been adopted by craftsmen specifically for blade production. For retouching, a different metallic tool – such as a disc hammer or a hammer made from a file – was used in Western Europe to strike the piece against a steel anvil, typically set in a wooden support (Barnes, 1937). In contrast, techniques documented in the Mediterranean basin at the end of the nineteenth century involved the use of the same angular, square-pointed hammer for both flaking and shaping via direct percussion (Evans, 1887).

In this context, we might expect different tools to have been used for producing blade-based rectangular gunflints compared to those shaped using invasive retouch. The retouching technique for blade-based gunflints involves a hammer-and-steel setup, while invasive retouch likely relied on direct percussion, possibly using a small hammer. These apparent differences in tool use suggest the presence of two distinct technical traditions at the site.