Grotte de l’Abbé is located on the left bank of the Tardoire river, on the townland of Vilhonneur, department of Charente, France (Fig. 1). In this area, the geology consists of Bajocian (Jurassic) age limestone, and yields unique features in the form of a complex karstic system (i.e. the “Karst of La Rochefoucauld”). Along the Tardoire river, between the towns of Montbron and La Rochefoucauld (a length of about ten miles), this karstic system contains a large concentration of caves and deep underground galleries. At about the center of this particular geological feature, the town of Vilhonneur is punctuated by several “mounds” where numerous caves are located. Some of these mounds hold meaningful names for prehistorians, such as Rochebertier, Le Bois-du-Roc, and Les Garennes (Airvaux et al., 1999, Balout, 1959, Debénath, 2006 and Debénath, 2014).

Indeed, archaeological pioneers (e.g., Jean-Louis Fermond, Arthur de Maret), followed by more experienced researchers (e.g. Gustave Chauvet, Henri Breuil), have invested in this region since the 1850s and this interest continued into the early 20th century. These researchers are thus responsible for revealing the prehistoric wealth of this region by carrying out more or less systematic explorations of the landscape and thus discovering numerous Palaeolithic sites (see Breuil, 1912, Chauvet, 1910, Fermond, 1873 and de Maret, 1879). Notable sites discovered include, at one location of the karstic system known as Rochebertier: Grotte de l’Ammonite, Grotte du Sureau, and most notably the famous cave of Le Placard (see Clottes et al., 2010). Another, known as Bois-du-Roc, yielded – among others – the two caves of Les Fadets, Abri André Ragout, and Abri du Chasseur (Balout, 1957, Balout, 1958 and Balout, 1959).

The hill of Les Garennes, wherein Grotte de l’Abbé is located, was little known until recently. The site of Grotte de l’Abbé is currently inventoried as “Grotte des Laurines” in the Regional Archaeological Services database, though both these names correspond to a single and same site: a complex network of galleries, with various areas of contact with the open air, two of which bear the names in use (Fig. 2). Unfortunately, while this site was discovered and explored quite early, no information regarding the excavators and their investigations during the 19th century and early 20th century is available.

For the past ten years, however, a group of spelunkers has extensively explored Les Garennes, and it was, in fact, their efforts which lead to the December 2005 discovery of the Aven du Charnier, a Gravettian painted cave (Boulestin et al., 2006 and Debénath, 2014: 266–267; Henry-Gambier et al., 2007). Their activities also allowed them to re-locate Grotte de l’Abbé in 2009 and, by some extraordinary luck, to also identify the supposed test pit of the old “excavations”. A preliminary observation of this heavily eroded “hole” revealed sediments in place rather brechified along the cave walls but quite loose in the middle of the gallery with a slight slope toward the south-east, and disturbed by burrows dug by badgers or rodents. At this initial stage of exploration it was difficult to identify a clear stratigraphy, or differentiate distinct litho-sedimentologial layers. Numerous bones and teeth (mandibles, scapulas, ribs, long bones) of hyena, bear, bison, (etc.), as well as several spear points fragments were observed scattered in the cut. These various lines of evidence gave us the feeling of a sporadic presence of humans associated with a carnivore den, on top of which the carcass of a brown bear was naturally “dragged” down. Since this excavation area is no larger than a few square feet, there appears to be quite a bit of sediment left untouched. Thus, there is undoubtedly the potential to conduct new excavations within this same area of the cave in the future.

Both HRR halves were present in the form of several fragments and splinters, covered with compacted sediments. It was therefore necessary to first conduct (by one of us, D.A.) the long and delicate work of artefact cleaning and preparation of these pieces before they could be glued back together to form these quite remarkable artefacts. During cleaning, we realized with surprise that these two HRR components shared very similar morphologies and engraved patterns. It quickly became obvious that the two pieces perfectly matched with each other in terms of their width at several places along the shaft, the character of the oblique lines engraved on the inferior surfaces, and the intricate incised pattern on the superior surfaces. These consistences suggest that they once made up the two halves of a single composite tool, in this case, a projectile point.

The Grotte de l’Abbé HRR consists of a mesial and a mesial-proximal fragment, which when matched together (by matching shaft width and the engravings on the superior surface; see Fig. 4) measure 262 mm in total length – around 100 mm shorter than the near complete example recovered from Isturitz by Passemard in 1913 (Passemard, 1916) and 100 mm short of the La Vache example (which measures some 360 mm) (Feruglio and Welté, 2003). While the proximal extremity of the mesial fragment (GdA 19) exhibits a clear splinter fracture caused by impact, the distal extremities of both this piece and the mesial-proximal fragment (GdA 20) appear to have fractured post-depositionally. The proximal (base) extremity of GdA 20 is formed by two wide facets on its superior surface, creating a roughly triangular (or sub-circular) section. This type of conical base form is that most commonly identified on HRR proximal fragments (Feruglio, 1987).

The superior surface of each of the two halves is covered with a carefully executed engraved design. This design begins (as seen on GdA 20) at the proximal end with what might be described as a pair of schematic projectile points or even (perhaps) people: a vertical line/s with a shorter oblique line off each of its left and right sides, topped with a rounded area formed by multiple closely incised oblique lines (Fig. 3 and Fig. 4). These hatched circles are also seen at the proximal end of GdA 19 (Fig. 4). The next (mesial) section of decoration is delineated by two parallel horizontal lines which run transversally across the shaft. What follows is a interwoven line formed by small chevrons. While the distal extremity of GdA 20 terminates in the second of the oval shapes formed by these marks, it can be seen on GdA 19 that the second oval terminates in what appears to be a stylized ‘fish’ (eel?) with gaping mouth (Fig. 4). Within these ovals are several short, oblique lines. The following section of decoration is less clear, though another oval/fish seems likely. The decoration then continues with three deep parallel curved lines before terminating at a post-depositional fracture. It is these decorations which allow us to confidently match the two halves as belonging to the same implement.

HRRs are the most highly decorated of all the Magdalenian osseous projectile points, with a range of figurative (primarily horse and reindeer motifs) and geometric designs (parallel lines, curved lines, chevrons, zig-zags, circles) observed (Feruglio, 1987). For example, analysis of the extensive La Vache (Ariège) collection of HRRs found that 68% exhibited engraved designs of one type or another (located on the superior face of the implements), though almost half (29%) of these displayed only one or two grooves running parallel to the point axis (Feruglio, 2003). In this dataset, 18% were found to display a unique geometric motif, 11% a figurative motif, and 9% a twisting pattern motif similar to that found on the Grotte de l’Abbé point (Feruglio, 2003). Further similarities of the Grotte de l’Abbé point with HRRs found elsewhere include the oval motifs constructed out chevrons and surrounding several oblique lines (seen at Mas d’Azil; Feruglio, 1987), while the use of two parallel horizontal lines to separate the motifs was found on 7% of the La Vache examples.

Previously, the most outstanding examples of engravings found on HRRs include the series of spiral designs recovered from Isturitz (Pyrénées-Atlantiques), while Mas d’Azil (Ariège) produced several excellent examples of reindeer motifs. These artefacts, however, are singular pieces lacking their complementary half, while the more complete points (where both halves were recovered) have featured less complex designs (such as the Isturitz example which exhibits paired curved lines). The Grotte de l’Abbé HRR with its extensive decorations is thus unique.

Also recovered were three bevel-based unbarbed points (sagaies) of exceptional length, distal, mesial, and proximal point fragments from additional HRRs and (likely) bevel-based points, blanks for point (or other antler tool) manufacture, burnisher (lissoir) fragments, and a single, perforated red deer tooth ornament.

Two fragments of additional HRRs (GdA 03 and GdA 09) were recovered alongside the matched and decorated pair, though in each of these cases no decoration was identified on the superior surfaces (Fig. 5B). These pieces include a mesial and mesial-proximal fragment respectively which may have belonged to a single implement as the same type of parallel oblique striations are present on each of the inferior faces and the pieces are of similar shaft dimensions (GdA 03: max. length: 74 mm, max. width: 11.1 mm; GdA 09: max. length: 55.4 mm, max. width: 10.2 mm). While GdA 03 displays a bevel fracture at both extremities, the mesial-proximal GdA 09 fragment has been broken post-depositionally at its distal end. Some crushing and rounding is visible on its curved proximal edge, however, indicating that it had seen use before discard.

Other antler-based projectiles recovered from this site included three near complete unbarbed points (sagaies) (GdA 17–18), four distal-mesial fragments probably from these same weaponry types (GdA 01, 02, 08, 10), and four mesial fragments with textbook impact (splinter and bevel) fractures (GdA 05, 06, 12) (Fig. 5A). While three of the distal-mesial fragments exhibit impact fractures on their distal and proximal extremities (crushing, chipping and rounding on the distal tip and splinter fractures on the proximal extremities), the fourth (GdA 02) terminates in what is known as ‘sectionnée par rupture après raclage’ (Chauvière and Rigaud, 2005) or ‘sectionnement par raclage en diabolo et détachement par flexion’ (Pétillon, 2006), an intentionally worked fracture. On this particular piece, both the superior and inferior faces have been cut and scraped (bifacially worked) in order to make the shaft thin enough to be snapped by flexion. The intentional removal of a point's distal tip was likely the result of either the weapon having been embedded in a carcass (to remove the weapon from the target) or to remove a damaged tip as part of maintenance activities. As this artefact also exhibits an uneven section along its distal half, resulting from the resharpening of the tip, either of these scenarios is plausible.

The three most complete points (GdA 16–18) are bevel-based points. The first (GdA 16) is a near complete single bevel-based point with an intact base and small bevel fracture at its distal tip (resulting from impact with a target). This weapon tip measures 260 mm in total length. The second artefact is a post-depositionally fractured mesial-proximal fragment of a double bevel-based point (GdA 17), measuring 255.3 mm in total length. The third artefact (GdA 18) measures 217.5 mm in total length and is a post-depositionally fractured mesial fragment which may have originally belonged to either a single or double bevel-based point. Interestingly, both GdA 16 and 18 have a series of closely knit slightly parallel and almost vertical lines engraved down their superior face. The fact that these artefacts exhibit the same decoration and were recovered together from this site suggests that they belonged to the same local technical tradition and perhaps even the same toolkit.

In that same box, were four pieces of tool blanks, including one complete example (Fig. 5: GdA 07, 13–15). These pieces display manufacturing stigmata in the form of grooves and cut marks resulting from the use of the groove-and-splinter technique to extract them from the antler beam. The groove-and-splinter technique was the most commonly used blank extraction method utilized during the Middle to Late Magdalenian (Pétillon and Ducasse, 2012). The complete example (GdA 13) appears long enough to be a projectile point blank at 198.6 mm in total length and 19 mm in maximum width.

Also identified were two distal fragments of burnishers (lissoirs), identifiable from their worn, smooth and rounded distal extremity: typical use wear for these leather working tools (Fig. 5: GdA 11 and 04). Measuring 67.7 mm (GdA 11) and 91.4 mm (GdA 04) respectively, each terminates in a splinter fracture at their proximal extremity indicating that they broke during use.

Finally, a heavily patinated fragment of a shouldered lithic point, which seems closer to the Magdalenian than Solutrean type (Fig. 5) along with a single perforated and engraved red deer canine was found (Fig. 6). This pendant (18.6 × 13.2 mm) exhibits a single perforation which has been drilled from both sides along with oblique lines engraved into its side and edge. There is a slight reddish tinge to the lines found on the side which may be traces of ochre, however, analysis to confirm this suggestion is yet to be undertaken. This bead is consistent with the many similar pieces of red deer teeth ornamentation found in Magdalenian contexts throughout Western Europe (e.g. Vanhaeren and d’Errico, 2005).

Owing to the circumstances surrounding the recovery of these artefacts, we cannot be sure of their direct association with one another, and thus, our interpretative capabilities are rather limited. Despite this setback, however, the fact that a matching pair of HRR components was recovered lends a little weight towards this small assemblage having been deposited over a restricted period of time. Moreover, and as mentioned earlier, there appears to be only one archaeological horizon recognizable in the stratigraphic cut left by the old investigations.

The material culture discovered at the museum may hold some clues regarding its age. As mentioned earlier, HRRs are considered as a fossile directeur for the Middle Magdalenian, even though they may appear in Early Magdalenian and early Late Magdalenian assemblages. The absence of barbed points (harpons) (present at the nearby Grotte de l’Ammonite) or other characteristic elements precludes a chrono-cultural attribution to the Late Magdalenian phase. By contrast, the strong similarity of features in this collection with other supposedly contemporaneous sites (e.g. Le Placard: Chauvet, 1910; Grotte de la Mairie at Teyjat: Capitan et al., 1908) supports an attribution to the later part of the Middle Magdalenian (Breuil's ‘Magdalenian IV’).

At Grotte de l’Abbé, the presence of the shouldered point is a bit ambiguous. With a very patinated appearance, such a tool, characteristic of either the earliest sub-phase of the Middle Magdalenian or the Late Magdalenian (see Demars and Laurent, 1989 and Lenoir, 2003) may be intrusive and reflect a more disturbed archaeo-stratigraphic context than initially expected. Nevertheless, in order to ultimately clarify the chronological position of these artefacts, some of them will need to be directly dated.

Overall, Grotte de l’Abbé and the finds presented here, are a welcome addition to an already impressive list of sites dating to the Magdalenian in this region. Indeed the area of Vilhonneur witnessed a dense human occupation during this period, with no less than nine sites already recorded: La Robinière, Grotte de l’Ammonite, Grotte du Bronze, Abri André Ragout, Grotte des Fadets n^o 1, Abri du Chasseur, Grotte du Loup, Grotte du Sureau, and Grotte du Placard (see Balout, 1959, Debénath, 2006, Debénath, 2014, Dujardin and Gomez de Soto, 2001 and Patte, 1941). Most of these sites are ill-known owing to the nature of early investigations which are usually associated with a lack of field notes and the loss of collections. Among these sites, Le Placard stands out as an exceptional site owing to its wealth and diversity of past human activities represented there along with the volume of artefacts exhumed by numerous generations of scholars. At this site, several chrono-cultural entities are documented, including the Upper Solutrean and Badegoulian, as well as the Lower and Middle Magdalenian. As noted above, Grotte de l’Abbé could also date to this latter phase. Thus, at the moment, these two sites may be perceived as contemporary. Additionally, it should be noted that about 15 miles away toward the south-east, Grotte de la Mairie in Teyjat (lowest layer, dated roughly to ca. 17,100–15,800 cal. BP) may constitute a third site in the region of this (still poorly known) cultural entity of the late Middle Magdalenian.

Clearly, additional work at Grotte de l’Abbé is justified with not only extraordinary artefacts having been already recovered, but also the potential of this site having functioned contemporaneously with nearby Le Placard and other sites.