The present study was carried out thanks to a grant from the Franco-Russian Study Center of Moscow and the support of the Artemir Associated International Laboratory.
We are thanking J.-J. Cleyet-Merle, director of the National Museum of Prehistory of Les Eyzies-de-Tayac, for allowing us to work in La Grèze cave, and G. Lévy, manager of the caves of Font -de-Gaume and Combarelles, for his logistical assistance.

Introduction

1S.A. Semenov is recognized as the inventor of traceology as an archaeological discipline (Семенов 1957; Semenov 1964), but its definition in Europe and the United States has been more restrictive than in Russia. Its field of application has long been limited to the use wear of lithic tools cutting edges, most often analyzed under the microscope at high magnification, whereas traceology was originally designed to take into account not only a great diversity of materials but also the traces of manufacturing produced by all kinds of techniques and therefore the use of different scales of observation (Longo and Skakun 2005). More generally, any archaeological object can potentially be subjected to a traceological reading since it is itself the material trace of a human action.

2Moreover, the traceological approach is implicit or latent in studies that do not claim it as such or call it technological, especially in the underground environment (traces of mining, e.g. Arles et al. 2013; graphical techniques, e.g. Tosello and Fritz 2005, D'Errico et al. 2016), the validity of the approach being fundamentally dependent on the relevance of the used reference sources.

3The study of prehistoric art offers a favorable field of application to traceology as conceived by S.A. Semenov, whether on portable support or on rock walls. In addition to the recognition of techniques, it lends itself to the study of the reconstruction of the chaînes opératoires, that is to say, to the reconstruction of the unfolding, over time, of the actions that marked the painted, scraped, engraved, pecked or carved support.

4The relative chronology of the lines constituting the figures, and that of the figures forming the compositions, are part of the register of questions that may be approached in the analysis of the representations. The main criteria considered are the chronology of the superimposition of the lines and their homogeneity, from which are deduced the order of the tracings and their belonging to the same ensemble, the technical synchrony contributing to the definition of graphic units. This approach is necessary to disentangle palimpsests (e.g. Aubry, Santos and Luis 2014), which are frequent in Paleolithic art, especially on portable supports. These, however, quickly buried because of their small dimensions, were not conducive to prolonged re-use, unlike the walls and rock surfaces potentially accessible during millennia, sometimes up to the present times. The same places could be occupied at different times of Prehistory and their graphic compositions completed or transformed by artists using rather similar codes, at our scale of perception, as in the Chauvet cave (Clottes 2010). They may also have been visited much later by individuals writing their marks for other reasons, be they graffiti marking their passage, easily distinguishable as, for example, Christian symbols superimposed on "pagan" figures (Hinout 1998) or, on the contrary, re-tracings the less obvious to discern, intended to revive the representations, as is known ethnographically (Miklashevich 2008; Миклашевич 2011), or to make them more visible during their discovery or exhibition to the public. The combinations are numerous and a stylistic analysis is not always sufficient to distinguish the original works from secondary works or later additions.

5In any case, the question of time remains at the center of preoccupations, whether it concerns the relationship between two lines, between different parts of a figure, between the figures of an ensemble or between the graphic compositions of the same space (from the wall to the site). In some contexts, isotopic dating can be carried out directly on the figures when they include organic pigments, or on the carbonate deposits that seal them (e.g. Genty et al. 2011, Benson et al. 2013). But most often the dating evidence are indirect and apply to the whole of a composition when it is in contact with a stratigraphic deposit itself dated in various ways (Aubry and Sampaio 2008). Also, it is on the scale of the tracings themselves that we must look for additional clues. This is a field of investigation brilliantly illustrated by the pioneering works of D'Errico, Sacchi and Vanhaeren (2002) and Gyria and Devlet (Гиря and Дэвлет 2010), from the signature of engraving and pecking tools, and more particularly from a distinction between lithic and metallic forms, which can sometimes be very useful in clarifying the relative chronology.

6Parallel to comparison reference sources, the efficiency of the traceological analysis is based on observation and recording techniques. These have progressed considerably in recent years, thanks to 3D imaging, which has become more accessible and usable in the field at scales compatible with reading the finest traces (Plisson and Zotkina 2015). This aspect is particularly important in the case of petroglyphs because the relief of the traces or shapes makes it necessary to take into account the 3^rd dimension. Until now, the use of imprints (silicone or other) was inevitable for measurements and profiles recording (D'Errico, Sacchi and Vanhaeren 2002; Гиря and Дэвлет 2010), with all the risks of pollution and tearing-off this represented for the surfaces. This method, henceforth prohibited, is nowadays advantageously replaced by photogrammetry.

7The traceological approach developed and implemented on engraved and pecked outdoor petroglyph sites (D'Errico, Sacchi and Vanhaeren op.cit.; Gyria & Devlet op.cit.; Zotkina 2012; Зоткина 2013; Черемисин et al. 2013; Zotkina et al. 2014; Avrillaud 2015; Plisson and Zotkina 2015), had not yet demonstrated its utility in underground environments where painted figures concentrate most of the analyzes. As a demonstration, we have applied it to the question of the re-tracing of figures, starting from the preliminary example of the bison of La Grèze cave.

Case study

8The cave-shelter of La Grèze is a small cavity in the Grande Beune valley, below the vicinal road connecting the RD 48, in the municipality of Marquay in the Dordogne (Patriarch 24 255 0003). It is known for its 60 cm by 40 cm bison figure (fig. 1), well preserved due to its position in a concave recess of the wall protecting it from erosion, unlike other animal representations of which only the lower parts remain.

9The cave and its engravings have been studied for more than a century, from 1904 (Capitan, Breuil and Ampoulange 1904a, 1904b) and several researchers succeeded one another (Ampoulange and Pintaud 1955; Aujoulat 1984; Breuil 1952; Cretin 2015; Delluc and Delluc 1991, 1992; Delluc and Bouvier 1984; Jaubert 2008; Peyrony 1949; Roussot 1965; Sonneville-Bordes 1960, 1965; Vidal 1967).

10Thanks to modern techniques we can now extend their work and answer some of their questions.

1 - Has the bison figure been over-engraved?

11The light color of the bison’s tracing attracts attention, and it was assumed that this fact could be explained by a re-tracing of the figure: "It is not impossible that the lines were recently over-engraved with a sharp tool, as the limestone is much lighter at the bottom of the lines than on their sides" (Delluc and Delluc 1991 - p.238); "... incidentally, the famous bison shows some traces of molding and perhaps its contour has been discreetly over-engraved" (Delluc and Delluc 1991- p. 245).

2 - Technique: pecking or deep engraving?

12In all the publications dedicated to La Grèze, the researchers mention the figuration of the bison as engraved (Peyrony 1949; Breuil 1952; Aujoulat 1984; Delluc and Delluc 1991). "The engraving is done by deep lines with angular section whose bottom is more or less blunt. The edges are often rounded, but this modeling of the inner edge may be related more to erosion than to human origin "(Delluc and Delluc 1991-p. 238).

13But Rigaud (1984 – p. 278) mentions another technique that may also have been used: "The other lines are several fine lines with angular sections, vigorous lines obtained by pecking, and some broad lines, probable remains of vigorously engraved lines."

14The technique of pecking can be considered as one of the explanations for the irregularity of the bison tracings.

3 - Does the alteration of the support modify the engravings?

15The first thing to consider when dealing with technical traces in rock art, and more generally in traceology, is the influence of the natural alteration of surfaces on the lines and on the features that are a priori seen as being typical, that is to say resembling tool marks (Kervazo et al. 2010). Without understanding the effects and mechanisms of the alteration, the risk of drawing false conclusions is not negligible.

16That is why this question must come first in the context of such research.

Techniques, methods and equipment

17This research focuses on traceological aspects. It changes the scale of observations and allows a deeper understanding of the mechanisms of transformation of the reliefs of the rock surface.

18Our observations were made with a binocular microscospe WILD M1A (x7, x14). The working distance of this microscope (about twenty centimeters) is high and allows avoiding accidental contact with the examined surface. Its lightness makes it possible to work without a tripod (fig. 2).

19For the recording and comparison of reliefs we used photogrammetry (Plets et al. 2012; Robert et al. 2012; Belarbi et al. 2012; Cassen et al. 2012; Plisson 2015). Photographs were taken using a Nikon 750D camera (35.9 x 24 mm sensor, 24.93 million pixel resolution), equipped with an AF-S Micro NIKKOR 60mm f / 2.8G lens and a ring flash. The models are generated and processed with the Agisoft Photoscan, MeshLab and GeoMagic software, from which cross-sections of the tracings are extracted.

20We have also made macros showing the characteristic details of the figure with a Canon 1000D case 22x15mm format with a Canon EF-S 60mm f / 2.8 Macro USM lens and extension rings attached to a rack rail and a robust tripod. The photographic sequences were conducted from a computer and then compiled with the Helicon Focus software to obtain high-resolution, hyper-focus views.

21During the observations and recordings, we have used Aurora Max 5600 LEDs lighting system whose luminous intensity and color temperature could be adjusted.

22The experimental references for comparison are those developed in the author's recent thesis (Zotkina 2013), established on different terrains, and enriched by archaeological and ethnographic examples related to different techniques (Plisson and Zotkina 2015), with lithic and metal tools. The observations carried out at La Grèze were supplemented by tests (fig. 7) on small blocks fallen outside at the foot of the shelter.

23But we must consider that the means used here are not always possible in the configuration of decorated caves (difficulties of lighting, difficulties in approaching the panels, fineness of certain engravings, extent of decorated surfaces, etc.).

Grooving, pecking or alteration of the surface?

24The lines of the bison and the silhouette identified as that of a mammoth, on the same wall, are very irregular and sometimes very deep. Some surface portions of the two figures are very alveolar. This makes it possible to speculate that these lines would have been done by the pecking technique. B. and G. Delluc indicate a fairly visible variation in the depth (between 2 and 7 mm) and the width (between 3 and 12 mm) of the lines of the bison (Delluc and Delluc 1991 – p. 238). The tracings are sometimes described as having a "recticurvilinear" section (op. cit., fig. 168). The technique of photogrammetry makes it possible to reproduce precisely their morphology (Plisson and Zotkina 2015). On our 3D models we can see sections that cannot be described as typical for engraving lines (D'Errico, Sacchi and Vanhaeren 2002), which usually rather have a regular profile more or less in V (fig. 3 a-g) .

25It is evident, however, that the support is fragile and quite altered, as already mentioned by Delluc (1991: 235).

26During our observations with the binocular microscope we observed the following characteristics:

27The conservation of tool marks in the representations of other caves in the region (e.g. Combarelles I and II, Font-de-Gaume) can be much better than in La Grèze; there are signatures much more visible and revealing. Indeed, on the edge of the broad and deep main groove of the bison, we have no finer traces that account for the specificities of the active part of the tool, nor a parasitic tracing revealing deviation in gesture. The high level of alteration does not allow to go beyond the survey of the profiles of the groove. This comparison leads us to believe that the state of preservation not only of the support itself but also of the prehistoric tracings on the walls of La Grèze is not conducive to a thorough traceological research.

28The comparison of some tracings of the bison (fig. 3, a, f) with the naturally altered parts that do not have an engraved line morphology but that resembles them at the level of the contour (fig. 3, d, e) shows similarities suggesting that areas considered to be pecked would in fact be altered (fig. 3).

29This is why we suppose that we are not observing tool marks but forms produced by alteration (fig. 3, 4).

30We are fortunate to have some more striking features, like those on the back of the bison (fig. 4 c-h): the line is rather narrow and the edges almost regular, which cannot be obtained by pecking. We also notice a line at the bottom of the tracing that is not really an imprint of the active part, because the surface is still quite altered, but that nevertheless shows the general relief of the line. This line evokes a gesture rather resembling grooving, and therefore deep engraving. We are in the part of the figuration where the surface is rather better preserved compared to the other zones. Above all, it seems to us that the gesture has not been very vigorous.

31We believe that another remarkable clue of the engraving technique at the level of the horns is important to describe: it is visible (fig. 5), when the lines change orientation in a rather sudden manner, certainly where the material has resisted the tool during the tracing. Indeed, when crossing harder parts of the material the tool has deviated. This is very typical of the technique of engraving, because the tool follows the surface while remaining almost constantly in contact, and is guided by the structure of this material, which is not at all the case with pecking.

32A further argument is the damaging of the representation of the mammoth (fig. 6) located about fifty centimeters to the left, on the same wall. We observe the consequences of the desquamation that almost destroyed the figure (fig. 6, a, b). There are nevertheless some preserved traces, for example an extremely narrow and deep line with regular edges, which is at the level of the leg (fig. 6c). This allows us to say that the alteration of the surface has widened the lines and modified their configuration.

33We therefore come to the conclusion that the representation of the bison was carried out by the technique of deep engraving, which may be described as grooving, but that the evolution of the wall has, in some places, given the tracing a pecked appearance .

Experimental verification

34Six geological facies were identified in the cave (Le Filiâtre in Cretin et al., 2015 -p.50, fig 1/23). Most of the figures, at least the most famous one, are engraved in the D facies, "pseudo-oolitic white limestone." It is a fairly homogeneous and tender material, easy to work by the engraving technique. For our experiments, we took small blocks from the same facies (fig. 7). We have tested a grooving method capable of producing a deep furrow like that of the silhouette of the bison, as it can be observed in the least altered portions of the tracing. We used a local flint flake, on the raw surface in the surrounding humidity of a spring undergrowth, but without the addition of water or other components.

35A deep groove was obtained by combining two gestures: first a vigorous gesture by repeating the movement 4 to 5 times in one direction, until a fairly marked groove makes it possible to guide the active end of the tool, then the deepening by a back and forth movement. When the tracing becomes deep enough, the tool then begins to smooth out the edges of a groove, which gradually becomes rounded (fig. 7, a, b). This corresponds well to the sections of the bison tracing at the level of the back (fig. 4c, f) where we recognize the characteristic traces of the grooving. The contour of its head is more drawn out, which involved tilting the tool to reduce the outer edge of the groove.

Lighter lines

36As mentioned above, researchers have raised the question of the renewal of the bison tracing (Delluc and Delluc 1991 – p. 238, 245). In our observations, we identified only two zones, reduced in size, where the bottom of the line is dark, that is to say the color of the wall: these two zones are located at the level of the tail and back. All the other parts of the engraving are light.

37To determine if the figure was over-engraved in contemporary times, two aspects are to be considered. The first relates to the natural alteration of the area concerned and its relation to the tracing, and the second to the marks that would be specific to a modern tool.

38As we have remarked, with the alteration, the surface becomes very irregular and the edges of the natural concavities are rough but these areas are quite light. If we compare them to the profiles of the areas of the bison figure that we consider to be very altered, and to those that show representative features of deep engraving technique, we see that the former are much closer to the shapes of degradation than the latter (fig. 3, fig. 4 a, b, fig. 6 a, b). In places it is seen that the dark and narrow groove is transformed very gradually into a wider groove, much less regular and lighter, which we consider to be the consequence of the alteration (fig.4c, e).

39Our observations with a binocular microscope showed no sign of a modern tool in the bison tracing, nor in the mammoth’s, which should be the case if the figures had been recently re-traced. We have examples of contemporary traces made during the excavation of the cave. On the one of figure 8, referenced as such by the managers of the cave, we can clearly see the marks of the tool that attacked the material: the bottom of the line was compacted and somewhat smoothed (fig. 8), and especially fine lines, which underline the movement, are clearly visible (fig. 8). These traces are also on the wall opposite to that of the bison but whose surface state is very similar. They can therefore be taken into consideration when comparing technical traces and taphonomic traces and, subsequently, be included in the argumentation: nothing supports the hypothesis of a modern re-tracing of the figures of the bison and the mammoth. The widening of the tracing and its irregularities are only due to the natural alteration of the wall.

40The argument that the light bottom of some portions of the tracing would show a recent re-tracing is ultimately not so decisive, as this is one of the consequences of the alteration. Several zones of natural desquamation are equally as light (fig. 3, fig. 4, a, b, fig 6, a, b). The phenomenon affects the degraded areas of the bison as well as the mammoth. It is active both on a macroscopic scale, with a very perceptible effect, and on a microscopic scale where a micro-desquamation, albeit less blatant, also inexorably alters the surface and thus the millenary patina that had darkened the tracings. However, the method applied here does not allow us to say at what period this alteration occurred or what its intensity was.

41We can therefore conclude that the present appearance of the figuration of the bison of La Grèze owes nothing to a recent intervention.

Conclusion

42If the silhouette of the freshly engraved La Grèze bison appeared at the time, in all probability, lighter than the wall, the perception we have of it today is not due to the freshness of the line but to an alteration of uneven intensity, having in some places over-dug the tracing, whitening it again giving it the appearance of re-tracing with a different technique from that used by the Paleolithic artist. A fine analysis of this tracing, using a stereoscopic magnifying glass, and two- and three-dimensional detail recordings, reveal only two categories of features: some relating to deep, old engraving and others to the natural degradation of the wall. No feature can be related to pecking or to any more recent intervention on the figure itself, whereas the wall shows, elsewhere, marks of tools probably dating from the initial excavations in the cavity.

43The traceological approach is an integral part of the study of painted prehistoric representations. The problematic of modern re-tracings, or largely subsequent to their initial making, deals here with engraved, carved and pecked forms; it opens up an additional field of expertise to traceology, which we have illustrated with the representation of the bison of La Grèze cave.