TRANSIT, An Experimental Archaeological Program in Periglacial Environment: Problem, Methodology, First Results

GEA(Wiley)

RIGHT

INTERACTIVE

TRANSIT, An Experimental Archaeological Program in Periglacial Environment: Problem, Methodology, First Results J.P. Texier,1 P. Bertran,1 J.P. Coutard,2 B. Francou,3 P. Gabert,2 J.L. Guadelli,1 J.C. Ozouf,2 H. Plisson,4 J.P. Raynal,1 and D. Vivent1 1

Universite´ de Bordeaux 1, Institut du Quaternaire, UMR 9933 CNRS, Avenue des Faculte´s, 33405-Talence, France 2 Centre de Ge´omorphologie, URA 1694 CNRS, 24, Rue des Tilleuls, 14000-Caen, France 3 ORSTOM-Bolivie, CP 9214, La Paz, Bolivia 4 Centre de Recherches Arche´ologiques du CNRS, Sophia Antipolis, 06560Valbonne Cedex, France

The aim of the experimental archaeological program TRANSIT is to improve the scientific study of Paleolithic sites. This program is based on studies and experiments carried out at high altitudes in the French Alps. One of its goals is to assess the effects of periglacial processes on spatial distributions within archaeological assemblages and on artifacts and bones. Research on sedimentary environments elucidates stratogenesis in periglacial contexts. In particular, sedimentary models that are useful for interpreting fossil deposits were identified. The first results obtained from experimental artifact test plots emphasize the importance and speed of changes that occur in such a climatic and morphodynamic context. Mean lateral displacements from 1.66 to 4.75 cm yr⫺1 were measured. Moreover, analysis of osseous and dental pieces showed that they experienced important damage. Works on pollen assemblages have shown that most of the pollen deposited on the ground surface is quickly removed and that, owing to the very uneven preservation of pollen taxa, an important distortion of the initial spectrum occurs. A model of spatial arrangement of artifacts linked to the action of solifluction processes is also proposed. Finally, application of these results to some Paleolithic sites in the Massif Central and southwest France are discussed. 䉷 1998 John Wiley & Sons, Inc.

INTRODUCTION At one time or another in their history, most European Paleolithic sites experienced periglacial climates. Studies carried out in this kind of environment (Washburn, 1979) show that numerous dynamic processes produce more or less important modifications in stratification of archaeological materials (Hopkins and Giddings, 1953; Mackay et al., 1961), spatial distribution of artifacts (Benedict and Olson, 1978, quoted by Wood and Johnson, 1978), and physical aspects of archaeological pieces (e.g., owing to frost shattering). Geoarchaeology: An International Journal, Vol. 13, No. 5, 433– 473 (1998) 䉷 1998 John Wiley & Sons, Inc. CCC 0883-6353/98/050433-41

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Although Wood and Johnson (1978) emphasized the importance of problems in interpretation due to these modifications, archaeologists working in Europe have, until now, largely underestimated and sometimes completely ignored the effects of these processes. Serious misinterpretations result from this situation: (1) creation of technological “synthetotypes” as a result of the mixing of several levels by sedimentary processes, or, on the other hand, identification of pseudoarchaeological levels in fact coming from a single human settlement; (2) natural retouch of lithic pieces and natural bone fragmentation wrongly attributed to human activity; (3) sedimentary structures interpretated as being of human origin; (4) erroneous taphonomic interpretation of bone assemblages; etc. To our knowledge, the only experimental work which aimed at the precise assessment of the consequences of periglacial processes on archaeological assemblages was that done by Bowers et al. (1983) in Central Alaska. These researchers were essentially concerned with lateral movements of lithic pieces and, as such, they identified their rate of movement and its importance. Mean displacements of 4 cm/yr were measured, with movements reaching 20 cm in a year in some instances. Nevertheless, relations between these modifications and specific geomorphologic mechanisms were not clearly established. Moreover, the work of Bowers et al. (1983) does not tackle the problem of physical alteration of osseous and lithic pieces. This near absence of analogs concerning such basic aspects of prehistoric site interpretation led us to devise a program to deal specifically with this problem. This program is referred to as TRANSIT (Transfert de Re´fe´rentiels Actuels de l’e´tage Nival aux SITes pale´olithiques). THE TRANSIT PROGRAM: OBJECTIVES AND MODUS OPERANDI Objectives The aim of the TRANSIT program is to improve upon current methods interpretating the formation of Paleolithic sites. One of its main objectives is to study the modifications of experimental archaeological materials placed in a periglacial context that include alterations (types and rates) of osseous and lithic pieces, analysis of artifacts displacements (strike, magnitude, and velocity), and assessment of the consequences of these movements on patterns of archaeological assemblages. Another goal is the examination of natural environments and, more specifically, periglacial processes, in order to better estimate their respective roles and to highlight diagnostic criteria for their identification in fossil deposits. The program also intends to study the nature of pollen assemblages in terms of the trapping, redistribution, and differential alteration of pollen grains. These assemblages will also be evaluated to assess their representativeness in comparison with regional vegetation. Finally, the results of this program will be used for better prehistoric site interpretation. In our first stage, we have focused upon prehistoric sites located in southwest and central France. Obviously, TRANSIT cannot address every kind of modification linked to peri-

434

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 1. Location of the studied area (star).

glacial contexts, but the program should, above all, be considered as a first approach to the important problem of site formation in periglacial contexts. Modus Operandi Choice and Main Characteristics of Experimental Site An experimental site was chosen based on its fulfillment of several requirements: (1) a periglacial environment; (2) a midlatitude location (as are the studied Paleolithic sites); and (3) good information on local climatic parameters in order to assess their consequences on morphodynamic processes and on archaeological assemblages. The massif of La Mortice (el. 3169), located in the southern part of the French Alps (Figure 1), satisfies all these requirements and is situated close to the ⫺ 3⬚C mean annual isotherm (see Coutard, 1985; Coutard and Francou, 1989; Mante´, 1985, 1988, 1989; Coutard and Ozouf, 1993; Coutard et al., 1996 for a detailed study of air and ground temperatures). On the east of the mountain, where the main experiment was located (cf. infra), the number of surface freeze-thaw cycles is estimated at 105 – 115 per year (measurements performed between August 1986 and August 1987) (Coutard et al., 1996). At 5 and 10 cm depth, the number of such cycles is only 12 and 5, respectively. On this mountainside the depth reached by frost was calculated to be 175 cm for the winter 1986 – 1987 using the formula proposed by Rouque`s and Caniard (1975): H ⫽ A · I1/2, where H is the depth (in cm) reached by frost, I is the freezing index (in ⬚C days yr⫺1 ) at the soil surface, and A is a coefficient depending on the sediment type (for uncemented rocks similar to those present on the site, A is about 5). Thus, the occurrence of a sporadic per-

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

435

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

mafrost is likely. These results also correspond with the measurements performed by King (1990) in the Swiss Alps where sporadic permafrost develops between ⫺ 3.5⬚C and ⫺ 1.5⬚C atmospheric isotherms. Experiments Carried Out on Archaeological Materials Two types of experiments were carried out. One specifically examines the frost heaving process (experiment no. 1), with a goal of assessing the consequences of this phenomenon on the vertical distribution of archaeological pieces. The other experiment is much more complex and deals with the modifications of lithic and bone objects placed on the ground surface in a solifluction-dominated periglacial environment (experiment no. 2). Experiment No. 1. Experiment no. 1 was set up on the northwest side of La Mortice (Figure 7). It simulates the modifications experienced by superimposed archaeological levels which would normally be separated from each other by sedimentary layers some centimeters thick. Three sets of 50 marked flints were buried in a line at different depths (10, 5, and 0.5 cm respectively), just downslope from a stone-banked solifluction lobe in a matrix-rich sediment of a slopeward stretched mud-boil. The flints were slightly elongated flat flakes with a mean length of 4 cm. The dip of the slope averages 25⬚ but locally varied considerably. On the experimental site itself, the inclination of the upslope solifluction front is 37⬚, whereas it is only 8⬚ on the top of the mudboil; it reaches 20⬚ 2 m downslope. The experiment was begun on August 22, 1990, and several measurements were made systematically every year: (a) number of frost-jacked artifacts; (b) magnitude and strike of lateral displacements of artifacts located at the ground surface; (c) number of flints turned over. Experiment No. 2. For this experiment, five homologous “archaeological” assemblages comprised of lithic pieces and bones were put together. Lithic pieces were manufactured from flint nodules of the same quality by a skilled knapper (M. Lenoir, Institut du Quaternaire, Universite´ de Bordeaux I). Each assemblage was composed of 100 pieces and small retouch-flakes ⬍ 5 mm wide; the 100 artifacts consisted of flakes, blades, and cores. Bones were selected in order to represent as many potentially real cases as possible, with bone assemblages containing bones from three main categories: (a) “fresh” bones, including long bones (proximal, mesial, and distal fragments), phalange of bovidae, fox rib, mandible and skull of sheep, fragments of maxilla and of palatum osseum of sheep; (b) fossil bones of reindeer and horse (diaphysis fragments; and (c) “fresh” teeth (sheep) and fossil teeth (Saı¨ga tatarica). Bones and teeth were not subjected to particular chemical processing. Fresh skulls and teeth were air-dried, and long bones of Bovidae were placed in boiling water for a few minutes in order to extract part of their fat and for cleaning. All these lithic and faunal objects were numbered except for the small retouch flakes.

436

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 2. Experiment 2: general view of artifact test plots.

Installation of the Experimental Test Plots Five experimental test plots were placed on the southeast facing slope of the Massif of La Mortice, on a bench near the 3090 m contour line (Figures 2 and 7). Here, the slope is between 8⬚ and 14⬚ and is analogous to most archaeological levels found in Paleolithic sites of southwest of France. The selected locality is situated in an area exhibiting uniform morphoclimatic characteristics and sedimentary processes, the latter dominated by solifluction (stone-banked lobe type 10 – 20 cm high). Nevertheless, other processes interfering with the solifluction do occur (see the next section). The test plots, corresponding to 1.5 m squares delimited by painted lines and numbered from 1 to 5, were placed on the distal part of solifluction lobes (near and on the stone front, Figure 2). The artifacts and bone pieces described above were placed in these squares. Distribution and orientation of the pieces were random in four test plots (1 – 4) (Figure 3). In contrast, test plot 5 reproduced some Paleolithic structures, including a hearth made of an accumulation of charcoal delimited by rounded quartz pebbles and concentrations of artifacts (Figure 4). Retouch flakes were arranged in small piles inside the test plots. A colored dot was painted on the dorsal face of flints in order to identify possible turning movement.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

437

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 3. View of artifact test plot 1 (experiment 2).

Measurements, Experiments, and Analysis Carried Out on “Archaeological” Pieces Before installation of the test plots, some lithic artifacts were used to scrape wood, bone, hide, and limestone for a duration long enough to develop wear characteristic of these activities. After a careful cleaning with ethyl alcohol for a few seconds and sodium phosphate for 12 h, two areas of the active edges were photographed under a reflected light microscope in order to appraise changes in microwear after exposure to periglacial conditions. Three examples of each of the four use-wear types were placed in each different test plot (i.e., 12 in each test plot). A separate series of 12 pieces was kept as reference. During installation of the test plots, the positions of all the pieces were registered by means of an electronic theodolite (accuracy of ⫾ 1 mm) in relation to an arbitrary horizontal grid system and a geodesic point situated at the top of the mountain. For pieces having no significant elongation index, only one measurement was taken at the location of the colored dot. For elongated pieces, two measurements were performed, one at each end. The limits of retouch flakes piles were also carefully located. In this way, plans of test plots were achieved and were complemented by photographs and video documents. At the end of each experimental period (i.e., every year, see below), the same operations were performed, as well as the following: (1) excavation of one test plot

438

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 4. View of artifact test plot 5 (experiment 2).

employing techniques similar to those used in Paleolithic sites (pieces collected in this way were packed up separately in plastic bags in order to avoid shocks and damage); (2) descriptive study of movements and physical alteration of the pieces; and (3) fabric measurements of elongated pieces contained in test plots 1 and 5. A use-wear study was carried out in the laboratory of some of the artifacts collected in the excavated test plots. These included artifacts used to work wood, bone, hide, and limestone, as well as some artifacts with raw edges, in order to examine possible damage of tool edges. Moreover, changes in the structure of bones and teeth were assessed by measurements of their porosity with a mercury porosimeter and by the study of thin sections with an optical microscope. The five experimental test plots were constructed in September 1991, and fieldwork was performed each subsequent year in August. In the first season (1992), field work consisted solely of the mapping of the pieces contained in each test plot. Excavation of test plots began in 1993 and has continued since then at a rate of one test plot every year. Excavation of the last test plot containing random distributed pieces was undertaken in August 1996. Test plot 5 has been left on the experimental site in order to follow its evolution for as long as possible. Palynological Studies Palynological research was not planned at the outset of the program, and as a result the palynologic experiments began only in August 1993 and ended in August

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

439

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

1996. The main goal was to study processes of trapping and preservation of pollen grains in a periglacial environment. Assemblages of dyed pollens were placed at the surface of a polygonal ground (nine assemblages) and of a stone-banked solifluction lobe (nine assemblages), and their locations were marked by painted lines. Pollens used for this experimentation belong to the families of Asteroidaceae, Cistaceae, Cichorioideae, and Cerealia, and for each a corresponding one used as reference and kept in the laboratory. Every year, six core samples (three in the polygonal ground and three in the solifluction lobe) were taken in order to study movements and alteration of experimental pollens. Moreover, several steps were taken to assess how well natural pollen assemblages found in sediments reflect local vegetation cover: (1) characterization and distribution of the flora on the Massif of La Mortice; (2) characterization of pollen assemblages trapped in pond muds and in bryophytes of the alpine zone, as well as in snowpatches of the periglacial zone; and (3) characterization of natural pollen assemblages contained in core samples taken in polygonal ground and in solifluction lobe (cf. supra). Environmental Studies Climatic parameters of the environment were defined before the launching of the program (see above). Therefore, the environmental studies have mainly involved qualitative and quantitative characterization of sedimentary processes acting not only in the experimental sites themselves but also in the different geomorphic zones of the Massif of La Mortice. We have also tried to identify diagenetic processes in order to optimize transfer of results to fossil deposits. Methods used for this purpose are essentially descriptive and experimental. Descriptive Studies Three different descriptive methods were employed. First, the morphodynamic units present in the periglacial zone the area were mapped at a scale of 1/5000; this was done from field observations and from an aerial photograph produced by the Institut Ge´ographique National. Second, morphostratigraphic studies entailed detailed analysis of superficial landforms (sizes, structures, dips, orientation, associations with other landforms, granulometry, morphometry of stones, etc.), and associated sediments at depth. The latter required excavation of trenches 1 m deep and 2 – 3 m long, orientated both parallel and transversally to the slope. The resulting exposures were then described and then sampled for granulometric and micromorphological analysis, and for fabric measurements of elongated stones (a/b ⬎ 2). The microstructure of deposits was studied from large (13.5 cm ⫻ 5.5 cm) thin sections prepared from orientated blocks of sediments that were impregnated with polyester resin under vacuum (Guillore´, 1980). The terminology used for the description of thin sections is adaptated from that of Bullock et al. (1985).

440

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Experimental Studies Experimental studies were essentially concerned with solifluction, which is the main process acting on experimental archaeological sites. In order to appraise surface and subsurface processes (lateral and vertical translocation of fine particles, displacement of stones), a red strip was painted on the front of a stone-banked lobe and a cylinder of white loamy sand (compositionally very different from local sediments) was buried in the upper part of the same lobe (Figure 5). Observations were made every year on movements of painted stones and on the transport of fine sediments on the surface of the lobe. At the end of the experiment, sectioning and sampling will be done in order to assess translocations experienced by exogenic sediments. The strong preferred orientation of stones in solifluction deposits has been stressed by all authors working in present-day periglacial areas (Brochu, 1978; Nelson, 1985; Matthews et al., 1986; Harris, 1987; Van Steijn et al., 1995). In order to better understand development and evolution of this fabric, the following experiment was carried out. Elongated and rounded white quartz pebbles were buried about 5 cm deep in the matrix-rich level of a stone-banked lobe. Elongated pebbles were positioned either in the stratification plane, their main axis making an angle of 45⬚ with the line of maximum dip, or in conformity with the line of maximum dip but with an angle of 40 – 45⬚ with the stratification plane (Figure 6). Moreover, the upward side of all the pebbles was marked by a painted dot. At the end of the program, strike and dip of the pebbles will be measured in order to appraise changes in fabric according to shape and initial position towards the stratification plane and the line of maximum dip. RESULTS AND DISCUSSION Research on the environment is less time-dependent, so it is currently more advanced than studies dealing with archaeological or pollen assemblages. Environment and Morphodynamic Processes Our activities have mainly concerned the spatial distribution of morphodynamic units, as well as the processes that can produce stratified deposits (stone-banked lobes and debris flows). Indeed, these lithofacies are widespread in Paleolithic sites of southwest France and the French Massif Central. Nevertheless, within the context of this article, we will focus especially on processes acting on the southeast and northwest-facing mountainsides of the Massif where the archaeological experiments were placed, that is, solifluction linked to runoff and wind. Spatial Distribution of Different Morphodynamic Units (P. Bertran and J.P. Texier) A large scale map (1/5000) of the main landforms existing on the massif of La Mortice was produced, revealing an altitudinal pattern for the area (Figure 7). Be-

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

441

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 5. Stone-banked lobe in which a cylinder of exogenic white fine sediment has been buried (dashed line) and the stone front of which has been painted in order to highlight the kind and the importance of displacements suffered by particles in this sedimentary environment (knife for scale).

short standard

442

VOL. 13, NO. 5

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 6. Experiment placed (before burying) in the upper part of a stone-banked lobe (matrix-rich layer) in order to assess the acquisition rate of an oriented fabric in a solifluction environment.

short standard

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

443

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 7. Map of the main morphodynamic units in the studied area and location of experimental sites (stars): (a) stone banked lobes; (b) alpine lawn; (c) temporary lake; (d) rockfall talus; (e) sorted polygons; (f) sorted stripes; (g) macroscale fronts; (h) experimental sites 1 and 2.

444

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 8. Stone-banked lobes on the east-facing mountainside of La Mortice (French Alps, 3100 m) (shovel for scale).

tween 3150 and 2600 m, the periglacial zone stricto sensu is developed, and stonebanked lobes (Figure 8) and mudboils dominate. On the west facing side, fronts are higher (0.5 – 1 m on average) than on the east facing side (0.2 – 0.3 m). These landforms are sometimes superimposed on larger lobes more than 1 m high which could be the result of the creep of permafrost zones. Locally, sorted stripes can be seen, along with sorted polygons on small benches (Figure 9). Furthermore, rockfall talus and cones develop downslope scarps and cornices. The latter sometimes show in their distal part a well-defined ridge which can also be interpreted as evidence for the creep of a permafrost perhaps still active at the present time. The area between 2600 and 2450 m is the limit between alpine and nival zones. This belt is mainly characterized by debris-flow cones (Figure 10), partly vegetated solifluction terracettes, and torrential fans. Ploughing blocks are also frequent. A comparative study in Vallon Laugier from air photographs shows that debris-flow activity was low between 1971 and 1990 (Bertran and Texier, 1994). Indeed, during this period, the 10 recorded events correspond to small debris flows of only some tens of cubic meters, and most of them did not reach the distal part of the slope. Relict landforms, whose the ages are indeterminate, are also present in this transition zone in the form of rock glaciers, lobed screes, and tills.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

445

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 9. Sorted polygons (about 1.5 m across) on a bench located on the east-facing mountainside of La Mortice.

Stone-Banked Lobe Type Solifluction (P. Bertran, J.P. Coutard, B. Francou, P. Gabert, J.C. Ozouf, and J.P. Texier) Our research concerning this process follows that performed in the same area by Coutard et al. (1988) and by Van Vliet-Lanoe¨ (1988a). A sedimentary model was established from our results (Bertran et al., 1993, 1995b; Van Steijn et al., 1995), as outlined below. 1. Displacement velocity of solifluction lobes varies from 2 to 5 cm yr⫺1 at ground surface and progressively decreases below the surface to cancel out at about 30 cm deep. Each dynamic unit from the top to the down comprises (a) a superficial pavement and a reverse graded stony front, (b) a diamicton type matrix-rich layer, and (c) an openwork normal graded stone layer (Figure 11). 2. Advance of the matrix-rich sheet is caused by cryoreptation and gelifluction processes. 3. Advance of the superficial pavement is mainly due to the combined work of needle ice and passive transport over the moving diamicton layer. Reverse grading is the result of both frostjacking and the sieve effect (Francou, 1990; Bertran et al., 1994a, 1994b, 1995a, 1995b).

446

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 10. Debris-flow tracks on a cone of the east-facing valley side of Vallon Laugier (La Mortice, 2600 m).

4. Each solifluction lobe moves like a caterpillar: the matrix-rich layer progressively overruns the stone front which is simultaneously restored by frost heaving and by washing of fine particles (Figure 11). 5. Displacement velocity is not steady in the course of a year, which produces in the lower part of the diamicton levels an undulating boundary that is later stretched by shearing (Figure 11). 6. As the solifluction lobes advance they are progressively eluviated, ensuing a downslope evolution of the microstructures (Figures 12 – 14). 7. Runoff caused by summer rainfalls results in transportation of fine particles towards the stone bank. This process partially compensates for the loss by eluviation and takes part in maintaining the movement of solifluction lobes. 8. Eluviation of subsurface layers is accompanied at depth by accumulations of fine particles which preferentially occur on the top of buried matrix-rich layers (Figure 11). These translocations give rise to specific microstructures: laminated silt infillings between 30 and 50 cm depth (Figure 15), then silty clay coatings at about 100 cm deep. 9. A frost susceptibility gradient arises from elu-illuviation processes. This favors cryoturbation (Van Vliet Lanoe¨, 1988a, and 1988b) and formation of mud

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

447

short standard

Figure 11. (A) Longitudinal section showing the main textural and structural characteristics of stone-banked lobes; (B) same section with the main processes involved in this kind of sedimentary environment. (a and b) coarse openwork layer; (c) matrix-rich layer; (d) silty pan.

GEA(Wiley)

448

LEFT INTERACTIVE

TEXIER ET AL.

short standard

VOL. 13, NO. 5

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 12. Photomicrograph (PPL) of a matrix-rich layer near the starting point of a solifluction lobe. Note the vesicular porosity.

10.

11.

12.

13.

boils, which are the starting points for the development of new solifluction lobes. A vesicular structure develops in subsurface layers (Figure 16), and is derived from the collapse of the soil aggregates during periods of oversaturation (melting of needle ice and superficial ice lenses, melting of snow, heavy summer rainfalls). Deeper in the profile a platy structure (Figure 17) appears due to the formation of ice lenses parallel to the freezing front. The gradual increase in size which affects this platy structure can be related to the progressive slower rate of freezing with depth. Fabric measurements show that the main axis of stones has a strong slopeward preferred orientation (Figure 18; Table I) and a dip close to the slope angle. However, a more isotropic arrangement exists in the front lobes where stones are imbricated upslope. The importance of translocation of fine particles explains the formation of numerous mud boils and of small mudflows at the lower part of the slope as well as the development of an upslope surficial pavement. These different processes finally lead to the genesis of irregularly stratified deposits showing the aforementioned specific characteristics.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

449

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 13. Photomicrograph (PPL) of a matrix-rich layer midway between the starting point and the front of a solifluction lobe. Matrix is less abundant than upslope (compare with Figure 12). Coarse grains are bridged and capped by fine material.

The Role of the Secondary Processes in Morphogenesis of the East-Facing Side of the Massif of La Mortice: Runoff and Wind (P. Bertran and J.P. Texier) Runoff and eolian processes seem to play a rather significant role in the morphogenesis of the east-facing side of La Mortice. As mentioned in the previous section, evidence of this process has also been noticed in the archaeological test plots (cf. infra), in the distal part of the painted solifluction lobe, and, more generally, in the downslope part of the mountainside where runoff results in the development of numerous flat bottom rills and the formation of small colluvial fans 1 m across (Figure 19). These features show that a locally concentrated wash occurs and that the volume of sediments entrained in this way is rather important. It is very likely that this process plays a noticeable role in the translocation of fine particles down the slope, and it seems to be especially active during summer rainfalls. However, water from the melting of snow and ground ice can also contribute to the saturation of sediments and thus promote overland flow. In this sedimentary environment, the fabric of stones shows no preferred orientation (Figure 18) (Table I).

450

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 14. Photomicrograph (PPL) of a matrix-rich layer near the front of a solifluction lobe. Eluviation processes are more important than in Figure 13. Washed sands infill the voids between clasts. Cappings are always present on schist fragments.

Wind, which also plays a role in the morphogenesis of this area, has been appraised by studying particles trapped in two residual snowpatches. One of them is located just behind the crest of the massif, on the leeward side (Figure 20); the other is situated downslope on the same side, near the experimental site no. 2. From granulometric analysis (Figure 21), aeolian particles contained in the upslope neve mainly consist of coarse to very coarse clasts (stones 9.1%, sands 75.1%); silts only account for 15.8% of the total amount of the sediment. Nevertheless, the particles trapped in the downslope snowpatch are mainly composed of fine silts (mode at 10 ␮m) (Figure 21). Hence, there is a clear grading of the aeolian particles according to the distance from the crest. The amount of sediment transported by wind in the upslope part of the mountain side of La Mortice has been assessed at 65.64 g/m2 for 1995. Moreover, diameters (⫽ b axis) of stones transported in this way may be very large, up to 6.2 cm. These stones are usually very platy and their elongation index (a/b) is very low (Figure 22).

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

451

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 15. Photomicrograph (PPL) showing a laminated silt coating in a void of a buried solifluction lobe (80 cm depth).

Debris Flows (P. Bertran and J.P. Texier) Although debris flow processes have not been implicated in these test plots, it is sufficient to recall from previous research (Bertran and Texier, 1994; Van Steijn et al., 1995) that debris-flow processes can give rise to stratified deposits with lithofacies similar to that which occurs by the stacking of stone-banked lobes. However, characteristics that permit discrimination of these two kind of deposits have been discussed previously (Bertran and Texier, 1994; Van Steijn et al., 1995).

452

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 16. Photomicrograph (PPL) showing vesicles occuring near the surface of a solifluction lobe.

Changes Undergone by Experimental Pollen Assemblages (D. Vivent) The first palynological results involve only samples collected in the solifluction lobe, 1 and 2 years after the experiment began. After 1 year, pollen remaining in test plots accounts only for a maximum of 7 – 8% of the initial amount. Moreover, the quantity decreases with depth. With the exception of Cistaceae, the relative frequency of pollen changed according to the taxon: Cichorioideae are slightly overrepresented, while Apiaceae and Cerealia are underrepresented. After 2 years of experimentation, the amount of pollen preserved in test plots decreased again: A maximum of 3 – 4% were preserved and pollen of some taxa (e.g., Cerealia) is now completely missing. On the ground surface as well as at depth, Cichorioideae are strongly overrepresented, Cistaceae and Apiaceae slightly overrepresented, and Asteroideae underrepresented. These initial results show that in such an environment, most of the pollen deposited on the ground surface is quickly removed; absence of pollen because of destruction is unlikely because preservation is excellent. Moreover, owing to the very uneven preservation of pollen according to taxa, an important distortion of the pollen spectrum occurs. Reworking seems to be completely independent of size of the pollen. As Bottema (1975) has indicated, the higher preservation poten-

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

453

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 17. Photomicrograph (PPL) showing platy structure developed at 70 cm depth in solifluction deposits.

tial of Cichorioideae could explain the specific abundance of this taxon often noticed in fossil pollen spectra. Changes Undergone by Experimental Archaeological Material Although the program has not ended, important data concerning behavior and preservation of archaeological pieces in periglacial contexts have already been obtained. Displacements of Archaeological Pieces and Their Causes (J.L. Guadelli, J.P. Raynal, P. Bertran, and J.P. Texier) In experiment no. 1, only 3.6% of artifacts (2 of 56) buried at 10 cm deep reappeared at the ground surface. These were ejected only 1 year after the beginning of the experiment. It was also noted that 22.7% of artifacts (10 of 44) buried at 5 cm deep progressively reached the ground surface. There were three after 1 year, five after 3 years, seven after 4 years, and ten after 5 years. All of the 53 pieces placed near the surface reappeared as soon as the first year. One of them disappeared after 4 years; it has probably been buried. The number of pieces turned over by needle ice action increased regularly from 5 in the first year to 13 in the fifth year. All of the lateral displacements are downslope. Maximal

454

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 18. Schmidt stereonets (lower hemisphere) showing fabric of pebbles in a rill (1) and in solifluction deposits; (2: buried stone pavement; 3: matrix-rich layer; 4: surface pavement) and fabric of artifacts in test plots 1 and 5; (5: test plot 1, 1994; 6: test plot 1, 1995; 7: test plot 5, 1994; 8: test plot 5, 1995). The arc of circle corresponds to the stratification plane.

movement occurred in a temporary rill and reached 82 cm (i.e., more than 16 cm yr⫺1). The mean of displacement is 6.2 cm yr⫺1; this high value can be explained by the fine texture of sediments (clayey silts), which promotes gelifluction processes (Harris et al., 1993). Frost jacking is thus very active in fine-grained deposits. This process can rapidly lead to a more or less complete mixing of archaeological levels which were originally stratified.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

455

short standard

456 p 0.087 3E ⫺ 04 0 0.002 0.693 0.213 0.46 0.549 0.716

L (%) 29 51.57 63.26 45.42 9.02 28.52 15.84 14.14 9.91

E1

0.597 0.557 0.557 0.53

0.547

0.634 0.733 0.758 0.654

E2

0.394 0.421 0.435 0.453

0.43

0.327 0.225 0.168 0.255

E3

0.009 0.022 0.009 0.018

0.024

0.039 0.042 0.074 0.111

r1

0.42 0.28 0.25 0.16

0.24

0.66 1.18 1.51 1.02

r2

3.80 2.94 3.63 3.24

2.91

2.13 1.67 0.82 0.75

K

0.11 0.10 0.06 0.05

0.08

0.31 0.71 1.85 1.36

0.463 0.542 0.460 0.500

0.413

0.60 0.55 0.40 0.62

SVAR

Test plot 1 (1994) Test plot 1 (1995) Test plot 5 (1994) Test plot 5 (1995)

Rill

Solifl.—matrix-rich layer Solifl.—matrix-rich layer Solifl.—surface pavement Solifl.—openwork layer

N: number of data; L: vector magnitude (Curray, 1956); p: probability test of Rayleigh (Curray, 1956); E1, E2, E3: normalized eigenvalues (Watson, 1965, 1966; Mark, 1973); r1 ⫽ In (E1/E2); r2 ⫽ In (E2/E3) (Woodcock, 1977); K ⫽ r1/r2 (Woodcock, 1977); SVAR: spherical variance (McEachran, 1990)

30 31 30 34

45

30 31 30 30

N

Table I. Fabric measurements carryed out on pebbles contained in solifluction deposits, lying on the bottom of a rill and on artifacts of test plots 1 and 5.

GEA(Wiley) LEFT INTERACTIVE

TEXIER ET AL.

short standard

VOL. 13, NO. 5

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 19. Small (about 1 m across) runoff fan (underlined by a dashed line) downwards from a lobe front.

Measurements performed since 1991 in experimental site no. 2 indicate that most of pieces have experienced a downslope translation at a mean rate of 2.5 cm yr⫺1 (Table II) (Figures 23 and 24). This movement was expected and can be linked to solifluction processes. Displacements of 0.5 m to several meters (Figures 23 and 24) were also noted, involving some bones and charcoal. They are directed downslope, upslope, or even laterally, and can be attributed to the work of the wind, or to rainfall and hail. Weaker upslope displacements (3 – 4 cm) of some flints located downward on the front lobes are probably the result of cryoturbation (mud boil formation). These movements do not always correspond to a simple lateral transfer. In this regard, we noticed the knocking over of archaeological pieces located on the stony front in connection with advance of the lobe and the sliding of small pieces into the voids between stones (the sieve effect). We also observed numerous artifact turnovers, which are very likely due to the formation and subsequent melting out of ice needles, the burying of flints under 1 – 2 cm of sediment, and a rather marked dispersal of retouch flakes in connection with runoff processes. However, in contrast to the random dispersal model devised by Bowers et al. (1983) in Alaska, our results emphasize a strong domination of unidirectional downslope displacements (Figure 25). We may also wonder if, over the long-term (i.e.,

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

457

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 20. Upslope snowpatch containing numerous aeolian rock-debris (east-facing mountain side of La Mortice).

some tens of years), multidirectional movements do not prevail and if a progressive diffusion of artifacts from high concentration zones to nil concentration zones does not occur, as in the model proposed by Culling (1963, in Kirkby, 1967) to describe creep action. Nevertheless, this last hypothesis seems unlikely. Indeed, except for the movement of flints due to cryoturbation (see above), upslope lateral displacements concern only pieces having a low density (some bones and the charcoals). Hence, most of the pieces situated on the back of the stone-banked lobes will continue their downslope translation. This movement will be accompanied by a deformation of the initial archaeological assemblages due to lateral variations in the displacement velocity of lobes. Displacement is higher in the axis of the lobes than laterally or at the front (Table III). Furthermore, certain pieces entrained by runoff processes will add to this deformation (cf. supra). Owing to the short duration of experiment no. 2, it is not possible to draw general conclusions on the vertical dispersal of archaeological pieces. Indeed, only some burying of flints (sometimes followed by exhumation) on the back of lobes as well as a selective sliding of small retouch flakes into the voids of the stony bank were noted. However, using recent information on stone-banked lobes (Francou, 1990; Bertran et al., 1993; Bertran et al., 1995a, 1995b; Van Steijn et al., 1995), we are able to describe some of the processes that may act on archaeological pieces. Runoff processes as well as some small mudflows will promote the burying of

458

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

Figure 21. Granulometry of particles trapped in upslope (cumulative curve 1) and in downslope (cumulative curve 2) snowpatch.

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

short standard

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

459

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 22. Morphology of pebbles trapped in an upslope snowpatch. Folk’s (1960) diagram. Star: mean morphology of the pebbles.

archaeological pieces, whereas frostjacking will produce the opposite effect (cf. experiment no. 1). The latter mechanism will quickly draw the most voluminous pieces toward the ground surface but will be far less active on small, flat flints that lie parallel to the soil surface. These processes will normally lead to a granulometric segregation similar to that noted for natural stones: pebbles are much smaller in the matrix-rich layer than in the surface pavement of the lobes. Moreover, a portion of the archaeological pieces located at the stony bank will be buried as the lobe advances and will participate in the formation of the lower openwork layer (cf. supra). The rest of these pieces will reach the back of the mud boil, which is formed downward from the stony bank. They will then travel

Table II. Mean yearly displacement of artifacts in each test plot. Artifact Test Plots

1

2

3

4

5

Experiment duration (yr) Mean displacement (cm yr⫺1) Mean displacement (displacements over 50 cm length not taken into account) (cm yr⫺1)

4 4.55 2.45

2 4.75 2.90

3 1.66 1.66

4 3.17 2.47

4 3.55 2.47

460

short standard

VOL. 13, NO. 5

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figures 23 and 24. Diagrams showing the amount and the relative frequency of artifact displacements recorded in each test plot of experiment 2 (abscissa: amount of displacements in cm; ordinates: relative frequency). 1, 2, 3, 4 and 5 refer to test-plot numbers.

short standard

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

461

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figures 23 and 24. Continued.

downslope and experience the same vicissitudes as those described earlier: differential burying of the remaining pieces in the matrix-rich layer, incorporation of some of these pieces in the buried openwork layer, etc. These processes can recur several times before the complete burial of archaeological material. Thus, such mechanisms will cause the redistribution of a single archaeological assemblage in several structurally and texturally different geological levels. We can also expect vertical and lateral granulometric segregation of archaeological pieces, which will be greater in openwork stony layers than in matrixrich layers, and also greater downslope compared to upslope (larger artifacts will tend to remain at the ground surface for longer than the smaller ones and thus will travel further). Lastly, a progressive downslope decrease in the density of the archaeological material can also be expected. In order to validate the proposed model, it must be tested in archaeological sites whose stratogenesis is similar to that of La Mortice. The Paleolithic site of Baume Valle´e appears to possess all the necessary qualities (discussed further on). Alteration of Archaeological Pieces (J.L. Guadelli, H. Plisson, and J.P. Raynal) Alteration mainly concerns faunal material, and, in some cases, this alteration can be quite spectacular. A number of significant observations have been made on our sample after 4 years of exposure to periglacial conditions. Long cracks, for example, have formed on mandibles under the dental implantation zone and on the ascending ramis. These pieces also showed more or less marked gum recession. Long bones showed the development of platy or scaly disintegration (Figure 26)

462

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figures 23 and 24. Continued.

short standard

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

463

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 25. Diagrams showing the strike frequency of artifact displacements in each test plot of experiment 2. Modal orientation in each test plot corresponds to that of local slope. 1, 2, 3, 4, 5: test-plot numbers— 0/20, 21/40, . . . : class boundaries (in grades)— an interval between two successive graduations on radii is equal to 10%.

464

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT Table III. Mean yearly displacement of artifacts in test plot 1 according to their location on the solifluction lobe.

Mean displacement (cm yr⫺1)

Center of the Solifluction Lobe

Side of the Solifluction Lobe

Stone Front

Downslope from the Stone Front

4

1.91

1.24

1.7

and then a “little stick” type fragmentation, which occured from the outside to the inside part of the bone. There was also a surficial peeling of “fresh” long bones (Figure 27). Frost shattering of some long bones occurred along with a relative displacement of the bone fragments (Figure 28). Finally, fossil teeth underwent a significant shattering, the fragments of which are no longer connected (Figure 29). These observations emphasize the importance of the rapidity of damage on faunal remains left in periglacial contexts. However, it should be noted that alteration processes may have been hastened by the preliminary drying and cleaning of the experimental osseous pieces. Similar results were obtained on analogous faunal

short standard

Figure 26. Long bone fragmentation of “plates pile” type.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

465

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

Figure 27. Surface peeling off of a “fresh” long bone (scale in cm).

Figure 28. Frost-shattering of a long bone and relative movement of bone fragments.

466

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

Figure 29. Frost-shaterring of a Saı¨ga tooth.

pieces placed in a cold chamber (at the Centre Ge´omorphologique de Caen) and submitted to freeze – thaw cycles. In these two samples, fragmentation of the same kind and degree occurred at the same places and in the same chronological order (Guadelli and Ozouf, 1994). Thus the specific environmental conditions of the site of La Mortice (e.g., importance of ultraviolet radiation) seem to have little effect on alteration processes. Further details about the observed processes will probably be obtained by microscopic studies as well as by analysis of porosity change. Until now, flints seem to have suffered no or little damage in this periglacial environment. Only one flint was frost-shattered. Moreover, use-wear studies show that raw edges and active edges of flints used to work bones, limestone, hide, and wood experienced no significant modifications. Whatever the scale of observation, traces of use-wear remain perfectly clear after 4 years of experimentation and it can be identified with confidence. Fabric of Archaeological Pieces (P. Bertran and J.P. Texier) In order to appraise the rate of acquisition of an oriented fabric in a solifluctiondominated context, measurements of strike and dip of the main axis (a-axis) of elongated pieces were performed in test plots 1 and 5. Results plotted on a Schmidt

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

467

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

diagram (Figure 18) show that, 3 years into the experiment, pieces remained randomly oriented in the stratification plane (vector magnitude ⫽ 22.52% and 14.14%; the hypothesis of a random distribution of orientations is accepted at the 0.05 level). Nevertheless, on test plot maps we note that some elongated pieces began to rotate and that their main axis tends to take a position parallel to the strike of the flow (Figure 30). These data suggest that acquisition of a high fabric strength typical of solifluction environments (Brochu, 1978; Matthews et al., 1986; Harris, 1987; Bertran et al., 1995a, 1995b) is slow and that talus shift is not very active in this part of the slope. Indeed, this process could rapidly (in one season) give rise to a strong preferred slope-aligned orientation of the clasts (He´tu, 1995). Weak influence of the superficial creep can be, above all, explained by the low gradient of the local slope (10 – 14⬚) (Kirkby, 1967; Abrahams et al., 1984; He´tu, 1995). However, other environ-

Figure 30. Test plot 3. Artifact plans plotted in 1991 and in 1994. Points: flints; stars: elements of fauna (elongated pieces are represented by two points or by two stars); arrows: displacements of artifacts.

468

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

mental elements act in the same way: Upslope from the experimental site, there is no cliff from which rock debris could fall; passage of animals is quite rare; the presence of permafrost hindering the snow creep is likely. Thus, the only mechanisms which could take part in the talus shift are needle-ice action, rain and hail splash, and, perhaps, thermic and hydric variations undergone by deposits. Transfer of Results to Paleolithic Sites (P. Bertran and J.P. Texier) Results concerning archaeological and palynological assemblages are not sufficiently advanced to use them in the interpretation of Paleolithic sites. It will be necessary to wait until the experiments are terminated to develop this aspect of research. Nevertheless, we have already begun to use the results from the geological aspects of the experiments beyond the archaeological framework stricto sensu, such as a morphodynamic study of the slope deposits of northern Aquitaine (Bertran et al., 1995a, 1995b; Ozouf et al., 1995). In this article, we will only report on their implications for interpretation of some Paleolithic sites and for the appraisal of the disturbance rate of archaeological assemblages. Among the different sites studied thus far, Baume Valle´e in the Massif Central (excavated by J.P. Raynal) exhibits the closest genetic similarity to our experimental site. Indeed, numerous criteria give evidence that these deposits were settled by the stacking of stone-banked lobes similar to those described at La Mortice. Such criteria include kind of stratification, dip of layers (10⬚ on average), grading of openwork layers, morphology of limits between the different layers, stones and artifacts strongly slope-oriented, and platy and granular microstructures in matrixrich layers. Thus, it is likely that burial and diagenetic conditions of archaeological levels (Mousterian) were quite similar to those noticed for test plots. Consequently, the same kinds of disturbances as those mentioned earlier and highlighted in active environments (see the previous subsections) are expected. Moreover, conditions suitable for solifluction were obviously maintained for a long time in this site. It should therefore be possible to test the model proposed in the previous section in order to see if the main diagnostic criteria related to solifluction can be identified: archaeological pieces coming from a single assemblage distributed in several texturally different layers; vertical and longitudinal sortings (larger artifacts being preferentially situated downslope and in openwork layers); density of pieces decreasing downslope; elongated pieces exhibiting strong slope orientation; deformation of anthropic structures due to shearing; etc. A new dynamic interpretation of deposits of some Paleolithic sites of Pe´rigord has also been possible using the data obtained within the context of the TRANSIT program. Thus the Mousterian site of Combe Capelle bas (excavated by H. Dibble and M. Lenoir) exhibits the main characteristics of debris-flow deposits: irregular stratification, specific morphology of openwork layers, and no vertical sorting (Texier and Bertran, 1995). Archaeological assemblages linked to these deposits have very likely been heavily disturbanced. In such a dynamic context, the only possibility for good preservation of archaeological levels would be if they were overrid-

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

469

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL.

den by debris flow deposits. Indeed, numerous studies carried out in present-day environments have emphasized that this kind of flow has no erosive action on the substratum on which it spreads (Johnson, 1970; Enos, 1977; Naylor, 1980; Johnson and Rodine, 1984; Hubert and Filipov, 1989). However, the isotropic distribution of archaeological material at Combe Capelle shows that it experienced the successive flows with natural sediments. Thus, the probability of homogeneity among archaeological assemblages collected in this site is very low. In the Upper Paleolithic sites of Laugerie-Haute Ouest (excavations of D. Peyrony, and then F. Bordes and P. Smith) and Combe Saunie`re (excavations of M. Geneste), the very high fabric strength of archaeological pieces and stones strongly suggests that solifluction probably played an important role in their formation processes (Bertran and Texier, 1995). CONCLUSIONS Among the studies carried out with the TRANSIT program, those concerning the natural environment are the most advanced. They permit us to make significant progress in understanding the dynamics and formation of deposits in a periglacial context. Two specific sedimentary processes have been studied: stone-banked solifluction lobes and debris flows. For these mechanisms, we were able to propose sedimentary models and identify diagnostic criteria useful for the interpretation of fossil deposits. Results began to be applied to some Paleolithic sites of the Massif Central and Aquitaine Regions in France. A more reliable reading of the sedimentary paleoenvironments and a better understanding of disturbance processes acting on archaeological assemblages are anticipated from these new data. Among the different Paleolithic sites considered thus far, Baume Valle´e exhibits characteristics very close to those noted in experimental sites. The modifications it suffered during its formation were probably similar to those observed in present environments. The first results on pollen assemblages have shown that in such a periglacial environment, most of the pollen deposited on the ground surface is quickly removed (by runoff and probably by leaching). Furthermore, owing to the very uneven preservation of pollen taxa, an important distortion of the initial spectrum occurs. Although archaeological experiments are still not finished, results obtained at present clearly emphasize the importance of periglacial processes on the distribution and alteration of archaeological data. After 4 years of study, lateral displacements from some centimeters to several meters were observed. Some pieces were buried whereas others were infiltrated in the voids between the stones. Numerous artifacts were turned over and their orientations have changed. Furthermore, the data obtained in active periglacial environments show that, on a more long-term basis, a single archaeological assemblage left in a solifluction-dominated environment will undergo vertical and longitudinal sorting processes and will be scattered in several texturally different geological levels. In the future, we will also focus on the very high alteration rate of bones in such a context. This rate varies with the

470

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT

kind of osseous and dental fragments, and a very significant distortion of the initial faunal assemblage can be expected.

The TRANSIT program is supported by the French CNRS (Action the´matique programme´e “Grands projets d’arche´ologie me´tropolitaine”), by the French Ministry of the Culture and by the Aquitaine Region. We are also grateful to Michel Lenoir who manufactured the experimental flint assemblages used in this program and to Paul Goldberg and Jan Simek, who kindly reviewed the English language of this article.

REFERENCES Abrahams, A.D., Parsons, A.J., Cooke, R.U., and Reeves, R.W. (1984). Stone Movement on Hillslopes in the Mojave Desert, California: A 16-Year Record. Earth Surface Processes and Landforms 9, 365– 370. Benedict, J.B., and Olson, B.L. (1978). The Mount Albion Complex: A Study of Prehistoric Man and the Altithermal. Report No. 1, Center for Mountain Archaeology. Bertran, P., and Texier, J.P. (1994). Structures Se´dimentaires d’un Coˆne de Flots de De´bris (Vars, Alpes Franc¸aises Me´ridionales). Permafrost and Periglacial Processes 5, 155– 170. Bertran, P., and Texier, J.P. (1995). Fabric Analysis: Application to Paleolithic Sites. Journal of Archaeological Sciences 22, 521– 535. Bertran, P., Francou, B., and Pech, P. (1993). Stratoge´ne`se Associe´e a` la Dynamique de Coule´es a` Front Pierreux en Milieu Alpin, La Mortice, Alpes Franc¸aises Me´ridionales. Ge´ographie Physique et Quaternaire 47(1), 93– 100. Bertran, P., Coutard, J.P., Francou, B., Ozouf, J.P., and Texier, J.P. (1994a). New Data on Gre`zes Bedding and Their Palaeoclimatic Implications. In D.J.A. Evans, Ed., Cold Climate Landforms, pp. 437– 454. New York: John Wiley. Bertran, P., Texier, J.P., Coutard, J.P., Ozouf, J.C., and Francou, B., (1994b). Contribution au De´bat sur l’Origine du Litage des Gre`zes. Quaternaire 5(1), 41– 46. Bertran, P., Coutard, J.P., Ozouf, J.C., and Texier, J.P. (1995a). De´poˆts de Pente Calcaires du Nord de l’Aquitaine. Re´partition Stratigraphique et Ge´ographique des Facie`s. Zeitschrift fu¨r Geomorphologie N.F. 39(1), 29– 54. Bertran, P., Francou, B., and Texier, J.P. (1995b). Stratified Slope Deposits: The Stone-Banked Sheets and Lobes Model. In O. Slaymaker, Ed., Steepland Geomorphology, pp. 147– 169. New York: John Wiley. Bottema, S. (1975). The Interpretation of Pollen Spectra from Prehistoric Settlements (with Special Reference to Liguliflora). Paleohistoria 17, 18– 34. Bowers, P.M., Bonnichsen, R., and Hoch, D.M. (1983). Flake Dispersal Experiments: Non Cultural Transformation of the Archaeological Record. American Antiquity 48(3), 553– 572. Brochu, M. (1978). Disposition des Fragments Rocheux dans les De´poˆts de Solifluxion, dans les E´boulis de Gravite´ et dans les De´poˆts Fluviatiles: Mesures dans l’Est de l’Arctique Nord-Ame´ricain et Comparaison avec d’Autres Re´gions du Globe. Biuletyn Peryglacjalny 27, 35– 51. Bullock, P., Fedoroff, N., Jongerius, A., Stoops, G., Tursina, T., et al. (1985). Handbook for Soil Thin Section Description. Wolverhampton: Waine Research Publications. Coutard, J.P. (1985). La Creˆte de Vars (Hautes-Alpes). Exploitation des Donne´es Thermiques. Bulletin du Centre de Ge´omorphologie du CNRS, Caen 30, 85– 98. Coutard, J.P., and Francou, B. (1989). Rock Temperature Measurements in Two Alpine Environments: Implications for Frost Shattering. Arctic and Alpine Research 21(4), 399– 416. Coutard, J.P., and Ozouf, J.C. (1993). Etude des Cycles Gel– De´gel sur les Creˆtes de Vars et des Couniets (Commune de Vars, Hautes Alpes). In Ge´omorphologie et Ame´nagement de la Montagne, pp. 41– 51. Hommage a` P. Gabert. Caen: CNRS.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

471

short standard

GEA(Wiley)

LEFT

INTERACTIVE

TEXIER ET AL. Coutard, J.P., Gabert, P., and Ozouf, J.C. (1988). Etude du Processus de Cryoreptation en Divers Sites de la Haute-Ubaye (Alpes du Sud). Bulletin du Centre de Ge´omorphologie du CNRS, Caen 34, 9– 28. Coutard, J.P., Ozouf, J.C., and Gabert, P. (1996). Modalite´s de la Cryoreptation dans les Massifs du Chambeyron et de La Mortice, Haute-Ubaye, Alpes Franc¸aises du Sud. Permafrost and Periglacial Processes 7, 21– 51. Culling, W.E.H. (1967). Soil Creep and the Development of Hillside Slopes. Journal of Geology 71, 127– 162. Curracy, J.R. (1956). Analysis of Two-Dimensional Orientation Data. Journal of Geology 64, 117– 134. Enos, P. (1977). Flow Regime in Debris Flows. Sedimentology 24, 133– 142. Folk, R.L. (1960). Petrology of Sedimentary Rocks. Austin, Texas: Hemphills. Francou, B. (1990). Stratification Mechanisms in Slope Deposits in High Subequatorial Mountains. Permafrost and Periglacial Processes 1, 249– 263. Guadelli, J.L., and Ozouf, J.C. (1994). Etudes Expe´rimentales de l’Action du Gel sur les Restes Fauniques. ´ labore´ en Os et Premiers Re´sultats. In Artefacts, 9, Actes de la 6e`me Table Ronde “Outillage Peu E en Bois de Cervide´s IV: Taphonomie/Bone Modification, Cedarc, Ed., Treignes, pp. 47– 56. Guillore´, P. (1980). Metho´de de fabrication me´canique ei en se´rie des lanies minces. Institut National d’Agronomie, Paris-Grignon, 22 p. Harris, C. (1987). Solifluction and Related Periglacial Deposits in England and Wales. In J. Boardman, Ed., Periglacial Processes and Landforms in Britain, pp. 209– 223. Cambridge: Cambridge University Press. Harris, C., Gallop, M., and Coutard, J.P. (1993). Physical Modelling of Gelifluction and Frost-Creep: Some Results of a Large-Scale Laboratory Experiment. Earth Surface Processes and Landforms 18, 383– 398. He´tu, B. (1995). Le Litage des E´boulis Stratifie´s Cryonivaux en Gaspe´sie (Que´bec, Canada): Roˆle de la Se´dimentation Nive´o-E´olienne et des Transits Supra-Niveaux. Permafrost and Periglacial Processes 6, 147– 171. Hopkins, D.M., and Giddings, J.L. (1953). Geological Background of the Iatayet Archaeological Site, Cape Denbigh, Alaska. Smithsonian Miscellaneous Collections 121(11), 1– 37. Hubert, F.J., and Filipov, A.J. (1989). Debris-Flow Deposits in Alluvial Fans on the West Flank of the White Mountains, Owens Valley, California, U.S.A. Sedimentary Geology 61, 177– 205. Johnson, A.M. (1970). Physical Processes in Geology. San Francisco: Freeman, Cooper. Johnson, A.M., and Rodine, J.R. (1984). Debris Flow. In D. Brunsden and D.B. Prior, Eds., Slope Instability, pp. 257– 361. New York: John Wiley. King, L. (1990). Soil and Rock Temperatures in Discontinuous Permafrost: Gornergrat and Unterrothorn, Wallis, Swiss Alps. Permafrost and Perglacial Processes 1, 177– 188. Kirkby, M.J. (1967). Measurement and Theory of Soil Creep. The Journal of Geology 75(4), 359– 378. Mackay, J.R., Mathews, W.H., and Mac Neish, R.S. (1961). Geology of the Engistciak Archaeological Site, Yukon Territory. Arctic 14, 25– 52. Mante´, C. (1985). Evolution du Champ de Tempe´ratures dans une Paroi Rocheuse Naturelle: La Cas de la Creˆte de Vars. Bulletin du Centre de Ge´omorphologie du CNRS, Caen 30, 99– 139. Mante´, C. (1988). Recherche sur les Situations Thermiques Types dans une Paroi Rocheuse. De´tection de Pe´riode Propices a` la Cryoclastie. Bulletin du Centre de Ge´omorphologie du CNRS, Caen 34, 181– 221. Mante´, C. (1989). Analyse en Composantes Principales d’un Processus Multiple Non Stationnaire: Une Application a` des Donne´es Me´te´orologiques. Statistique et Analyse des Donne´es 14(2), 25– 53. Mark, D.A. (1973). Analysis of Axial Orientation Data, including Till Fabrics. Geological Society of America Bulletin 84, 1369– 1374. Matthews, J.A., Harris, C., and Ballantyne, C.K. (1986). Studies on a Gelifluction lobe, Jotunheimen, Norway: Chronology, Stratigraphy, Sedimentology, and Palaeoenvironment. Geografiska Annaler 68A(4), 345– 360. McEachran, D.B. (1990). Stereo, the Stereographic Projection Programme. Apple Macintosh Computer, version 1.3.

472

VOL. 13, NO. 5

short standard

GEA(Wiley)

RIGHT

INTERACTIVE

EXPERIMENTAL ARCHAEOLOGY IN A PERIGLACIAL ENVIRONMENT Naylor, M.A. (1980). The Origin of Inverse Grading in Muddy Debris Flow Deposits— A Review. Journal of Sedimentary Petrology 50, 1111– 1116. Nelson, F.E. (1985). A Preliminary Investigation of Solifluction Macrofabric. Catena 12, 23– 33. Ozouf, J.C., Texier, J.P., Bertran, P., and Coutard, J.P. (1995). Quelques Coupes Caracte´ristiques dans les De´poˆts de Versant d’Aquitaine Septentrionale: Facie`s et Interpre´tation Dynamique. Permafrost and Periglacial Processes 6(2), 89– 101. Rouque`s, G., and Caniard, L. (1975). Gel et De´gel des Chausse´es. Note d’Information Technique, Laboratoire Central des Ponts et Chausse´es, Paris, 34 p. Texier, J.P., and Bertran, P. (1995). Les De´poˆts du Site Mouste´rien de Combe-Capelle Bas: Leur Signification Dynamique et Pale´oenvironnementale. Pale´o 7, 27– 48. Van Vliet-Lanoe¨, B. (1998a). Approche Morphologique et Micromorphologique de la Dynamique de Solifluxion dans les Massifs de la Font Sancte et du Chambeyron. Bulletin du Centre de Ge´omorphologie du CNRS, Caen 34, 29– 45. Van Vliet-Lanoe¨, B. (1988b). Le Roˆle de la Glace de Se´gre´gation dans les Formations Superficielles d’Europe de l’Ouest. Processus et He´ritage, The`se d’Etat, Universite´ de Paris I and Centre de Ge´omorphologie du CNRS, Editec, Caen. Van Steijn, H., Bertran, P., Francou, B., He´tu, B., and Texier, J.P. (1995). Models for the Genetic and Environmental Interpretation of Stratified Slope Deposits. A Review. Permafrost and Periglacial Processes 6(2), 125– 146. Washburn, A.L. (1979). Geocryology. A Survey of Periglacial Processes and Environment. London: Edward Arnold. Watson, G.S. (1965). Equatorial Distributions on a Sphere. Biometrika 52(1– 2), 193– 201. Watson, G.S. (1966). The statistics of orientation data. Journal of Geology 74, 786– 797. Wood, R.W., and Johnson, D.L. (1978). A Survey of Disturbance Processes in Archaeological Site Formation. In M. B. Schiffer, Ed. Advances in Archaeological Method and Theory, Vol. 1, pp. 315– 381. New York: Academic Press. Woodcock, N.H. (1977). Specification of Fabric Shapes Using an Eigenvalue Method. Geological Society of America Bulletin 88, 1231– 1236.

Received October 8, 1996 Accepted for publication January 24, 1998

short standard

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL

473