We acknowledge D. Augier (Musée d’Angoulême), R. Vacant (Muséum National d’Histoire Naturelle) M. Boivin and C. Augier for technical assistance, material acquisition and help during preparation of experiences. We also thank the tramplers J. Rouillard, H. Hanus, F. Lafuma, J. Pruvost, S. Allain, E. Allain, G. Allain, T. Legrand, G.r Legrand and P. Rivière.

1 - Introduction

1Trampling by animals is an important taphonomical agent that can considerably modify a bone assemblage. Depending on several parameters (e.g. trampling duration, trampler weights, soil content and viscosity, bone shape…), it can disarticulate skeletons, disperse, reorientate, break and crack bones and create marks on bone surface (Courtin & Villa, 1982 ; Fiorillo, 1984, 1989 ; Andrews and Cook, 1985 ; Behrensmeyer et al., 1986 ; Olsen and Shipman, 1988 ; Auguste, 1994 ; Domίnguez-Rodrigo et al., 2009 ; Blasco et al., 2008 ; Denys & Patou-Mathis, 2014). The resulting bone modifications can be mistaken with other taphonomical agents, such as breakage pattern and surface marks due to predation or scavenge (e.g. crocodile tooth marks) and human tool manufacture or tool-assisted butchery (e.g. cut marks ; Fiorillo, 1984 ; Behrensmeyer et al., 1986 ; Buc, 2011 ; Njau, 2012). The latter can have great consequences on archeological interpretations (e.g. McPherron et al., 2010, 2011; Domínguez-Rodrigo et al., 2010, 2011).

2Several taphonomical trampling experiments on bones have been done to study and identify spatial movements, breakage and surface marks patterns from other taphonomical agents (Courtin and Villa, 1982; Fiorillo, 1984, 1989; Andrews and Cook, 1985; Behrensmeyer et al., 1986; Olsen and Shipman, 1988; Blasco et al., 2008; Domίnguez-Rodrigo et al., 2009). However, most of them took place in coarse sands and knowledge of trampling effects on a bone assemblage in swampy environment sediments is lacking (A.K. Behrensmeyer, pers. com., 2016).

3Here we report the results of two exploring experiments done to understand the effects of trampling on bones in soft clay and fine sandy sediments. Experiments took place during the fourth excavation campaign in August 2013 at the Lower Cretaceous fossiliferous Angeac-Charente locality (Charente, France). They have been done on the way to compare with the modifications observed on the dinosaur bone assemblage fossilized in the swampy-deltaic environment of Angeac (Néraudeau et al., 2012; Rozada, 2014; Rozada et al., 2014).

2 - Material and methods

4Bone material has been collected in a mass grave in a disused quarry near Angeac. They all belong to adult individuals, not weathered and devoid of periosteum. Elements have been chosen on the way to represent a diversity of shape and size (scapulae, long bones and ribs). Location of bones before and after experiments have been recorded with photographs.

2.1 - Experiment 1

5The first experiment focuses on the study of spatial displacements of bones in soft clayey sediments. Ten sheep bones (two scapulae, articulated humerus-ulna-radius, one tibia and four ribs) have been horizontally placed close together at the surface of soft clays in the Angeac site (fig. 1). The soft, watery clay layer was about 50 cm thick, overlying harder clays. For one hour, five persons (one adult and four children of different ages and weight) wearing boots trampled the area freely (fig. 1).

2.2 - Experiment 2

6The aim of the more rigorous second experiment was to study marks on bone surface left by trampling in different soft fine sediments. Four sets of ten fresh bones have been chosen: one scapula, one humerus, one metatarsal, one metacarpal, and four ribs (one complete and large, one proximally broken and large, one thin and one small) of sheep and two small and thin dog ribs (fig. 2). One metatarsal shows a pathology (see fig. 2, before experiment C1). Additionally, a fossilized dinosaur bone splinter has been added in each set to test if trample marks can appear on mineralized bone surfaces. The material has been cleaned with soft brush and acetone. As control, sheep bone surfaces were replicated before experiment with silicone RTV silastic 3481 and fossil bone splinter surfaces with RTV silicone paste.

7Four different sediments were used in this experiment. They correspond to the four lithological units described at Angeac-Charentes locality, in which dinosaur bones have been preserved: fine light gray sands (C1), fine yellow sands (C2), brown clays (C3) and dark-gray clays (C4). Four holes have been dug in the gray clays (C4) in the Angeac site, covered by hermetic plastic bags, and filled by each sediment saturated in water. Bones of each set were positioned horizontally at the surface and covered by less than one centimeter of sediment (fig. 2). During experimentation, four people wearing boots or trainers have taken successively ten steps in each hole (fig. 2). To avoid any sediment pollution, the tramplers washed their soles before stepping each hole. After the experiment, bones were excavated, cleaned with soft brush and water, and surfaces were examined with a binocular magnifier and compared with silicone molds. 1 kg of each sediment was finely sieved to assess sediment components and grain size.

3 - Results

3.1 - Experiment 1

8During the experiment, children firstly tried to trample the bones deliberately. After a few seconds of trampling, the bones were not visible any more, buried in a layer of very fluid clays (fig. 1). During the rest of the experimentation, children were playing together in a large area, walking, running, jumping and sliding.

9Ones excavated, bones appeared in a horizontal plane, at the limit between liquid and more resistant clays (fig. 1). Three bones have been highly horizontally displaced: the tibia (T), one rib also highly reoriented (R2), and one rib never recovered (R1). The two other ribs (R3, R4) and one scapula (S2) were broken. The rib R4 shows a transversal fracture, with the two fragments very close to each other and showing a slight vertical reorientation. For the rib R3, the proximal and distal fragments are present, with the proximal one showing oblique (spiral) fracture and distal one being turned over (R3 prox, R3 dist). However, the central fragment is missing (fig. 1, R3). For the scapula (S2), one fragment stayed in place but rotated to a vertically direction, and some bone splinters have been found on and in contact with the main part of the scapula (fig. 1, S2). Finally, a fragment of the superficial layer of the proximal epiphysis of the tibia has been pulled off (fig. 1, T).

10Photographs: cL. Cazes, MNHN, 2013.

3.2 - Experiment 2

11During experiment 2, we have observed that one metapodial from C4 stayed stuck on the boot of the trampler and was deposited outside the hole (fig. 2). After the experiment, bones trampled in C1, C3 and C4 showed a lot of displacements and orientation change, while they are limited in C2 (fig. 2). One rib from C1 has been broken (transversal fracture), some bones are cracked, and several little splinters has been found after sieving. One metapodial from C4 shows an oval-shaped crushed (fig. 3).

12Numerous typical trample marks have been observed on bone surfaces: stripes, parallel striations and scratches (fig. 3A-E, tab. 1). They appear preferentially on rounded, convex, flat or prominent surfaces of bones. Stripes are very fine, and shallow grooves often linear and simple, but sometimes doubled and/or slightly curved (fig. 3A). Parallel striations are sets of parallel fine linear stripes (fig. 3B-D). They can appear as isolated patches, but often as elongated patches along bone shafts and ridges, with striations being oriented perpendicular or oblique to the long axis of the shaft (fig. 3B, D; tab. 1). Scratches are shorter, deeper and wider marks. They are sometimes associated with (that means parallel to or as a continuation of) stripes or parallel striations (fig. 3E, see arrow). In some cases, two or more marks are crossing each other (fig. 3C; tab. 1).

1329 of the 33 sheep bones or bone fragments (85 %), exhibit trample marks. 15 of them (39 %) show more than one or two isolated marks (tab. 1). Most of the trample marks are stripes (67 %), then parallel striations (45 %) and scratches (42 %). On 6 bones (18 %), marks have been found intercrossed (fig. 3C). The surface of almost all bones trampled in fine sands (C1, C2) and dark clays (C4) have been marked, while few isolated marks have been observed on 4 bones trampled in yellow clays (C3). Dog ribs and fossil dinosaur splinters does not exhibit any surface modifications.

14Results of sediment sieving show that the composition of each sediment includes mostly clays and fine silts (< 50 µm), silty to fine sand quartz grains (50-100 and 100-250 µm) and few pyrite, calcareous, lignite and quartz coarser grains (>250 µm; tab. 1).

4 - Discussion

15Trampling done in the two experiments has considerably modified the artificial bone assemblage. Indeed, bones have been horizontally and vertically displaced, their vertical and horizontal orientations have changed, they have been broken, knocked, and their surfaces have been affected by trample marks.

16Bones have been horizontally displaced from several centimeters to decimeters, and probably more than one meter for the missing bones of the experiment 1 (R1, R3). Horizontal reorientation has reached at least 180° (fig. 1, R2). However, vertical displacements (few centimeters) and reorientation (tens of degrees) of bones is more limited, probably due to the thin thickness of the fluid sediment layer in the two experiments. These displacements and position changes occurred by driving a bone, carrying a stuck bone to the foot (fig. 2, C4) or moving the soft soil. The limited displacements and orientation changes in C2 compared to other sediments are probably due to the weaker fluidity of this sandy sediment. Trampling in soft clays can considerably displace and reorient bones, and even transport them over short distances.

17Only ribs and scapulae have been broken. That preferential breakage pattern is due to their flattened shape and less compact/density compared to limb bones (Eberth et al., 2007). The recurrence of fragments of the same broken bone preserved in close proximity indicates an in-situ breakage. An unexpected result is the breakage of the superficial part of the proximal epiphysis of the tibia. As a result, the proximal epiphysis seems to be abraded and mimic erosion due to transport. Thus, trampling in soft clays can break in-situ preferentially less compact bones and can mimic abrasion.

18Morphology, localization and orientation of traces found on bone surfaces are very similar to previously described trample marks (Andrews and Cook, 1985 ; Behrensmeyer et al., 1986 ; Olsen and Shipman, 1988). They appear preferentially on convex, flat or prominent parts, where the siliceous grains are more able to be pushed across the bone (Behrensmeyer et al., 1986). They are often oriented perpendicular or oblique to the long axis of the bones, especially on metapodial shafts, because long bones tend to turn around their long axis during trampling.

19The interesting thing that was not expected is the high number of these marks, even created in dominant-clayey sediments (tab. 1). The fraction of sediment used during experiment liable to create marks are mostly the silty and fine sandy quartz grains, present as little isolated aggregates in clay and observed inside the marks after the experiment. That thinness and low proportion of quartz grains are enough to considerably modify bone surfaces.

20The lower occurrence of marks on bones trampled in C3 is probably due to the lower proportion of quartz grains in C3 (< 3 %, tab. 1). The presence of coarse pyrite minerals can also have created some marks and can explain the high proportion of scratches on bones. The absence of traces on dog ribs is probably the result of a double bias of size and the peripheral position in the hole (fig. 2). As expected, no traces have been found on the surface of fossil bone splinters due to the harder mineralized matter.

5 - Conclusion

21In conclusion, these two trample experiments give preliminary information about trampling of a bone assemblage in fine soft sediments. They show that trampling of bones in fine sandy and clayey sediments considerably displace and reorient bones can transport them to short-distances, create in-situ breakage, knocking in, erosion-like patches and numerous trample marks. For a fossilized bonebed study application, trampling in swampy environment can explain in-situ disarticulation, dispersion and reorientation of bones originally in connection. Further studies of bone trampling in swampy environment are needed, including a better control of the viscosity, grain composition and clay mineral proportion of sediments.

22The results of these experiments are in comparison with modifications observed on fossilized dinosaur bones from Angeac locality. These experimental and fossilized marks are also in the process of microscopic study with confocal microscopy, and compared with crocodile tooth marks and cut-marks on the way to formalize the identification of these three taphonomical agents.