Vertebrate Ichnology Behavioural, Sedimentological and Palaeoecological Aspects of Vertebrate Ichnology

Home , Damatria, Excirolana chiltoni, Festningen Sandstone, Navahopus, Ornithichnites, Parabrontopodus

Jesper Milàn

Ph.D. Thesis Department of Geography and Geology, Geology Section Faculty of Science University of Copenhagen April 2007 Jesper Milàn – Vertebrate Ichnology Title ______

Behavioural, Sedimentological and Palaeoecological Aspects of Vertebrate Ichnology

Ph.D. thesis

Jesper Milàn Department of Geography and Geology Faculty of Science University of Copenhagen April 2007

Superviser: Docent Nanna Noe-Nygaard Department of Geography and Geology University of Copenhagen

i Jesper Milàn – Vertebrate Ichnology Title ______

“A footprint is not an organism but the by-product of dynamic contact between an organism and its environment” - Donald Baird 1957.

ii Jesper Milàn – Vertebrate Ichnology Preface ______

Preface

This thesis is the result of a three-year Ph.D. programme supported by the Faculty of Science, University of Copenhagen and conducted at the Department of Geography and Geology – Geology Section, University of Copenhagen, under supervision of Docent Nanna Noe-Nygaard, Institute of Geography and Geology – Geology Section, University of Copenhagen.

In accordance with the guidelines from the Faculty of Science, University of Copenhagen, the thesis comprises an introduction, overview and discussion and the following 12 papers.

1. Milàn, J. & Gierlinski, G. 2004. A probable thyreophoran (Dinosauria, Ornithischia) footprint from the Upper Triassic of southern Sweden. Bulletin of the Geological Society of Denmark, v. 51, p. 71–75. 2. Milàn, J., Clemmensen, L.B. & Bonde, N. 2004. Vertical sections through dinosaur tracks (Late Triassic lake deposits, East Greenland) – undertracks and other subsurface deformation structures revealed. Lethaia, v. 37, p. 285–296. 3. Milàn, J., Avanzini, M., Clemmensen, L.B., Garciá-Ramos, J.C. & Piñuela, L. 2006. Theropod foot movement recorded from Late Triassic, Early Jurassic and Late Jurassic fossil footprints. In Haris et al. (eds.). The Triassic-Jurassic Terrestrial Transition. New Mexico Museum of Natural History and Science Bulletin, v. 37, p 352–364. 4. Milàn, J., Loope, D.B. & Bromley, R.G. (submitted). Dinosaur tracks showing evidence of individual behaviour from the Lower Jurassic Navaho Sandstone, Coyotes Buttes locality, Utah, USA. Acta Palaeontologia Polonica. 5. Milàn, J. & Loope, D.B. 2007. Preservation and erosion of theropod tracks in eolian deposits; examples from the Middle Jurassic Entrada Sandstone, Utah, USA. The Journal of Geology, v. 115, p. 375–386. 6. Milàn, J. & Bromley, R.G. 2005. Dinosaur footprints from the Middle Jurassic Bagå Formation, Bornholm, Denmark. Bulletin of the Geological Society of Denmark, v. 52, p. 7–15. 7. Mateus, O. & Milàn, J. (Submitted). Ichnological evidence for giant ornithopod dinosaurs in the Late Jurassic Lourinhã Formation, Portugal. Oryctos.

iii Jesper Milàn – Vertebrate Ichnology Preface ______

8. Hurum, J.H., Milàn, J., Hammer, Ø., Midtkandal, I., Amundsen, H. & Sæther, B. 2006. Tracking polar dinosaurs - new finds from the Lower Cretaceous of Svalbard. Norwegian Journal of Geology, v. 83, p. 397–402. 9. Milàn, J., Bromley, R.G., Titschack, J. & Theodouru, G. (in press). A diverse vertebrate ichnofauna from a Quaternary eolian oolite from Rhodes, Greece. In: Bromley, R.G., Buatois, L.A., Márango, M.G., Genise, J.F. & Melchor, R.N. (eds.), Sediment-organism interactions: A multifaceted ichnology. SEPM Special Publications, v. 88. 10. Milàn, J., Clemmensen, L.B., Buchardt, B. & Noe-Nygaard, N. (in prep). A Late Holocene tracksite in the Lodbjerg dune system, northwest Jylland, Denmark. New Mexico Museum of Natural History and Science Bulletin. 11. Bromley, R.G., Uchman, A., Milàn, J. & Hansen, K.S. (in press). Rheotactic Macaronichnus, and human and cattle trackways in Holocene beachrock, Greece: reconstruction of palaeoshoreline orientation. Ichnos. 12. Milàn, J. & Bromley, R.G. (in press). Do shod humans leave true tracks?. Ichnos.

iv Jesper Milàn – Vertebrate Ichnology Acknowledgements ______

Acknowledgements

During my work on this thesis I have had the fortune to work together with a number of different researchers, each internationally recognized expert in their respective fields, and who in many cases appear as co-authors on my works. I thank my superviser Nanna Noe-Nygaard, Department of Geography and Geology, University of Copenhagen, for her patient supervision and her open mindedness to my many wild ideas. Richard G. Bromley, University of Copenhagen, has been a constant source of support and inspiration to new projects and his efforts in proof reading and improving my English has been endless. Octávio Mateus and all the staff and various volunteers at Museu da Lourinhã have given me numerous unforgettable excavation seasons and fruitful co-operations during the years. David B. Loope, University of Lincoln, Nebraska, are thanked for his hospitality and excellent co-operations on the fieldwork with tracks from the Jurassic sandstones of Utah. Andrew S. Gale, University of Portsmouth, is thanked for a very enjoyable track hunt around the Isle of Wight. I thank Jørn H. Hurum, Natural History Museum, Oslo for his hospitality and great co-operations. Marco Avanzini, Museo Tridentino di Scienze Naturali, Italy; Gerard Gierlinski, Polish Geological Institue, Poland; Jürgen Titschack, University of Erlangen, Georgios Theodorou, University of Athens, Alfred Uchman, Institute of Geological Sciences, Poland; Lars B. Clemmensen, Niels Bonde and Bjørn Buchardt all from Department of Geography and Geology, University of Copenhagen, have all been important partners in the various papers and projects. Eckart Håkansson, Department of Geography and Geology, University of Copenhagen, has always been helpful with good advises, proof reading and coffe drinking. I would like to thank the staff and students of the Department of Geology and Geography, Section for Geology, and the Geological Museum, University of Copenhagen for making my time here as enjoyable and memorable as it has been. If anybody feels that I have forgotten them in the previous listing please enter your name here ______. The music of Lumsk, Tiamat, Rammstein and Theatre of Tragedy has been a great relief during the long hours of writing. And finally I sincerely thank my wonderful wife, Inken Mueller-Töwe for her support and patience with me, and my parents and family who always supported my choice to become a palaeontologist instead of getting a real job.

v Jesper Milàn – Vertebrate Ichnology ______

vi Jesper Milàn – Vertebrate Ichnology Contents ______

Contents

Title…………………………..………………..…………….……………….…………….………..i

Preface………………………………………..…………………………………….………………iii

Acknowledgements………………………...…………………….……………….…………………v

Contents……………………………………..…………………………………………..…………viii

Abstract……………………………………………………..……...……………..………………….1

Introduction………………………………………..…………………………………………………3

Summary of papers………………………………………………………………….……………….7

Material and methods…………………………………………………………….…………………21

Discussion…………………………………………………………………………………………..23

Conclusions…………………………………………………………………………………………30

References………………………………………………………………..…………………………32

Appendix:

Paper 1: A probable thyreophoran (Dinosauria, Ornithischia) footprint from the Upper Triassic of southern Sweden Paper 2: Vertical sections through dinosaur tracks (Late Triassic lake deposits, East Greenland) – undertracks and other subsurface deformation structures revealed. Paper 3: Theropod foot movement recorded from Late Triassic, Early Jurassic and Late Jurassic fossil footprints. Paper 4: Dinosaur tracks showing evidence of individual behaviour from the Lower Jurassic Navaho Sandstone, Coyotes Buttes locality, Utah, USA. Paper 5: Preservation and erosion of theropod tracks in eolian deposits; examples from the Middle Jurassic Entrada Sandstone, Utah, USA. Paper 6: Dinosaur footprints from the Middle Jurassic Bagå Formation, Bornholm, Denmark. Paper 7: Ichnological evidence for giant ornithopod dinosaurs in the Late Jurassic Lourinhã Formation, Portugal. Paper 8: Tracking polar dinosaurs - new finds from the Lower Cretaceous of Svalbard.

vii Jesper Milàn – Vertebrate Ichnology Contents ______

Paper 9: A diverse vertebrate ichnofauna from a Quaternary eolian oolite from Rhodes, Greece. Paper 10: A Late Holocene tracksite in the Lodbjerg dune system, northwest Jylland, Denmark. Paper 11: Rheotactic Macaronichnus, and human and cattle trackways in Holocene beachrock, Greece: reconstruction of palaeoshoreline orientation. Paper 12: Do shod humans leave true tracks?.

viii Jesper Milàn – Vertebrate Ichnology Abstract ______

Abstract

Twelve case studies, each representing different aspects of vertebrate ichnology, are published in order to examine the applications of vertebrate ichnology in reconstructing behaviour, sedimentological properties, and palaeoecology. Tracks of sauropods and thyrophorean dinosaurs are for the first time described from the Upper Triassic of southern Sweden and the Middle Jurassic of Bornholm, and a gigantic track from the Late Jurassic of Portugal is evidence of a hitherto unknown large-sized ornithopod. New finds of theropod and ornithopod tracks from the Early Cretaceous of Svalbard are important biogeographical markers for the distribution of dinosaurs living within the polar latitudes. A Late Pleistocene to Holocene trackfauna described from an aeolinite on the southern part of Rhodes, provides evidence of a diverse fauna, including the first ever evidence of camels on prehistoric Rhodes. Vertical sections cut through theropod tracks from the Late Triassic of Greenland exposed undertracks and sub-sediment deformation structures reflecting foot movements during the time of contact between the animal and the substrate. When compared with similar sub-sediment deformation structures in and around tracks from the Early- and Late Jurassic, a difference in walking dynamics from the Late Triassic and Early to Late Jurassic theropods is demonstrated. Two examples of dinosaur trackways, showing evidence of unique individual behaviour are described from the Lower Jurassic Navajo Sandstone of southern Utah. A sauropodomorph trackway heading up the sloping front of a dune, show the animal to have walked sideways upward for the first part of the trackway, and then to have changed orientation to walk head-on upward for the last part of the trackway. Another trackway showing the complete trace of a theropod walking up the sloping front of a dune, lying down producing impressions of hands, the elongated metatarsi, ischial callosity and tail, before it rises up again and continues upslope. The effects of erosion and undertrack formation on the morphology of theropod tracks are described from the Middle Jurassic Entrada Sandstone of Utah. The tracks were originally emplaced in fine, horizontally laminated sand-sheet deposits, and are today exposed in a wide range of track morphologies as a result of preservational styles and degrees of erosion. The Late Holocene Lodbjerg dune system from the north-western coast of Denmark has revealed a diverse trackfauna comprising a wide spectre of tracks from both wild and domesticated species. The faunal composition from the tracksite is comparable with faunal remains from archaeological excavations of nearby contemporary settlements. Lithified Roman-aged beachrock sediments from Pefkos, Rhodes, has preserved shore-parallel tracks and trackways of humans and livestock together. The

1 Jesper Milàn – Vertebrate Ichnology Abstract ______tracks and trackways occur together with the polycheate tracefossil Macaronichnus, which orientation supports the shore-parallel orientation of the trackways. Among the fossil human tracks are examples of tracks from humans wearing shoes. This raises the problem of ichnotaxonomical nomenclature of shod human tracks, and it is proposed that they should not be treated ichnotaxonomically as wearing of shoes represents both a behavioural choice and reflects technological developments in shoe-wear. ______

2 Jesper Milàn – Vertebrate Ichnology Introduction ______

Introduction

The first scientific records of vertebrate ichnology dates back to the early – mid 1800, where Edward Hitchcock in a series of lengthy papers and books described the ichnofauna of Connecticut Valley (Hitchcock 1836, 1848, 1858, 1865), a work that was followed up by several large works by Richard S. Lull (1904, 1915, 1953). Up through the latter half of 1900 vertebrate ichnology lived a somehow subdued existence in the shadow of vertebrate palaeontology, with only sporadic larger works popping up (Haubold 1971; Sarjeant 1974). Within the last 30 years, however, vertebrate ichnology has experienced an enormous rise in interest, initiated by the First International Symposium on Dinosaur Tracks and Traces at the New Mexico Museum of Natural History and Science in Albuquerque, New Mexico (Lockley 1987), and followed up by several important books on the topic (Leonardi 1987, 1994; Gillette & Lockley 1989; Thulborn 1990; Lockley 1991; MacDonald 1994; Lockley & Hunt 1995; Tresise & Sarjeant 1997; Lockley & Meyer 2000) and an explosive increase in number of publications on the topic. While the majority of the works on vertebrate ichnology hitherto have dealt with Mesozoic, and especially dinosaur ichnology, recent initiatives have focused on Quaternary, and especially hominid ichnology (Kim et al. 2004), and this year New Mexico Museum of Natural History and Science will host the first Cenozoic Vertebrate Track Symposium. A fossil vertebrate track is much more than just the mere impression of the trackmakers foot in the substrate. In reality a track is a complex three-dimensional structure whose morphology is dependent on the local sedimentary conditions plus any simultaneous foot movements exercised by the trackmaker during the time of contact between the animal and the substrate (e.g. Allen 1989, 1997; Gatesy et al. 1999; Manning 2004; Milàn 2006; Milàn & Bromley 2006, in press b; Milàn et al. 2004, 2006). In contrast to body-fossils, tracks and trackways are biogenic sediment structures and are dependent on different taphonomical processes than skeletons in order to be preserved. Tracks are therefore likely to be preserved in sedimentary environments where no body fossils are preserved, and further, an animal potentially leaves millions of tracks during its lifetime, but ultimately one skeleton. Tracks are thus very important sources of additional information about past biodiversity and animal behaviour, soft-tissue morphology and distribution in the pedal parts; information that in many instances are unobtainable from the study of skeletons alone.

3 Jesper Milàn – Vertebrate Ichnology Introduction ______

One important factor to take into consideration when studying fossil tracks and trackways is the phenomenon of undertracks. When a track is emplaced in a layered heterogeneous substrate, the weight of the trackmakers foot is transmitted down and outward into the surrounding sediment creating a stacked succession of undertracks along the subjacent horizons (Fig. 1).

Figure 1. True track and undertracks. A, Experimental setting with an emu track emplaced in an artificial layered heterogeneous substrate, allowing the package to be split along several subjacent horizons. B, The true track at the surface is the direct impression of the trackmakers foot, and has preserved fine anatomical details like number and arrangement of digital pads, claw imprints and skin texture. C, The track is still easily recognizable as an undertrack along the horizon one cm below the tracking surface, but appears more rounded, less detailed and with a shallower relief. D – F, The undertracks becomes successively shallower and less detailed downward along each subjacent horizon, until it no longer is possible to recognize the track. Figure modified after Milàn & Bromley (in press b).

4 Jesper Milàn – Vertebrate Ichnology Introduction ______

The phenomenon of undertracks was already noted by Hitchcock (1958), who depicts the same track exposed at successive, subjacent sediment surfaces. If not realized, undertracks can be the source of confusion and misinterpretation, as they make the track seem larger, less detailed and more rounded than the true track. Experimental work with track and undertrack formation has helped to illuminate the spectre of morphological variation of undertracks (Allen 1997, Manning 2004; Milàn & Bromley 2006, in press b). Another factor able to strongly alter the appearance of a track is erosion. When a track becomes exposed to sub aerial erosion, the shape will gradually disintegrate and fine anatomical details will be lost (Henderson 2006) and a track exposed to severe erosion can be hard to distinguish from an undertrack (Milàn & Bromley 2006). Furthermore, erosion can alter the total size and morphology of a track. This is especially the case with tracks originally emplaced in deep soft substrates where the trackmakers foot have sunk to a considerable depth below the tracking surface, thereby reflecting not only the shape of the foot but also the actual movements of the lower parts of the limb, creating an elongation of the track at the tracking surface (Gatesy et al. 1999). In soft sediments, also the different parts of the foot penetrates the sediment to different depths due to differentiated weight loads on the different parts of the foot. When a track comprising all these factors are exposed to erosion, the morphology will vary considerable according to which depth it is exposed (Fig. 2).

Figure 2. The changes in morphology of a track exposed to different degrees of erosion. The example is a plaster cast of an emu track emplaced in soft mud, and afterward sectioned horizontally to simulate erosion of a track with the sedimentary infill still in place. A, section cut just below the tracking surface. B, section cut 14 mm below the tracking surface. C, section cut 25 mm below the tracking surface. D, section cut 38 mm below the tracking surface. Notice how the overall dimensions of the track becomes smaller with depth, and that the individual parts of the track becomes separated with depth, until only the most deeply parts are present, in this case the distal part of the impression of the middle digit. Figure based on experimental data from Milàn & Bromley (2006).

5 Jesper Milàn – Vertebrate Ichnology Introduction ______

When all the these factors are taken into consideration - i.e. that the track encountered can be the result of the foot morphology of the animal, the foot movements exercised by the animal, the consistency of the substrate at the time of tracking, the formation of undertracks, erosion and finally can turn out to have been emplaced by a hitherto unknown animal - then it becomes really difficult to interpret and determine the original morphology and origin of the track. The aim of this project is to explore the utilizations and applications of vertebrate ichnology, in using tracks to (1) reconstruct palaeofauna compositions and to identify hitherto unknown fauna elements; (2) to identify examples of trackmaker behaviour from tracks and trackways; and (3) to describe the influence erosion and sedimentology has on the present day appearance of a vertebrate track. This is carried out as a series of 12 case studies representing vertebrate tracks spanning in time from the Late Triassic to the Late Holocene, and sedimentological settings including aeolian systems, floodplain/lacustrine deposits, and beach deposits. Each case has been treated as a separate study. Six of the cases are published, three are accepted for publication, two manuscripts are at present day still in review and one manuscript is in the final stage of preparation, still awaiting the last 14C datings needed in order to finish the sedimentological descriptions and finalize the manuscript for submission. The style of formatting and language will vary from manuscript to manuscript as they are reproduced in the style of the individual journals (see appendix).

6 Jesper Milàn – Vertebrate Ichnology Summary of papers ______

Summary of Papers

Paper 1 A probable thyreophoran (Dinosauria, Ornithischia) footprint from the Upper Triassic of southern Sweden Jesper Milàn & Gerard Gierlinski Bulletin of the Geological Society of Denmark, 2004. v. 51, p. 71–75.

Summary An unusual tridactyl dinosaur track from the collection of the Geological Museum, University of Copenhagen is described. The track was found 50 to 55 years ago in the Rhaetian coal-bearing strata in the Gustav Adolf Mine, near Höganäs in the southern Sweden. The track is preserved as a natural cast of silty sandstone in the roof of the mine and was originally one of four tracks retrieved from the mine. However, attempts to locate the last three specimens have so far been futile. The track is tridactyl with short rounded digits with blunt claws, and measures 26 cm in length and 28.5 cm in width. The morphology of the track, being broader than it is long, with short, broad digits is consistent with that of tracks from thyreophoran dinosaurs from the Early Jurassic of Poland. The track shows no evidence of having been altered by penecontemporaneous erosion, excluding the possibility that the track is the eroded remains of a theropod track. Further, the trackwalls are vertical and well-defined around the digits, indicating that it is indeed a true track and not an undertrack, as undertracks have shallower relief, and sloping trackwalls. Thus, the Late Triassic age of the described footprint suggests that the track represents the hitherto oldest known footprint of a large thyreophoran dinosaur, and it is the first evidence of non- theropod dinosaurs in southern Sweden.

7 Jesper Milàn – Vertebrate Ichnology Summary of papers ______

Paper 2 Vertical sections through dinosaur tracks (Late Triassic lake deposits, East Greenland) – undertracks and other subsurface deformation structures revealed. Jesper Milàn, Lars B. Clemmensen & Niels Bonde Lethaia, 2004, v. 37, p. 285–296.

Summary The Late Triassic lake sediments of East Greenland contain abundant tracks and trackways of small to medium sized theropod dinosaurs. The fine-grained nature of the sediment allows preservation of fine anatomical details as well as evidence of any foot movements occurring during the time of contact between the animal and the substrate. This study examines two theropod tracks; a rather diffuse theropod track preserved on the upper surface of a red heterolithic mudrock, but in a badly eroded and vaguely recognizable state. The other track is a better preserved track from the upper surface of a slab of greyish mudrock. In order to study the formation of undertracks and other subsurface deformation structures resulting from foot movements during the stride, both slabs were sectioned vertically at closely-spaced intervals, perpendicular to the length of the axis of the impression of digit III. Each of the sections through the tracks was subsequently polished and the sections were scanned in order to reveal internal structures. Below the digit impressions of both tracks were well-defined undertracks formed to considerable depth below the digit impressions. The undertracks were cut by deep and narrow claw imprints at the distal end of each digit impression. Marginal ridges surrounding the track at the tracking surface can be recognized around the undertracks along the subjacent horizons. The marginal ridges around the tracks were asymmetrically developed on each side of the individual parts of the track. This is caused by an outward movement of the proximal part of the foot, probably occurring when the theropod was kicking off towards the next step. This asymmetry is also evident in the undertracks, demonstrating that it was caused by a powerful consistent movement of the foot. The method of studying vertebrate tracks in vertical section has proved to be a very useful tool for acquiring information crucial to the correct interpretation of both the trackmaker and the substrate consistency at the time of trackmaking, information that normally is concealed below the tracking surface. One of the tracks was initially identified as an undertrack due to its shallow, rounded and undefined appearance. By studying the track in vertical section, however, it was demonstrated that the undefined appearance of the tracks was due to present day subsurface erosion

8 Jesper Milàn – Vertebrate Ichnology Summary of papers ______of the track, and that the real undertracks were present in the subjacent layers. In both the studied tracks, the impression of the claws were present as deep cuts down through the sediments, and the narrowness of the cut shows the claws to be laterally compressed sharp claws. The study shows that dinosaur tracks can display large structural variation, when studies in different cross sections of which many are not recognizable as tracks at all. This is important to bear in mind when studying exposures of continental deposits, as horizons with apparent disturbance of the layers may be the result of trampling by vertebrates inhabiting the area.

Paper 3 Theropod foot movement recorded from Late Triassic, Early Jurassic and Late Jurassic fossil footprints. Jesper Milàn, Marco Avanzini, Lars B. Clemmensen, Jose Carlos Garciá-Ramos & Laura Piñuela In Harris et al. (eds.). The Triassic-Jurassic Terrestrial Transition. New Mexico Museum of Natural History and Science Bulletin, 2006, v. 37, p 352–364

Summary This study is a continuation and expansion of the study in paper 2, this time including analysis of additional Late Triassic material and description of and comparisons with tracks from Early Jurassic and Late Jurassic. This is done in order to examine whether the tracks reveal any differences in theropod walking dynamics from the Late Triassic to the Late Jurassic. The methods from paper 2 of studying the tracks in vertical sections are supplemented by the study of the geometry of well preserved natural casts of tracks. Careful study of the deformation structures occurring in and around dinosaur tracks provides valuable information about the walking kinematics and gait of the track maker. The method of studying tracks in vertical section has previously proven successful in obtaining additional details about the walking kinematics that rarely could be obtained from studying the true track at the surface, especially if these have been exposed to erosion or are poorly preserved Thirteen theropod tracks including ages from the Late Triassic, Early Jurassic and the Late Jurassic were examined for deformation induced by foot movements during a stride. The tracks from the Late Triassic and Early Jurassic are preserved as true tracks except one that is preserved as

9 Jesper Milàn – Vertebrate Ichnology Summary of papers ______a natural cast. All the studied Late Jurassic tracks are preserved as deep, three-dimensionally preserved natural casts. The studied tracks from the Late Triassic of Greenland and the Early Jurassic of Italy both show an outward deformation resulting from an outward rotation of the foot during the stride. This deformation falls into two distinct types. One type is formed by digit III being twisted outward during the kick-off and the second type is formed by an outward translocation of the proximal parts of the foot and especially digit IV, which becomes deeply impressed into the substrate. The Late Triassic and Lower Jurassic tracks are impressed into the substrate in a way that is different from the Late Jurassic tracks. In the former, the base formed by the bottom of digits II and IV is almost horizontal, with digit IV being only slightly more deeply impressed than digit II. In the Late Jurassic tracks the base is sloping inwards as digit II is significantly more deeply impressed than digit IV. In addition to that, the Late Jurassic tracks, which are preserved as deep natural casts, show an even, outward deformation of the whole track. The study of the tracks shows that theropods adopted different walking strategies at different times. Taken all together or singularly, true tracks, undertracks, natural casts, and subsediment deformation structures observed in vertical sections, provides unique information about the walking kinematics of track makers, and are thus important factors to incorporate in the study of fossil tracks and trackways.

Paper 4 Dinosaur tracks showing evidence of individual behaviour from the Lower Jurassic Navaho Sandstone, Coyotes Buttes locality, Utah, USA. Jesper Milàn, David B. Loope & Richard G. Bromley Submitted to Acta Palaeontologia Polonica.

Summary The Lower Jurassic Navajo Sandstone of northern Arizona and southern Utah contain a diverse reptilian ichnofauna, dominated by tracks and trackways of small theropod dinosaurs. The theropod tracks are particularly abundant along a single horizon within the aeolian cross strata. Among the tracks are two unique examples of individual dinosaur behavior expressed in the fossil tracks. During reexamination of the trackway of a quadropedal trackmaker, previously thought to be

10 Jesper Milàn – Vertebrate Ichnology Summary of papers ______made by a synapsid reptile, it became clear that the trackway was that of a sauropodomorph (prosauropod) dinosaur. The prosauropod affinity was proven by the presence of a large inward directed pollex claw. The trackway is similar to the trackway Navahopus falcipollex from Arizona, and based on morphological differences in the pes, the new ichnospecies N. coyotensis is erected for the trackway. The trackway is located on the sloping front of a fossil dune and, in the preserved part of the trackway, the arrangement of the tracks show that the animal walked sideways at an angle up the slope for the first part, and then changed direction and continued directly up the slope. The parts where the animal was walking sideways had preserved traces of the digits sliding down slope in the steps. A trackway segment from a small theropod is unique in that it show tracks from the theropod walking and then lying down in resting posture and then rising up again and walking away. The resting trace comprises two impressions of the elongated metatarsi, impressions of the small hands, an impression of the ischial callosity and a curved impression of the animal’s tail. This is only the fourth example worldwide of a theropod crouching down, and the first in the World to include both tail impression and tracks leading to and away from the crouching site. The trackway was orientated up the slope of a dune, and the theropod was crouching down near the top of the dune before it continued uphill. The behavior captured in the prosauropod and the theropod trackway gives a unique view into the lives of dinosaurs in the Early Jurassic deserts.

Paper 5 Preservation and erosion of theropod tracks in eolian deposits; examples from the Middle Jurassic Entrada Sandstone, Utah, USA. Jesper Milàn & David B. Loope The Journal of Geology 2007, v. 115, p. 375–386.

Summary The Middle Jurassic Entrada Sandstone exposed near the town of Escalante, southern Utah contains numerous dinosaur tracks and trackways, predominatly from medium to large-sized theropods. The tridactyl theropod tracks range in length from 4 to 45 cm and some of the tracks and trackways show evidence of a semi-plantigrade walk where the elongated metatarsus has been partly impressed into the substrate during the walk. In addition to the theropod tracks are two

11 Jesper Milàn – Vertebrate Ichnology Summary of papers ______sauropod trackways, unique in that they are the oldest evidence of sauropod dinosaurs in North America, and that they have preserved a sinuous drag trace from the tail. The exposure of the Entrada Sandstone consists of large-scale cross-bedded aeolian deposits that are interbedded with horizontally laminated sand sheets and thin sets of aeolian cross- strata, representing periods with a moister climate. The majority of the tracks is found in the horizontal sandsheets and are especially abundant in the uppermost members of the Entrada Sandstone at the Entrada – Summerville transition zone in Utah and can be traced for about 1000 square kilometres as part of the “Moab Megatracksite”. The studied tracks show that the impact of the dinosaurs’ feet onto the thin-bedded sand deposits generated a zone of disturbance up to 20 cm wide radiating outward and downward from the tracks. Present day erosion has exposed the tracks in a wide range of preservational states, ranging from well-preserved with the sedimentary infilling of the true tracks still in place, to deeply eroded, with the only evidence of the animal’s passage being the circular structures created by erosion through the deep undertracks. This allows detailed studies of track and undertrack formation in aeolian sediments. Tracks that originally were emplaced on sloping surfaces show, in their present day erosional state, a morphology distinct from those originally emplaced on horizontal surfaces. In theses case the main deformation created by the foot is directed downslope. The wide spectrum of morphological variation of the tracks and undertracks due to the present day erosion, demonstrate that great care should be taken when describing fossil footprints that have been exposed to subaerial erosion. The shape, dimensions and general appearance of the tracks are seriously altered by erosion. Such erosion could take place on the modern outcrop, as in the present case, or in the ancient setting, before final burial. When studying aeolian sediments, special attention should be paid to horizons with deformation structures, as the present study demonstrates that what might look like random deformation structures in the sand layers is very likely to be the eroded remains of vertebrate tracks.

12 Jesper Milàn – Vertebrate Ichnology Summary of papers ______

Paper 6 Dinosaur footprints from the Middle Jurassic Bagå Formation, Bornholm, Denmark. Jesper Milàn & Richard G. Bromley Bulletin of the Geological Society of Denmark, 2005, v. 52, p. 7–15.

Summary The first dinosaur tracks from Denmark are described from sandstone blocks discarded from the now flooded clay pit Pyritsøen, south of Hasle, on the Baltic island Bornholm. The sandstone layer where the tracks were found belongs to the Middle Jurassic Bagå Formation, but the exact stratigraphic horizon is not known as the blocks were found loose and the pit is now flooded with water. To ensure the origination of the sandstone blocks, palynological samples taken from the blocks confirmed the Middle Jurassic age of the blocks and that they did originate from the Bagå Formation. The surface of several of the discarded sandstone blocks had a distinct highly sculpted surface, suggesting trampling from large vertebrates. Among the studied blocks were two with clearly recognizable dinosaur tracks. The tracks were preserved as natural casts, and were formed by dinosaurs trampling in the soft clay of the floodplain. The clay had a soft consistency, allowing the feet of the dinosaurs to be emplaced to a depth of 20–30 cm. Subsequent flooding of the floodplain filled the tracks with sand and deposited a sand layer that now constitutes the sandstone bed. One block contained two large oval tracks, measuring respectively 69 and 68 cm in length and 45 and 48 cm in width. The tracks were tetradactyl with impressions of short blunt toes. When the foot was emplaced in the soft mud, the toes formed deep ridges in the front of the tracks as they sank through the mud. The tracks are identified as tracks from sauropod dinosaurs. A third, smaller track measuring 26 cm in length and 19.5 cm in width has the impressions of five short forward facing digits and is suggested to be from an early armoured dinosaur. The find of dinosaur tracks along a clay-sand interface in the Bagå Formation, opens the possibility for more finds, and special attention should be paid to horizons with deformation structures in the other Mesozoic terrestrial deposits on Bornholm.

13 Jesper Milàn – Vertebrate Ichnology Summary of papers ______

Paper 7 Ichnological evidence for giant ornithopod dinosaurs in the Late Jurassic Lourinhã Formation, Portugal. Octávio Mateus & Jesper Milàn Submitted to Oryctos.

Summary The Late Jurassic Lourinhã Formation is situated in the Lusitanian Basin, central west Portugal, and is prominently exposed along the coast near the town of Lourinhã. The Lourinhã Formation consists of alternating beds of red and green clays, interbedded with massive, laterally extensive sandstone beds, representing periods of increased channel activity. Fossils of dinosaurs are particularily abundant in the formation, and skeletal remains of a diverse fauna comprising theropods, sauropods, stegosaurs, ankylosaurs and several genera of ornithopods have so far been described from the area. The sedimentology in the area, with its rapid lithological shifts, favours preservation of tracks and trackways, and so far tracks from most of the dinosaurs represented by skeletal remains have been found in the area. During fieldwork carried out in the summer of 2003 a large sized tridactyl track was found approximately 6 kilometres north of Lourinhã, at the beach of Vale Frades. The track was situated on a clay bed within the tidal zone, and due to the immediate danger that the track would be destroyed by erosion, the track was excavated by Museu da Lourinhã. The preservation of the track is unusual in that it is preserved as sandstone cast standing on a pedestal of clay. This is the result of the original track being infilled with sand which became lithified, while the clay remained unconsolidated. The sandstone cast of the track protects the subjacent clay layers from the present day wave erosion. The track is 70 cm long and 69 cm wide, and the toes are short and broad, with indications of short blunt claws. The angle of divarication between the outer digits is high, and the dimensions of the track identified the track as being made by an ornithopod dinosaur with an estimated hip height around three meters. Large sized theropod tracks from the Lourinhã Formation are morphologically distinctive, in that they have long slender digits, with indications of sharp claws and a low angle of divarication between the outer digits.

14 Jesper Milàn – Vertebrate Ichnology Summary of papers ______

Large sized ornithopods are mostly known from the Early Cretaceous and onwards, while no really large sized ornithopods so far have been found in the Late Jurassic. The largest ornithopod known from the Jurassic is Camptosaurus from North America, and the largest known from Portugal and Europe is the camptosaurid Draconyx loureiroi. Neither of these reached the body size suggested by the new track. The described track is so far the only evidence for a Late Jurassic ornithopod of that size.

Paper 8 Tracking polar dinosaurs - new finds from the Lower Cretaceous of Svalbard. Jørn H. Hurum, Jesper Milàn, Øyvind Hammer, Ivar Midtkandal, Hans Amundsen & Bjørn Sæther Norwegian Journal of Geology 2006, v. 83(4), p. 397–402.

Summary Several tracks from ornithopod dinosaurs are described from the Early Cretaceous (Barremian) section at Festningen, Isfjorden, Svalbard, which was located within the polar latitudes at about 65° N during the Lower Cretaceous. The sediments at Festningen consist of sandstones interbedded with alluvial floodplain sediments and occasional coal seams. The tracks were found preserved as natural casts on the wall of a vertical crevasse left by a coal seam that had been washed away by the sea, and special homemade laser scanning equipment was necessary in order to scan and map the trackbearing surface. The Dinosaur tracks from the Festningen are found in three distinct modes of preservation. 1. Preserved as concave low-relief impressions in sandstone, most likely undertracks due to their shallow and undetailed appearance. 2. Natural sandstone casts in the top of the eroded coal seam, and 3. Preserved as isolated sandstone casts within organic rich shales. The new finds of tracks from ornithopod dinosaurs at Svalbard show that the previous finds of dinosaur tracks from Svalbard was not an isolated phenomenon. Instead it shows that both theropod and ornithopod dinosaurs thrived there for long intervals of time as evidenced by the findings of their tracks along several stractigraphic levels from the Lower Cretaceous. During the beginning of the Early Cretaceous a “greater Wealden” dinosaur fauna were present throughout Laurasia. This dinosaur fauna consisted of iguanodontids, basal euornithopods, nodosaurid ankylosaurians, brachiosaurids and various theropods. Together with finds of dinosaurs from other

15 Jesper Milàn – Vertebrate Ichnology Summary of papers ______areas that were located within the polar latitudes during the Mesozoic, the tracks from Svalbard represent important biogeographical information about the distribution and composition of the “greater Wealden” polar dinosaur faunas. Furthermore, the tracks are direct evidence that dinosaurs were thriving within the polar latitudes.

Paper 9 A diverse vertebrate ichnofauna from a Quaternary eolian oolite from Rhodes, Greece. Jesper Milàn, Richard G. Bromley, Jurgen Titschack & Georgios Theodorou Accepted for publication in: Bromley, R.G., Buatois, L.A., Márango, M.G., Genise, J.F. & Melchor, R.N. (eds.), Sediment-organism interactions: A multifaceted ichnology. SEPM Special Publications 2007, v. 88.

Summary A diverse mammalian ichnofauna is described from a newly discovered outcrop of aeolian, oolithic sediments in the southwestern end of the Greek island of Rhodes. The main exposure of the aeolianite is exposed in a road cutting, and an additional exposure in a nearby quarry. The ooids have sand-sized nuclei and are lightly cemented by vadose meniscus cement. The dune deposits have ramp morphology and within the dunes are three palaeosols that divide the aeolian sediments into three units. The palaeosols contain rhizoliths, gastropod shells and poorly preserved invertebrate bioturbation. The three units are considered to have been deposited during glacial periods where sea level low-stands exposed the oolitic seafloor sediment, which became blown inland and deposited. The soil horizons formed during the warmer interglacial periods. This dates the eolinite to Late Pleistocene to Early Holocene age. The aeolianite contains ichnological evidence of a diverse vertebrate fauna, comprising footprints ranging from 4 to 37 cm in diametric cross section. A total of 79 tracks are described in cross section and additional 4 tracks were found and described on the limited exposures of horizontal bedding planes. The tracks are poorly preserved due to the oolitic nature of the sediment, and only the overall dimensions of the tracks are measurable. The tracks fall into five distinct size-classes where the three smallest sizes of tracks probably are made by different smaller types of artiodactyls. The next largest size class of tracks is

16 Jesper Milàn – Vertebrate Ichnology Summary of papers ______represented both by cross sections and well-preserved tracks on a horizontal surface. The morphology of these tracks is consistent with the morphology of camel tracks, and comparisons with camel tracks from living camels confirm the cameloid origin of the track. This is the first evidence of Pleistocene or Early Holocene camels on Rhodes. The largest of the tracks are interpreted to have been made by proboscideans, especially the medium sized endemic Pleistocene elephants that inhabited the island. The combination and diversity of size groups of tracks differs in the three units, demonstrating differences in faunal composition during the different depositional periods of the oolite formation. Unit 3 has the highest diversity of tracks sizes, which suggests that the animal diversity was highest during that period.

Paper 10 A Late Holocene tracksite in the Lodbjerg dune system, northwest Jylland, Denmark. Jesper Milàn, Lars B. Clemmensen, Bjørn Buchardt & Nanna Noe-Nygaard Manuscript in preparation for New Mexico Museum of Natural History and Science Bulletin.

Summary A diverse mammalian trackfauna is found on the exposed surface of lake/bog deposits, dated to 1265 ± 145 cal. yr. BC. The track-bearing lake/bog deposit is found at the Holocene Lodbjerg dune system at beach level at the northwest coast of Jylland, Denmark. The Lodbjerg dune system is bounded seaward by a prominent coastal cliff with partly fixed cliff-top dunes. The trackbearing lake/bog deposit rests directly on Weichselian till, and is covered by the aeolian Lodbjerg dune system. The Lodbjerg dune system is composed of alternating aeolian sand units and peaty palaeosols recording alternating periods of aeolian activity and stabilization of the dunefield. The trackbearing surface is part of a system of locally developed lake basins developed in the structural lows of the underlying Weichlian till, which subsequently became buried by aeolian sands. The track-bearing surface contains tracks from at least six different types of animals, with tracks of different artiodactyls being the most abundant, represented by tracks from cattle, sheep or goat, and red deer. Among the artiodactyl tracks are also preserved tracks from unshod horse and domesticated dog or wolf. The tracks are all found in the peaty surface of the Lake/bog deposit and are well-preserved, except where the surface has been modified by waves from the sea.

17 Jesper Milàn – Vertebrate Ichnology Summary of papers ______

All the studied tracks were infilled with the overlying aeolian sand, suggesting that the track assemblage was rapidly buried during the onset of aeolian movements, and thus represent a single event. In addition to the tracks from the surface of the lake/bog deposit, there are also several tracks to be found preserved in cross section within the overlying aeolian sands, demonstrating that animals continued to inhabit the area during the periods of aeolian activity. The studied locality consists of two exposed surfaces, presumably parts of the same lake basin, situated approximately 600 metres apart. The main locality, Site 1, contains tracks of cattle, sheep or goats, red deer, horse and canids and the smaller exposure, Site 2, contains tracks of sheep or goats and red deer. The faunal assemblage indicated by the tracks, is consistent with skeletal material excavated from similar aged settlements in the nearby area, and most of the way along the west coast of Jylland down to Holland. A part of site 1 was mapped in detail, and the directions of 42 tracks from the mapped are was determined. This showed the main direction of progression to be perpendicular to the slope of the lake/bog deposit, i.e. shore-parallel. The study of the trackfauna from Lodbjerg, shows that ichnological methods can provide important additional data about faunal compositions, palaeoecology and sedimentary conditions at the time of trackmaking.

Paper 11 Rheotactic Macaronichnus, and human and cattle trackways in Holocene beachrock, Greece: reconstruction of palaeoshoreline orientation. Richard G. Bromley, Alfred Uchman, Jesper Milàn & Klaus Steen Hansen Accepted for publication in Ichnos.

Summary The Roman aged beachrocks found approximately 2 kilometres west of Pefkos on the Greek Island of Rhodes, contain both invertebrate and vertebrate trace fossils. This study describes the different tracefossils encountered in the beachrock and uses their orientation to reconstruct palaeoshoreline orientation. The polychaete trace fossil Macaronichnus segregatis occurs in three different morphs introduced as new ichnosubspecies, M. segregatis lineiformis, M. segregatis maeandriformis and M. segregatis spiriformis, respectively. M. segregatis lineiformis is a shore-slope parallel, almost

18 Jesper Milàn – Vertebrate Ichnology Summary of papers ______straight tracefossil; M. segregatis spiriformis forms a spiral-like pattern and M. segregatis maeandriformis is an intermediate form forming winding patterns. Along some laminae in the beachrock, M. segregatis is strongly oriented perpendicularly to the Roman-age seashore. In other laminae, the trace fossil displays a winding to spiral-shaped course. It is suggested that the trace fossils were produced under conditions of different porewater flow. During higher energy periods in the winter, water movement was perpendicular to the coastline. This caused M. segregatis to be oriented parallel to the coastline. During periods of more or less stationary pore-water, predominantly in the summer period, M. segregatis assumed a spiral like form, and the winding forms represents intermediate conditions. In the same beachrock as the polycheate traces, are abundant tracks and trackways of artiodactyls and humans. The tracks are all preserved in various states of erosion and some as undertracks. The artiodactyl tracks are interpreted as tracks made by pigs or cattle, due to their size and shape. One single human track show a morphology indicating that the trackmaker was wearing shoes, and a short trackway consisting of three consecutive tracks was made by a small or juvenile person estimated to be around 130 cm in height. All the tracks and trackways are orientated eastward and westward parallel to the palaeoshore, the same orientation observed in M. segregatis lineiformis at the locality. The morphology of many of the tracks showed evidence of being modified by the gravitational effect of the sloping beach deposits. The beachrock locality was studied in 2001 and revisited again in 2005 for further studies. In that period 2/3 of the original described surface was eroded away, and the rest was strongly modified by erosion. However, the erosion exposed new features of the existing trackways and exposed new trackbearing layers subjacent to the original surface. That all the tracks and trackways were orientated both eastward and westward along the palaeoshoreline plus the fact that tracks were found on several subjacent bedding planes suggests that the trampleground represents multiple use of a coast-parallel beach pathway. The beachrock at Pefkos is thus a unique example, where both invertebrate and vertebrate tracefossils are used to determine the palaeoshore orientation.

19 Jesper Milàn – Vertebrate Ichnology Summary of papers ______

Paper 12 Do shod humans leave true tracks? Jesper Milàn & Richard G. Bromley Accepted for publication in Ichnos

Summary The recent interest in hominid ichnology raises some interesting ichnophilosophical questions. One of the main problems is the incorporation of footprints from humans wearing shoes into ichnotaxonomical nomenclature. In the strictest sense, a true track is only formed by the sediment grains that come directly in contact with the trackmakers foot. The problem of wearing of shoes is that the trackmaker does not actually touch the sediment it walks upon, and then strictly speaking is not producing a true track but an undertrack, which should not form basis for a valid ichnotaxonomical name. Furthermore, the wearing of shoes represents as much a behavioural choice as well as it is evidence of technological developments in shoemaking and the whims of fashion. The sheer variety of different morphologies human footwear exhibits like different sole patterns and shape and proportions of the different parts of the shoe, could be interpreted as widely different ichnospecies if treated by standard ichnological methods. If tracks of shod humans were to be treated ichnotaxonomically, they should then be regarded as a compound trace of behaviour that comprises bipedal walking, the choice of wearing, and the technological development of shoe production. For these reasons, tracks of shod humans should not be the subject of ichnotaxonomic nomenclature, while they by all means still should be registered and studied when encountered. Tracks of naked feet however, offer no such restrictions.

20 Jesper Milàn – Vertebrate Ichnology Material and methods ______

Material and methods

The majority of the work has been conducted on original material found and collected in the field or described in situ at the localities. The following is a listing of the material and specimens studied in the individual papers. Interpretations about the morphology and taphonomy of the tracks will be based on the methods previously developed by Milàn (2006), and Milàn & Bromley (2006, in press b). The specific methods and techniques used in the individual studies are described in detail in the individual papers. The dinosaur track from the Upper Triassic of southern Sweden (Milàn & Gierlinski 2004 [Paper 1]) is stored in the collection of the Geological Museum, University of Copenhagen (MGUH 27219). The Late Triassic tracks from East Greenland (Milàn et al. 2004 [Paper 2], 2006 [Paper 3]) were collected by Niels Bonde and Lars Clemmensen and are now included in the collection of the Geological Museum, University of Copenhagen (MGUH 278811 – 27815). Some of the tracks were sectioned vertically, perpendicular to the length axis of digit III. The sections were polished and scanned directly on a flatbed scanner in 600dpi. The scans were enhanced using Corel Draw 10-12. The Lower Jurassic tracks (Milàn et al. 2006 [Paper 3]) are from the collections of Museu Tridentino di Marco (MTSN 5210 – 5213) and the Late Jurassic tracks are from the collection of Museo del Jurásico de Asturias (MUJA 1080, 1199 and 1261). All tracks and trackways from the Navaho and Entrada Sandstone in Utah (Milàn et al. submitted [Paper 4], Milàn & Loope 2007 [Paper 5]) were described in situ at the localities. The dinosaur tracks from Bornholm (Milàn & Bromley 2005 [Paper 6]) was collected in the field and are now part of the collection of the Geological Museum, University of Copenhagen (MGUH 27754 & 27755) and placed on permanent display at NaturBornholm. The gigantic ornithopod track from the Late Jurassic of Portugal (Mateus & Milàn, submitted [Paper 7]) was collected and is now part of the collection of Museu da Lourinhã (ML 1000). To confirm the ornithopod identity of the track, the dimensions of the track was compared with the multivariate analysis of theropod and ornithopod tracks proposed by Moratalla et al. (1988). The track material from Svalbard (Hurum et al. 2006 [Paper 8]), was studied in the field and casts of selected tracks are now stored in the Palaeontological Museum in Oslo, (PMO

21 Jesper Milàn – Vertebrate Ichnology Material and methods ______

210.570 & PMO x621). Most of the tracks were located down in a crevice and were scanned, by lowering a custom-built scanner down the crevice (Hurum et al 2006). The trackfauna from the Pleistocene/Holocene oolite from Rhodes (Milàn et al. in press [Paper 9]) was described in the field, and the tracks from recent camels were obtained in the Zoological Gardens, Copenhagen. The trackfauna from the Holocene Lodbjerg tracksite (Milàn et al. in prep. [Paper 10]), was mapped directly in the field on large sheets of transparent plastic, and later redrawn and digitized. 16 selected tracks were cast and will be stored at the Geological Museum, University of Copenhagen. Tracks from the Roman-aged beachrock (Bromley et al. in press [Paper 11]), were recorded and described in situ at the locality at the beach near Pefkos, Rhodes. The holotypes of the three new subspecies of Macaronichnus were collected and are now stored in the Geological Museum, University of Copenhagen (MGUH 27804 – 27806).

22 Jesper Milàn – Vertebrate Ichnology Discussion ______

Discussion

The 12 individual papers produced during this study deal with different aspects of vertebrate ichnology, which can be grouped into three major themes; Tracks as indicators of biodiversity, Tracks as indicators of behaviour and Tracks and sediment interactions. As each paper was conducted as a separate case study, most of the papers deal with more than one of the three topics.

Tracks as indicators of biodiversity A track is direct, in situ evidence of the presence of the trackmakers in contrast to skeletal remains, which can be transported for a considerable distance before burial; the track is therefore an important tool for reconstructing past biodiversity. Tracks help expand the knowledge of palaeofaunas in areas where skeletal remains are scarce or absent, and provide information about size and distribution of the trackmakers. Many of the tracks from the present study are such examples where tracks have implications for the palaeobiodiversity. The Late Triassic Höganäs Formation from southern Sweden, has only preserved fragmentary and indeterminable skeletal remains of dinosaurs (Bölau 1954), whereas tracks and trackways of theropod dinosaurs are abundant (Bölau 1952; Pleijel 1975; Ahlberg & Sivertson 1991; Gierlinski & Ahlberg 1994). The herein described track from an indeterminate ornithischian dinosaur, presumably of thyreophorean origin, is thus the only evidence of herbivorous dinosaurs in an assemblage otherwise dominated by theropod tracks (Milàn & Gierlinski 2004 [Paper 1]). The Middle Jurassic is characterized by a worldwide scarcity of dinosaurian body fossils (Romano & Whyte 2003; Whyte et al. 2007). In Europe the scarcity of body fossils is especially pronounced and ichnological data thus become an important source of information about the dinosaurian diversity and biogeography during the Middle Jurassic. The Middle Jurassic Cleveland Basin of the Yorkshire coast has preserved one of the most diverse track assemblages from the Middle Jurassic of Europe, with abundant tracks from theropods, ornithopods, stegosaurians, sauropods, theropods and crocodilians (Romano & Whyte 2003; Whyte et al. 2006, 2007). In contrast to that, the only skeletal remains from that area are a single, partial vertebra (Romano & Whyte 2003). Middle Jurassic sauropod tracks are known from Portugal (Santos et al. 1994), and sauropod tracks together with tracks from theropods are known from Oxfordshire in England (Day et al. 2004), and isolated theropod tracks are found on the Isle of Skye, Scotland

23 Jesper Milàn – Vertebrate Ichnology Discussion ______

(Clark et al. 2005). The new finds of sauropod and thyreophorean tracks from the Middle Jurassic of Bornholm is from a formation where, to date, no skeletal remains have been discovered, (Milàn & Bromley 2005 [Paper 6]). The Late Jurassic Lourinhã Formation is known for an extensive skeletal record of dinosaurs and numerous tracks and trackways that are attributable to the dinosaur fauna (Antunes & Mateus 2003). However, the newly found gigantic ornithopod track is considerable larger that any hithero known ornithopod from the Late Jurassic, both from skeletal and track material (Mateus & Milàn submitted [Paper 7]). A similar situation is described from the Lower Jurassic Kayenta Formation of Arizona, where theropod tracks are found that are at least 1/3 longer that any theropod known from skeletal material from the Lower Jurassic of North America (Morales & Bulkley 1996). That skeletal remains so far have not been discovered of these “over-sized” trackmakers, is most likely due to the fact that the larger an animal is, the fewer individuals are present in the animal community, and thus likely to leave a fossil behind. This is similar to the situation from the Quaternary aeolianite (Milan et al. in press, [Paper 9]) where, among many tracks and traces attributable to animals well known from skeletal remains, there are tracks attributable to camels, who never have been found as fossils on Rhodes. The tracks from the Lower Cretaceous of Svalbard are especially important, as they are indisputable proof that both theropod and ornithopod dinosaurs inhabited the polar regions during the Mesozoic (Hurum et al. 2006 [Paper 8]). When dinosaur tracks were first recorded from Svalbard in 1960 it was a sensation as it was the first record of dinosaurs living within the polar latitudes (Lapparent 1960, 1962). Ironically, the finds of dinosaur tracks from the polar latitudes did not seem to effect the predominant perception at that time, that dinosaurs were cold blooded. Since then reports of polar dinosaurs have now been known from more than 10 arctic areas from the Late Jurassic to the Late Cretaceous (Rich et al. 1997, 2002; Fiorillo 2004, 2006). The dinosaur tracks from Svalbard are found along several horizons, demonstrating that the presence of dinosaurs was not a singular phenomenon (Hurum et al. 2006 [Paper 8]). Taken altogether, these reports all represents cases where ichnological data help to expand the knowledge of the palaeofauna, either by being the only evidence of the existence of the fauna, or by providing additional data about hitherto unknown faunal elements. In short; animals leave millions of tracks during their lifetime, while they only leave one skeleton behind. Taken the widely different taphonomical factors controlling track and bone preservation, the chances that just a few of the tracks gets preserved instead of the skeleton is high.

24 Jesper Milàn – Vertebrate Ichnology Discussion ______

Tracks as indicators of behaviour Fossil tracks and trackways are more than just the mere impression of the feet of the animals, they are also important sources of information about the behaviour of the animals that made the tracks, information that it is not possible to obtain from the study of skeletal remains. Several examples are known where trackways from dinosaurs are running parallel without crossing each other. Such trackway evidence for gregarious behaviour is known from sauropods, theropods and ornithopods (e.g. Ostrom 1972; Barnes & Lockley 1994; Lockley et al. 1994, 2002; Lockley & Matsukama 1999; Barco et al. 2006). Trackways can further provide data about the specific behaviour and habits of an individual animal. Since Alexander (1976) developed the formula for calculating the progressing speed of animals form the tracks, it has been used regularly to estimate the progression speed of especially dinosaurs (e.g. Farlow 1981; Thulborn 1982; Mazzetta & Blanco 2001). One interesting example is the Middle Jurassic tracksite of the Ardley Quarry, England, where one theropod trackway shows the animal to have gradually increased its speed along the captured trackway (Day et al. 2004), and in an example from Late Triassic of Virginia, USA, a theropod burst from walking speed to full run in the span of three paces (Weems 2006). When the formula for progression speed is applied to the trackway form the subadult human from the Roman-aged beachrock of Rhodes (Bromley et al. in press [Paper 11]), the resulting walking speed was 2.59 km/h, which is a fair and comfortable progression speed when walking in beach sand (pers. obs. 2005). Some trackways further reveal information about the behaviour of the individual trackmaker. Examples could be trackways showing alternating long and short steps from limping dinosaurs (e.g. Dantas et al. 1994; Lockley et al. 1994) or a dinosaur that accelerates to run (Day et al. 2004). The sauropodomorph trackway described from the Navajo Sandstone (Milan et al. submitted [Paper 4]), show the dinosaur to first walk sideways up the sloping front of a dune and then continue straight up the dune face. This is a clear example of the behaviour of a trackmaker walking up the sloping front of a sand dune reflected in a trackway. The trackway from the crouching theropod dinosaur poses an ichnotaxonomical problem as it in reality is a compound trace resulting from several different behaviours, including walking, squatting and rising up again (Lockley et al. 2003; Lockley 2007). Most tetrapod ichnotaxa are diagnosed on the tracks alone, and do not reflect behaviour. An exception is the Permian ichnotaxon Moodeichnus (Sarjeant 1972), which is a behavioural variant of Dromopus and which appears didactyl instead of pentadactyl (Lockley 2007). Giving different ichnotaxonomical names to different parts of a trackway because

25 Jesper Milàn – Vertebrate Ichnology Discussion ______of change in behaviour would result in over splitting of names and confusion. Based on these considerations, we refrain from erecting new ichnotxonomical names for the trace of the resting theropod from the Navajo Sandstone and simply refer to it as the trace of a squatting theropod (Milàn et al. Submitted [Paper 4]). Although we erect the new ichnospecies Navahopus coyotensis for the prosauropod trackway based on direct morphological differences in the imprints of the pes, it would have been possible to erect a name for further behaviour based ichnotaxa for the part of the trackway where the animal was progressing sideways. Even if this were possible, the result would be a highly confusing example where half the trackway would belong to one ichnotaxon and the other half would be a different ichnotaxon. The problems of applying ichnotaxonamical names to traces that incorporate behaviour is put to a test in the discussion about naming of tracks from shod humans, where it is argued that tracks from shod humans in reality represent a compound trace comprising bipedal walking, foot morphology and technological developments in foot wear (Milàn & Bromley in press a [Paper 12]). If such tracks were to be given ichnotaxonomical names, then the possible list of ichnonames for human tracks would be endless. The consistency of the substrate influences the mode of progression adapted by the trackmaker. In soft sediments, where the feet of the animals sink to a considerable depth, the animal adopts a mode of walking where the feet are emplaced and lifted as close to vertically as possible to prevent drag from the foot. This is especially seen in a deep, three-dimensionally preserved, natural cast of a sauropod manus from the Late Jurassic of Portugal, where the dimensions of the cast are constant all the way from top to bottom of the cast, with clear striations from the skin preserved on the sides of the cast (Milàn et al. 2005). In firmer mud where the feet did not sink as deep, the feet were dragged forward through the mud, leaving the tracks as elongated grooves (Day et al. 2004). The sauropod tracks from Bornholm were impressed to a depth of 25 to 29 cm in the best preserved of the specimens (Milàn & Bromley 2005 [Paper 6]). The trackwalls are close to vertical, demonstrating that these animals did lift their feet vertically out of the soft sediment. Tracks from small theropod dinosaurs walking in deep mud, however, have been described as highly elongated collapsed structures, showing that these dinosaurs did indeed drag their feet forward through the mud instead of lifting them vertically out of the mud (Gatesy et al. 1999). Two of the studies presented herein bears evidence of the trackmakers progressing parallel to the direction of the palaeoshoreline. The human and animal tracks described from the fossilized Roman-aged beachrocks of Pefkos, Rhodes are predominantly running parallel to the palaeoshoreline orientation, a fact supported by the orientation of the invertebrate trace fossil

26 Jesper Milàn – Vertebrate Ichnology Discussion ______

Macaronichnus (Bromley et al. in press [Paper 11]). The majority of the animal tracks described from the lake/bog deposit of Lodbjerg, Denmark are progressing perpendicular to the slope of the lake/bog deposits, with only few of the tracks heading towards or away from the basin (Milàn et al in prep. [Paper 10]). Shore parallel progression is both a behavioural and a practical choice for animals. Often the shores of lakes, rivers or seas constitutes natural paths with less vegetation cover, and less relief that the inland settings. Numerous examples of shore parallel progression and tracks and trackways made in shorefacies are known from the fossil record, especially from the shores of the Western Interior Seaway, which existed in North America up through the Cretaceous (Lockley & Hunt 1995).

Tracks and sediment interactions Sediment properties exercise a strong influence on the appearance and morphology of the track. Experimental work with tracks emplaced in sediments of different water contents, have demonstrated that tracks from the same animal alter their appearance dramatically as the water content of the sediment increases (Brand 1996; Milàn 2006; Milàn & Bromley 2006, In press b): Fossil examples of tracks and trackways that change morphology as a result of changes in the original water content of the sediment are described from the Late Triassic of Greenland (Gatesy et al. 1999), the Middle Triassic of Germany (Diedrich 2002) and in sauropod trackways from the Upper Cretaceous of Bolivia (Lockley et al. 2002). The dinosaur tracks described from the Middle Jurassic of Bornholm (Milàn & Bromley 2005 [Paper 6]) were originally emplaced in relatively soft mud that allowed the dinosaurs’ feet to sink to a considerable depth and creating elongated grooves from the toes being dragged up through the sediment. In this case however, the sediment was firm enough to preserve the shape of the track after the foot was lifted, and retain the shape while the tracks were infilled with sand without collapsing, which would have rendered the track unrecognizable. When tracks are emplaced in finely laminated sediments, like the case of the large theropod tracks from the Middle Jurassic Entrada Sandstone (Milàn & Loope 2007 [Paper 5]) or the Late Triassic lake deposits of East Greenland (Milàn et al. 2004 [Paper 2], 2006 [Paper 3]) the weight of the trackmakers foot has formed recognizable deformation in the sediment below and around the direct imprint of the trackmakers foot. Especially the horizontally laminated sand-sheets of the Entrada Sandstone offer the perfect conditions to study the formation of undertracks in finely laminated sediments (Milàn & Loope 2007 [Paper 5]). Here the full extent of the deformation created by the

27 Jesper Milàn – Vertebrate Ichnology Discussion ______trackmakers foot becomes evident, as the present day erosion has exposed zones of deformation radiating as far as 20 cm out from the edges of the true track, and vertical erosional cuts through tracks show the deformation to extend for at least 10 cm below the true track as undertracks. The preservation of the tracks from the Entrada Sandstone (Milàn & Loope 2007 [Paper 5]) differs from that of the tracks from the Navajo Sandstone (Milan et al. submitted [Paper 4]) in that there is no apparent undertrack formation connected to the tracks described from the Navajo Sandstone, although both localities are aeolian sand deposits. One explanation for this could be that the tracks from the Navajo Sandstone in most instances are emplaced in the looser, more structureless sands of the sloping fronts of the dunes and the tracks from the Entrada Sandstone are found in horizontally laminated interdune deposits. Furthermore, the tracks described from the Navajo Sandstone are from considerably smaller animals with typical foot lengths of 5 to 15 cm in contrast to the large animals from the Entrada Sandstone with typical footlengths of around 40 cm. Such larger animals would exercise a considerably larger pressure on the substrate, displace more sediment and create deeper undertracks. Compared with the tracks from the Late Pleistocene to Holocene aeolianite from Rhodes (Milàn et al. in press, [Paper 9]), the tracks from Rhodes are much more indistinctly preserved, due further to the fact that the majority are only exposed in cross section. However, the tracks from Rhodes are all relatively deep compared to their size, suggesting that the oolithic sand they are emplaced in was relatively loose at the time of tracking, but firm enough to preserve the sub-vertical trackwalls. The preservation of the tracks from Rhodes is more similar to the tracks and trackways described from Pleistocene coastal aeolinties from Mallorca, where the majority of the tracks are preserved as indistinct holes in the sediment surface and associated with the formation of displacement discs around the tracks (Fornós et al. 2002). When the Mallorcan tracks are exposed in cross section, they also show that the trackmakers feet had sunk to a considerable depth into the sand. The thyreophorean track from the Upper Triassic of Skåne (Milàn & Gierlinski 2004 [Paper 1]) and a number of the tracks found at Svalbard (Hurum et al. 2006 [Paper 8]) are preserved as natural casts of sand/siltstone protruding down into a coal seam. This is a commonly occurring form of preservation and, through time, numerous tracks and trackways have been reported, especially from the roofs of coalmines, where the casts of the tracks often have been left after the coal was quarried out (e.g. Peterson 1924; Lockley & Jennings 1987; Parker & Balsby 1989; Parker & Rowley 1989). The high degree of preservation of tracks from coal mines is due to several

28 Jesper Milàn – Vertebrate Ichnology Discussion ______factors. The tracks were originally emplaced in soft peat, which had good moulding properties and was firm enough to sustain the shape of the foot after the foot was lifted again. In all cases the peat horizon was rapidly buried by fluvial deposits (sand or mud) infilling the tracks. After diagenesis of the sediments, the coal and sand/mudstone facies readily splits along the lithological border, exposing the tracks. The Holocene track assemblage from Lodbjerg (Milàn et al. in prep [Paper 10]), are preserved on the surface of a peaty lake/bog deposit and have been rapidly covered by aeolian sand. This setting, if allowed to fossilize under the right circumstances, would be preserved as a thin coal seam, with natural casts of tracks protruding down from the overlying sandstone beds. The Lodbjerg tracksite is thus a good modern example to demonstrate the origin of tracks preserved in the roofs of coal seams.

29 Jesper Milàn – Vertebrate Ichnology Conclusions ______

Conclusions

Each of the 12 papers conducted as part of this study represents different aspects and applications of vertebrate tracks and trackways, concerning tracks as indicators of biodiversity, tracks as records of behaviour and track formation and preservation. The main conclusions to be drawn from the present study are: Tracks are important tools in reconstructing prehistoric biodiversity, as they in many cases are the only evidence that the animals inhabited the area, this includes: The first record of early thyreophorean dinosaurs are described from the Upper Triassic of southern Sweden, and the first dinosaur tracks from Denmark are found and described from the Middle Jurassic Bagå Formation of Bornholm. The tracks derive from a large sauropod with a foot length of 70 cm and from a smaller armoured dinosaur with a foot length of 26 cm. These tracks are important data about the distribution and diversity of dinosaurs in the Middle Jurassic, where the skeletal records of dinosaurs are scarce. A large track from an ornithopod dinosaur, from the Late Jurassic of Portugal, is evidence of a hitherto unknown ornithopod dinosaur with a body size considerably larger than what is known from contemporary skeletal remains. Dinosaur tracks from the Lower Cretaceous of Svalbard provide important data about the biogeographical distribution of dinosaurs. The tracks are indisputable proof that both ornithopod and theropod dinosaurs inhabited the Polar Regions during the Cretaceous. The late Pleistocene to Holocene Kattavia oolite from southern Rhodes represents three different periods of aeolian activity with abundant tracks from animals. The tracks range in size from 4 to 37 cm in cross section, the latter being made by proboscideans. A well-preserved track from a camel is the first evidence of pre-historic camels on the island of Rhodes. A diverse mammalian trackfauna is described from the Late Holocene (early Bronze Age), eolian system of north-western Denmark. The trackfauna comprises tracks of both domesticated species as cattle, sheep or goat, horse and dog/wolf and tracks from wild animals as red deer. The tracks are all preserved on the surface of a lake/bog deposit covered in aeolian sands, which has preserved the tracks. Fossil tracks and traces of animal provide important data about animal behaviour, knowledge which it is impossible to obtain from the study of body fossils. Dinosaur foot movements and walking dynamics are reconstructed by study of the deformation structures in the sediment around and below the tracks, and a comparative study of theropod tracks from Late Triassic, Early Jurassic and Late Jurassic, show that different theropods at different times adopted

30 Jesper Milàn – Vertebrate Ichnology Conclusions ______different walking strategies. Especially the method of studying vertical sections cut through tracks has proven useful for this purpose. Two trackways are described from the Lower Jurassic Navajo Sandstone which shows unique evidence of individual behaviour of the trackmakers; one show a sauropodomorph dinosaur walking sideways up the sloping front of a dune before it changes orientation and continues head-on up. The other shows a theropod that crouched down on the sloping front of a dune, leaving impressions of the elongated metatarsi, hands, ischial callosity and the tail, before it stood up again and continued to walk up the dune front. A diverse mammalian trackfauna is described from the Late Holocene (early Bronze Age), eolian system of north-western Denmark. The trackfauna comprises tracks of both domesticated species as cattle, sheep or goat, horse and dog/wolf and tracks from wild animals as red deer. The tracks are all preserved on the surface of a lake/bog deposit covered in aeolian sands, which has preserved the tracks. Lithified Roman-aged beachrock from Pefkos, Rhodes contain tracefossils from polycheate worms, Macaronichnus, and human and livestock together, along several subjacent horizons. The human and animal tracks are predominantly progressing parallel to the shoreline, as inferred from the orientation of the Macaronichnus. The Late Holocene trackfauna from Lodbjerg, show a similar trend in shore-parallel progression, as the majority of the tracks are found orientated perpendicular to the slope of the lake/bog basin. Tracks from shod humans is a composite trace fossil comprising both a behavioural choise of wearing shoes as well as technological developments in shoe wear, and should not form the basis for new ichnotaxa. The sedimentological conditions at the time of tracking and the subsequent taphonomical processes have a strong influence on the morphology of the track. The theropod tracks from the Middle Jurassic Entrada Sandstone were originally emplaced in thin bedded sand- sheet deposits and an extensive set of undertracks were formed below them. Today the tracks are exposed in a wide range of degrees of erosion and erosional cuts through tarcks and undertracks. These range from easily recognizable true tracks to broad zones of concentric rings of distortion representing deeply eroded undertracks. All in all these case studies demonstrate that vertebrate ichnology is an important tool to incorporate in palaeoecological reconstructions. Furthermore, the similarities between the presented cases spanning in time from Late Triassic to Late Holocene, demonstrate the independence of time in ichnological applications.

31 Jesper Milàn – Vertebrate Ichnology References ______

References

Ahlberg, A. & Siverson, M. 1991. Lower Jurassic dinosaur footprints in Helsingborg, southern Sweden. Geologiska Föreningens i Stockholms Förhandlinger 113: 339–340. Alexander, R. McN. 1976. Estimates of speeds of dinosaurs. Nature 261: 129–130. Allen, J. R. L. 1989. Fossil vertebrate tracks and indenter mechanics. Journal of the Geological Society, London 146: 600–602. Allen, J.R.L. 1997. Subfossil mammalian tracks (Flandrian) in the Severn Estuary, S.W. Britain: mechanics of formation, preservation and distribution. Philosophical Transactions of the Royal Society of London, B 352: 481–518. Antunes, M.T. & Mateus, O. 2003. Dinosaurs of Portugal. Comptes Rendus Palevol 2: 77–95. Baird, D. 1957. Triassic Reptile Footprint Faunas from Milford, New Jersey. Bulletin of the Museum of Comparative Zoology 117: 449–520. Barco, J.L., Canudo, J.I. & Ruiz-Omeñaca, J.I. New data on Therangospodus oncalensis from the Barriasian Fuentesalvo Tracksite (Villar del Rio, Soria, Spain): an example of gregarious behaviour in theropod dinosaurs. Ichnos 13: 237–248. Barnes, F.A. & Lockley, M.G. 1994. Trackway evidence for social sauropods from the Morrison Formation, Eastern Utah (USA). Gaia 10: 37–41. Brand, L.R. 1996. Variations in salamander trackways resulting from substrate differences. Journal of Paleontology 70: 1004–1010. Bromley, R.G., Uchman, A., Milàn, J. & Hansen, K.S. (In press). Rheotactic Macaronichnus, and human and cattle trackways in Holocene beachrock, Greece: reconstruction of palaeoshoreline orientation. Ichnos. Bölau, E. 1952. Neue Fossilfunde aus dem Rhät Schones und ihre paläogeographisch-ökologische Auswartung. Geologiska Föreningens i Stockholms Forhandlingar 74: 44–50. Bölau, E. 1954. The first finds of dinosaurian skeletal remains in the Rhaetic-Liassic of N.W. Scania. Geologiska Föreningens i Stockholms Forhandlingar 76: 501–502. Clark, N.D.L., Ross, D.A. & Booth, P. 2005. Dinosaur tracks from the Kilmalaug Formation (Bathonian, Middle Jurassic) of Score Bay, Isle of Skye, Scotland, UK. Ichnos 12: 93–104. Dantas, P., Santos, V.F., Lockley, M.G. & Meyer, C. 1994. Footprint evidence for limping dinosaurs from the upper Jurassic of Portugal. Gaia 10: 43–48.

32 Jesper Milàn – Vertebrate Ichnology References ______

Day, J.J., Norman, D.B., Gale, A.S., Upchurch, P. & Powell, H.P. 2004. A Middle Jurassic dinosaur Trackway site from Oxfordshire, UK. Palaeontology 47: 319–348. Diedrich, C. 2002. Vertebrate track bed stratigraphy at new megatrack sites in the Upper Wellenkalk Member and orbicularis Member (Muschelkalk, Middle Triassic) in carbonate tidal flat environments of the western Germanic Basin. Palaeogeography, Palaeoclimatology, Palaeoecology 183: 185–208. Farlow, J.O. 1981. Estimates of dinosaur speeds from a new trackway site in Texas. Nature 294: 747–748. Fiorillo, A.R. 2004. The dinosaurs of Arctic Alaska. Scientific American 291, 85–91. Fiorillo, A.R. 2006. Review of the dinosaur record of Alaska. With comments regarding Korean dinosaurs as comparable high-latitude fossil faunas. Journal of the Paleontological Society of Korea 22: 15–27. Fornós, J.J., Bromley, R.G., Clemmensen, L.B. & Rodriguez-Perea, A. 2002. Tracks and trackways of Myotragus balearicus Bate (Artiodactyla, Caprinae) in Pleistocene aeolianites from Mallorca (Balearic Islands, Western Mediterranean). Palaeogeography, Palaeoclimatology, Palaeoecology 180: 277–313. Gatesy, S.M., Middleton, K.M., Jenkins, F.A., & Shubin, N.H. 1999. Three-dimensionel preservation of foot movements in Triassic theropod dinosaurs. Nature 399: 141–144. Gierlinski, G. & Ahlberg, A. 1994. Late Triassic and Early Jurassic dinosaur footprints in the Höganäs Formation of southern Sweden. Ichnos 3: 99–105. Gillette, D.G. & Lockley, M.G. (eds.) 1989. Dinosaur tracks and traces. Cambridge University Press. 454 pp.

Haubold, H. 1971. Ichnia Amphibiorum et Reptiliorum fossilium. Handbuch der Paläeoherpetologie Teil 18. 124 pp. Henderson, D.M. 2006. Simulated weathering of dinosaur tracks and the implications for their characterization. Canadian Journal of Earth Sciences 43: 691–704. Hitchcock, E. 1836. Ornithichnology,-Description of the Foot marks of Birds, (Ornithichnites) on new Red Sandstone in Massachusetts. American Journal of Science 29: 307–340. Hitchcock, E. 1848. An attempt to discriminate and describe the animals that made the fossil footmarks of the United States, and especially of New England. American Academy of Arts & Sciences Memoir (n.s.) 3: 129–256.

33 Jesper Milàn – Vertebrate Ichnology References ______

Hitchcock, E. 1858. Ichnology of New England, A report on the Sandstone of the Connecticut Valley Especially its Fossil Footmarks. Boston: William White. 220 pp + plate I-LX. Hitchcock, E. 1865. Supplement to Ichnology of New England. Commonwealth of Massachussetts, 96 pp + plate I-XX. Hurum, J.H., Milàn, J., Hammer, Ø., Midtkandal, I., Amundsen, H. & Sæther, B. 2006. Tracking polar dinosaurs - new finds from the Lower Cretaceous of Svalbard. Norwegian Journal of Geology 83: 397–402. Kim, J.Y., Kim, K.-S., Park, S.I. & Shin, M.-K. (eds.). 2004. Proceedings of International Symposium on the Quaternary Footprints of Hominids and Other Vertebrates. Namjejugun, 175 pp. Lapparent, A.F. 1960. Decouverte de traces de pas de dinosauriens dans le Crétacé de Spitzberg. Comptes Rendus de l’Academie des Sciences 251: 1399–1400. Lapparent, A.F. 1962. Footprints of dinosaur in the Lower Cretaceous of Vestspitsbergen – Svalbard. Norsk Polarinstitutt, Årbok 1960: 14–21. Leonardi, G. (ed.). 1987. Glossary and manual of tetrapod footprint palaeoichnology. Dept. Nacional de Producao Mineral, Bracil. 75 pp. Leonardi, G. 1994. Annotated Atlas of South America tetrapod footprints (Devonian to Holocene). Ministry of Mines and Energy, Brasilia, Brazil, 274 pp. Lockley, M.G. 1987. Dinosaur tracks symposium signals a renaissance in vertebrate ichnology. Palaeobiology 13: 246–252. Lockley, M. 1991. Tracking Dinosaurs. Cambridge University Press. 238 pp. Lockley, M. 2007. A tale of two ichnologies: different goals and potentials of invertebrate and vertebrate (tetrapod) ichnotaxonomy and how they relate to ichnofacies analysis. Ichnos 14: 39– 57. Lockley, M. & Hunt, A.P. 1995. Dinosaur tracks and other fossil footprint of the Western United States. Columbia University Press. 338 pp. Lockley, M.G., Hunt, A.P., Moratella, J.J., & Matsukawa, M. 1994. Limping Dinosaurs? Trackway evidence for abnormal gaits. Ichnos 3: 193–202. Lockley, M.G. & Jennings, C. 1987. Dinosaur Tracksites of Western Colorado and Northern Utah, Late Cretaceous coal mine tracks. In Averett, W. R. (ed.) Paleontology and Geology of the Dinosaur Triangle. The Museum of Western Colorado.

34 Jesper Milàn – Vertebrate Ichnology References ______

Lockley, M.G. & Matsukawa, M. 1999. Some observations on trackway evidence for gregarious behaviour among small bipedal dinosaurs. Palaeogeography, Palaeoclimatology, Palaeoecology 150: 25–31. Lockley, M., Matsukawa, M. & Jianjun, L. 2003. Crouching theropods in taxonomic jungles: ichnological and ichnotaxonomical investigations of footprints with metatarsal and ischial impressions. Ichnos 10: 169–177. Lockley, M. & Meyer, C. 2000. Dinosaur tracks and other fossil footprints of Europe. Columbia University Press. 323 pp. Lockley, M.G., Schulp, A.S., Meyer, C.A., Leonardi, G., Mamani, D.K. 2002. Titanosaurid trackways from the Upper Cretaceous of Bolivia: evidence for large manus, wide-gauge locomotion and gregarious behaviour. Cretaceous Research 23: 383–400. Lull, R. S. 1904. Fossil footprints of the Jura-Trias of North America. Memoirs of the Boston Society of Natural History 5(11): 461-558. Lull, R.S. 1915. Triassic life of the Connecticut Valley. Connecticut State Geological and Natural History Survey, Bulletin 24: 1-285. Lull, R.S. 1953. Triassic Life of the Connecticut Valley. State Geological and Natural History Survey, Bulletin 81: 1- 336. MacDonald, J. 1994. Earths first steps: tracking life before the dinosaurs. Johnson Printing, Colorado. 290 pp. Manning, P. 2004. A new approach to the analysis and interpretation of tracks: examples from the dinosaurian. In McIlroy, D. (ed.). The application of ichnology to palaeoenviromental and stratigraphic analysis. Geological Society, London, Special Publications 228: 93–123. Mateus, O. & Milàn, J. (Submitted). Ichnological evidence for giant ornithopod dinosaurs in the Late Jurassic Lourinhã Formation, Portugal. Oryctos. Mazzetta, G.V. & Blanco, R.E. 2001. Speeds of dinosaurs from the Albian-Cenomanian of Patagonia and sauropod stance and gait. Acta Palaeontologica Polonica 46: 235–246. Milàn, J. 2006. Variations in the morphology of emu (Dromaius novaehollandiae) tracks, reflecting differences in walking pattern and substrate consistency: ichnotaxonomical implications. Palaeontology 49: 405–420. Milàn, J., Avanzini, M., Clemmensen, L.B, Garciá-Ramos, J.C. & Piñuela, L. 2006. Theropod foot movement recorded from Late Triassic, Early Jurassic and Late Jurassic fossil footprints. In

35 Jesper Milàn – Vertebrate Ichnology References ______

Harris et al. (eds.). The Triassic-Jurassic Terrestrial Transition. New Mexico Museum of Natural History and Science Bulletin 37: 352–364. Milàn, J. & Bromley, R.G. 2005. Dinosaur footprints from the Middle Jurassic Bagå Formation, Bornholm, Denmark. Bulletin of the Geological Society of Denmark 52: 7–15. Milàn, J. & Bromley, R.G. 2006. True tracks, undertracks and eroded tracks, experimental work with tetrapod tracks in laboratory and field. Palaeogeography, Palaeoclimatology, Palaeoecology 231: 253–264. Milàn, J. & Bromley, R.G. (In press a). Do shod humans leave true tracks?. Ichnos. Milàn, J. & Bromley, R.G. (In press b). The impact of sediment consistency on track- and undertrack morphology: experiments with emu tracks in layered cement. Ichnos. Milàn, J., Bromley, R.G., Titschack, J. & Theodouru, G. (In press). A diverse vertebrate ichnofauna from a Quaternary eolian oolite from Rhodes, Greece. in: Bromley, R.G., Buatois, L.A., Márango, M.G., Genise, J.F. & Melchor, R.N. (eds.), Sediment-organism interactions: A multifaceted ichnology. SEPM Special Publications 88. Milàn, J., Christiansen, P. & Mateus, O. 2005. A three-dimensionally preserved sauropod manus impression from the Upper Jurassic of Portugal: Implications for sauropod manus shape and locomotor mechanics. Kaupia, Darmstädter Beiträge zur Naturkunde 14: 47– 52. Milàn, J., Clemmensen, L.B. & Bonde, N. 2004. Vertical sections through dinosaur tracks (Late Triassic lake deposits, East Greenland) – undertracks and other subsurface deformation structures revealed. Lethaia 37: 285–296. Milàn, J., Clemmensen, L.B., Buchardt, B. & Noe-Nygaard, N. (in prep). A Late Holocene tracksite in the Lodbjerg dune system, northwest Jylland, Denmark. New Mexico Museum of Natural History and Science Bulletin. Milàn, J. & Gierlinski, G. 2004. A probable thyreophoran (Dinosauria, Ornithischia) footprint from the Upper Triassic of southern Sweden. Bulletin of the Geological Society of Denmark 51: 71– 75. Milàn, J. & Loope, D.B. 2007. Preservation and erosion of theropod tracks in eolian deposits; examples from the Middle Jurassic Entrada Sandstone, Utah, USA. The Journal of Geology, 115: 375–386.

36 Jesper Milàn – Vertebrate Ichnology References ______

Milàn, J., Loope, D.B & Bromley, R.G. (Submitted). Dinosaur tracks showing evidence of individual behavior from the Lower Jurassic Navajo Sandstone, Coyotes Buttes locality, Arizona-Utah, USA. Acta Palaeontologia Polonica. Morales, M. & Bulkley, S. 1996. Paleoichnological evidence for a theropod dinosaur larger than Dilophosaurus in the Lower Jurassic Kayenta Formation. In Morales, M. (ed.). The Continental Jurassic. Museum of Northern Arizona, Flagstaff 60: 143–146. Moratalla, J.J., Sanz, J.L. & Jimenez, S. 1988. Multivariate analysis on Lower Cretaceous dinosaur footprints: Discrimination between ornithopods and theropods. Geobios 21: 395–408. Ostrom, J. H. 1972. Were some dinosaurs gregarious? Palaeogeography, Palaeoclimatology, Palaeoecology 11: 287–301. Parker, L.R. & Balsby, J.K. 1989. Coal mines as localities for studying dinosaur trace fossils, p 353–359. In Gillette, D.G. & Lockley, M.G. (eds.). Dinosaur tracks and traces. Cambridge University Press. 454 pp.

Parker, L.R. & Rowley, R.L. 1989. Dinosaur footprints from a coal mine in east-central Utah, p. 361–366. In Gillette, D.G. & Lockley, M.G. (eds.). Dinosaur tracks and traces. Cambridge University Press. 454 pp.

Peterson, W. 1924. Dinosaur tracks in the roofs of coal mines. Natural History 24: 388–397. Pleijel, C. 1975. Nya dinosauriefotspår från Skånes Rät-Lias. Fauna och Flora 3: 116–120. Rich, T.H., Gangloff, R. A. & Hammer, W.R. 1997. Polar dinosaurs. In Currie, P. & Padian, K. (Eds.): Encyclopedia of Dinosaurs, 562–573. Academic Press. New York. Rich, T.H., Vickers-Rich, P. & Gangloff, R.A. 2002: Polar dinosaurs. Science 295: 979–980.

Romano, M. & Whyte, M.A. 2003. Jurassic dinosaur tracks and trackways of the Cleveland Basin, Yorkshire: preservation, diversity and distribution. Proceedings of the Yorkshire Geological Society 54: 185–215. Sarjeant, W.A.S. 1972. Vertebrate tracks from the Permian of Castle Peak, Texas. Texas Journal of Science 22: 343–366. Sarjeant, W.A.S. 1974. A history and bibliography of the study of fossil vertebrate footprints in the British Isles. Palaeogeography, Palaeoclimatology, Palaeoecoclogy 16: 265–378. Santos, V.F., Lockley, M.G., Meyer, C.A., Carvalho, J., Carvalho, A.M.G.D., Moratella, J.J. 1994. A new sauropod tracksite from the Middle Jurassic of Portugal. Gaia 10: 5–13.

37 Jesper Milàn – Vertebrate Ichnology References ______

Thulborn, R.A. 1982. Speeds and gaits of dinosaurs. Palaeogeography, Palaeoclimatology, Palaeoecology 38: 227–256. Thulborn, T. 1990. Dinosaur tracks. Chapman and Hall. 410 pp. Tresise, G. & Sarjeant, W.A.S. 1997. The tracks of Triassic vertebrates. The Stationary Office. 204 pp. Weems, R.E. 2006. Locomotor speeds and patterns of running behaviour in non-maniraptoriform theropod dinosaurs. In Harris et al. (eds.). The Triassic-Jurassic Terrestrial Transition. New Mexico Museum of Natural History and Science Bulletin 37: 379–389. Whyte, M.A., Romano, M. & Elvidge, D.J. 2007. Reconstruction of Middle Jurassic dinosaur- dominated communities from the vertebrate ichnofauna of the Cleveland Basin of Yorkshire, UK. Ichnos 14: 117–129. Whyte, M., Romano, M., Hudson, J.G., & Watts, W. 2006. Discovery of the largest theropod dinosaur tracks known from the Middle Jurassic of Yorkshire. Proceedings of the Yorkshire Geological Society 56: 77–80.

38 Jesper Milàn – Vertebrate Ichnology Appendix ______

A probable thyreophoran (Dinosauria, Ornithischia) footprint from the Upper Triassic of southern Sweden

JESPER MILÀN & GERARD GIERLINSKI

Milàn, J. & Gierlinski, G. 2004–10–22: A probable thyreophoran (Dinosauria, Ornithischia) footprint from the Upper Triassic of southern Sweden. Bulletin of the Geological Society of Denmark, Vol. 51, pp. 71–75, Copenhagen. © 2004 by the Geological Society of Denmark. ISSN 0011–6297.

A curious blunt-toed tridactyl footprint of a relatively large trackmaker is stored in the Geological Museum in Copenhagen. The footprint was found nearly 50 years ago in the Rhaetian coal-bearing strata mined in the Gustav Adolf Mine, near Höganäs, Scania, Southern Sweden. The morphology of the specimen suggests that it was left by an early advanced thyreophoran dinosaur, in this case the earliest known.

Key-words: Footprint, Thyreophoran dinosaur, Late Triassic, Scania, Sweden.

Jesper Milàn [[email protected]], Geological Institute, Østervoldgade 10, 1350 Copenhagen K, Denmark. Gerard Gierlinski [[email protected]], Polish Geological Institute, ul. Rakowiecka 4, 00-975 Warszawa, Po- land. 16 June 2004.

Dinosaur footprints from the Upper Triassic, Rhae- tian, Höganäs Formation, in southern Sweden were first reported by Bölau (1952), who described them as theropod footprints without assigning them to any ichnotaxon. Other tracks, also of theropod origin, were found in the Early Jurassic strata of Helsingborg and Valläkra (Pleijel 1975; Ahlberg & Siverson 1991). Gierlinski & Ahlberg (1994) assigned Late Triassic tracks from southern Sweden to the ichnotaxon Grallator (Eubrontes) giganteus Hitchcock, 1845, and the Lower Jurassic tracks to Grallator (Eubrontes) soltykovensis Gierlinski, 1991 following subichnogeneric designation proposed by Olsen and Galton (1984). Re- cently, these ichnotaxa have been revised as Eubrontes giganteus and Kayentapus soltykovensis (see Gierlinski 1996; Olsen et al. 1998). All tracks so far described from the Höganäs Formation have been clearly of theropod origin, characterized by a functionally tridactyl pes, the third digit being the longest, and having relatively narrow claws (Fig. 1). However, the specimen exhibited on display at the Geological Museum in Copen- hagen (MGUH 27219) (Fig. 2), differs from all other Early Mesozoic tracks hitherto described from southern Sweden. The track was found in the roof of the Fig. 1. Typical theropod track, Grallator (Eubrontes) giganteus Gustav Adolf coal mine, Höganäs, between 1950 and from the Gustaf Adolf coal mine in Höganäs, southern Swe- 1955 and was purchased by the Geological Museum den. Notice the long slender digit impressions and the sharp in Copenhagen from a relative to one of the workers laterally compressed claw imprints. Scalebar 5 cm. Figure from the mine. Vertebrate tracks are often found in modified after Gierlinski & Ahlberg 1994.

Milàn & Gierlinski: A probable thyreophoran footprint · 71

DGF Bulletin 51-1.pmd 71 26-10-2004, 14:47 Fig. 2. The probable ornithischian footprint (MGUH 27219), from the Bjuv Member mined in the Gustav Adolf Mine in Scania, Sweden. Scale bar is 5 cm. Photo Ole Bang Berthelsen.

the roofs of coal mines, because the techniques used trackwalls sensu Brown (1999) in the middle digit in mining exposes large surfaces at lithological impression and vertical trackwalls in parts of the left boundaries, which are ideal for track preservation and digit. A now repaired fracture running through the recognition (Parker & Balsley 1989; Parker & Rowley slab has caused the loss of the distal end of the right 1989). In the Gustav Adolf mine, the Rhaetian Bjuv digit. Faint colour variation in the surface of the slab Member of the Höganäs Formation was mined. Ac- indicates the approximate extend of the missing digit cording to the Geological Museum, Copenhagen, the (Fig. 3). specimen was found together with three other foot- The footprint is 26 cm long, measured from the tip prints, presumably forming a short trackway, but a of the middle digit to the proximal edge of the meta- recent attempt to locate the storage of the other three tarsal-phalangeal area. The width of the track is 28.5 ichnites from the presumed trackway was unsuccess- cm; including the estimated extent of the missing part ful. of the left digit. The track is uniformly impressed to a depth of 3–3.5 cm. The digit impressions become shallower in the distal parts and the metatarsal area uniformly slopes down to the level of the original sedi- Footprint description ment surface during the proximal 7 cm of the track. The faint contour of a distal phalangeal pad and the The footprint (Fig. 2) is preserved as a natural cast possible trace of a short, wide claw, are preserved in filled with sandy heterolithic siltstone. Parts of the the impression of the right digit. The divarication track are still covered with a thin layer of coal origi- angle between the two outer digits equals 76º. nating from the coal-seam in which it was originally impressed. The track is tridactyl with short, blunt digits, which are subequal in length. The shapes of the digit impressions are well defined, forming steep to vertical

72 · Bulletin of the Geological Society of Denmark

DGF Bulletin 51-1.pmd 72 26-10-2004, 14:47 Discussion The morphology of the Copenhagen specimen clearly differs from the grallatorid footprints hitherto reported from Höganäs Formation. Theropod tracks in general are characterized by being longer than wide, and having long, slender digits terminating in sharp claw impressions (Lockley 1991). The Copenhagen specimen has short, broad and blunt-ended digits, and the track is broader than long. Such morphology is more consistent with the footprints of large ornithopods from the Cretaceous, than those known from the Jurassic. Among the Jurassic forms there are only few, relatively rare exceptions, which correspond with the discussed specimen. Those exceptions represent sup- posedly Stegosaurian footprints from the Late Juras- sic, Oxfordian (Gierlinski & Sabath 2002) and the Early Jurassic, Upper Hettangian, ichnotaxon of Moyenisau- ropus karaszevskii Gierlinski, 1991, which was proposed to be of proto-stegosaurian origin (Le Loeuff et Fig. Interpretative drawing of the specimen with measure- al. 1999; Gierlinski 1999). ments given. One possible explanation that should be considered for the ornithischian-like shape of this footprint is that it is an undertrack. Undertracks form when an animal left tracks in layered sediments. In that case the weight of the animal’s foot not only forms a track on the trampled surface, i.e. the tracking surface, but the layers subjacent to the tracking surface are also subject to deformation. Experimental work with track

Fig. 4. Formation of undertracks demonstrated with an emu (Dromaius novaehollandiae) footprint formed in a layered package of artificial sediments. A, a track formed in layered sediment not only leaves an impression on the trampled surface, but the subjacent layers are also subject to deformation. When the layered package is split horizontally, B, an impression of the track can be found on the subjacent horizons, each undertrack being shallower, more rounded, and less detailed downward. The length of the true track is 19.5 cm. Figure modified from Milàn & Bromley (in press).

Milàn & Gierlinski: A probable thyreophoran footprint · 73

DGF Bulletin 51-1.pmd 73 26-10-2004, 14:47 formation in artificial substates (Milan & Bromley Håkansson, for critical reading and suggestions to an 2003, in press), demonstrate how the morphology of early draft of the manuscript, and to the reviewers undertracks differs from that of true tracks, in that Richard G. Bromley and Hans Jørgen Hansen for their the digit impressions become shallower, wider, more positive and constructive reviews. rounded and preserve fewer anatomical details, for each successive layer downward (see Fig. 4). The broad, rounded digits in the track discussed do have a superficial undertrack-like appearance. References However, the presence of steep to vertical trackwalls in parts of the track makes the undertrack hypothesis Ahlberg, A. & Siverson, M. 1991: Lower Jurassic dinosaur foot- unlikely. Undertracks will have sloping trackwalls prints in Helsingborg, southern Sweden. Geologiska without the well-defined edges between the tracking Föreningens i Stockholm Förhandlingar 113, 339–340. Brown Jr., T. 1999: The science and art of tracking. 219 pp. surface and the track as present in the Copenhagen Berkley Books, New York. specimen. Bölau, E. 1952: Neue Fossilfunde aus dem Rhät Schones und In conclusion the track is a true track, unaltered by ihre paläogeographisch-ökologische Auswartung. Geolo- pen contemporaneous erosion. The morphology of the giska Föreningens i Stockholm Förhandlingar 74, 44–50. track, being broader than long, with short, broad dig- Gierlinski, G. 1991: New dinosaur ichnotaxa from the Early its is consistent with that of supposed thyreophoran Jurassic of the Holy Cross Mountains, Poland. Paleogeog- tracks from the Early Jurassic of Poland. Thus, the raphy, Palaeoclimatology, Palaeoecology 85, 137–148. Gierlinski, G. 1996: Dinosaur ichnotaxa from the Lower Juras- Late Triassic age of the described footprint suggests sic of Hungary. Geological Quarterly 40, 119–128. that the track represents the hitherto oldest known Gierlinski, G. 1999: Tracks of a large thyreophoran dinosaur footprint of a large thyreophoran dinosaur. from the Early Jurassic of Poland. Acta Palaeontologica Polonica 44, 231–234. Gierlinski, G. & Ahlberg, A. 1994: Late Triassic and Early Ju- rassic dinosaur footprints in the Höganäs Formation of southern Sweden. Ichnos 3, 99–105. Danish Summary Gierlinski, G. & Sabath, K. 2002: A probable stegosaurian track from the Late Jurassic of Poland. Acta Palaeontologica Po- Et bemærkelsesværdigt fossilt dinosaur fodspor fra lonica 47, 561–564. Øvre Trias, udstillet på Geologisk Musum, Køben- Hitchcock, E.H. 1845: An attempt to name, classify, and de- havn har vist sig at være en hidtil ukendt form. Spo- scribe the animals that made the fossil footmarks of New ret blev fundet i loftet af Gustaf Adolf kulminen ved England. Proceedings of the 6th Annual Meeting of the As- Höganäs omkring 1950–55. Alle hidtil beskrevne di- sociation of American Geologists and Naturalists, New nosaur fodspor fra det skånske område stammer fra Haven, Connecticut 6, 23–25. Lockley, M. 1991: Tracking Dinosaurs. 238 pp. Cambridge rovdinosaurere. Rovdinosaur spor (Fig. 1) er karak- University Press, Cambridge. teristisk ved at de tretåede, har aftryk af lange slanke Leouff, J.L., Lockley, M., Meyer, C. & Petit, J.-P. 1999. Discov- tæer med skarpe kløer, samt at den midterste tå er ery of a thyreophoran trackway in the Hettangian of cen- betydelig længere end de to ydre tæer, hvilket gør tral France. C. R. Académie des Sciences Paris/Sciences de sporet længere end det er bredt. Det heri beskrevne la Terre et des Planetas/Earht and Planetary Science 328: spor (Fig. 2) afviger fra denne morfologi i at være bre- 215–219. dere end det er langt at og have korte, tykke, afrun- Milàn, J. & Bromley, R.G. 2003: How to distinguish between true tracks and undertracks – experimental work with arti- dede tæer. Denne spormorfologi er fundet i spor fra ficial substrates. 1st EAVP (European Association of Verte- panserøgler, Thyreophora, fra nedre Jura i Polen. Den brate Palaeontologists) Meeting 15th–19th of July, Basel, øvre Triassiske alder af det heri beskrevne spor, tyder Switzerland. Abstracts of papers and posters with program, på at sporet repræsenter det hidtil ældste kendte spor p 30. fra en pansret dinosaur. Milàn, J. & Bromley, R.G. (in press). The impact of sediment consistency on track- and undertrack morphology: experiments with emu tracks in layered cement. In Rainforth, E.C. & McCrea, R.T. (eds): Fossil Footprints. Indiana University Press. Acknowledgements Olsen, P.E. & Galton, P.M. 1984: A review of the reptile and amphibian assemblages from the Stormberg of South Af- We are grateful to Arne Thorshøj Nielsen for his im- rica, with special emphasis on the footprints and the age of mense help in locating information about the origin the Stormberg. Palaeontographica Africana (Haughton of the footprint, to Ole Bang Berthelsen for photo- Memorial Volume) 25, 87–110. graphic work, to Steen Lennart Jacobsen for provid- Olsen, P.E., Smith, J.B. & McDonald, N.G. 1998: The material of the species of the classic theropod footprint genera ing a plastercast of the track. A special thank to Eckart

74 · Bulletin of the Geological Society of Denmark

DGF Bulletin 51-1.pmd 74 26-10-2004, 14:47 Eubrontes, Anchisauripus and Grallator (Early Jurassic, Hartford and Deerfield basins, Connecticut and Massachu- setts, U.S.A.). Journal of Vertebrate Paleontology 18, 586– 601. Parker, L.R. & Balsley, J.K. 1989: Coal mines as localities for studying dinosaur trace fossils. In Gillette, D.D., & Lockley, M.G. (eds): Dinosaur Tracks and Traces, 353–359. Cam- bridge University Press, Cambridge. Parker, L.R. & Rowley, R.L. 1989: Dinosaur footprints from a coal mine in east-central Utah. In Gillette, D.D. & Lockley, M.G. (eds.): Dinosaur Tracks and Traces, 361–366. Cam- bridge University Press, Cambridge. Pleijel, C. 1975: Nya dinosauriefotspår från Skånes Rät-Lias. Fauna och Flora 3, 116–120.

Milàn & Gierlinski: A probable thyreophoran footprint · 75

DGF Bulletin 51-1.pmd 75 26-10-2004, 14:47 Jesper Milàn – Vertebrate Ichnology Appendix ______

Vertical sections through dinosaur tracks (Late Triassic lake deposits, East Greenland) ± undertracks and other subsurface deformation structures revealed

JESPER MILAÁ N, LARS B. CLEMMENSEN AND NIELS BONDE

MilaÁn, J., Clemmensen, L.B. & Bonde, N. 2004 09 15: Vertical sections through dinosaur tracks (Late Triassic lake deposits, East Greenland)±undertracks and other subsurface deformation structures revealed. Lethaia, Vol. 37, pp. 285±296. Oslo. ISSN 0024-1164.

Tracks and trackways of theropod dinosaurs (Grallator footprints) are abundant in the Late Triassic lake sediments of East Greenland. For this study we selected a rather diffuse theropod track preserved on the upper surface of a red heterolithic mudrock, and a better preserved track seen on the upper surface of a greyish mudrock. In order to examine undertracks and other subsurface deformation structures, both slabs were sectioned vertically at closely-spaced intervals, perpendicular to the length of the axis of the impression of digit III. Each section was subsequently polished and internal structures revealed. The digit impressions of both tracks were associated with well-de®ned undertracks which were cut by deep and narrow claw imprints at the distal end of the digit impressions. Marginal ridges at the tracking surfaces were typically associated with subsurface marginal folds. The marginal ridges were asymmetrically developed suggesting an outward movement of the proximal part of the foot, probably during the kick-off; this is in contrast to what is observed in tracks from Lower Jurassic theropods. The study shows that cross-sections through dinosaur tracks display large structural variation and it is suggested that some disturbed layers in continental deposits could be the result of trampling by vertebrates. & Lake sediments, Late Triassic, Theropod tracks, undertracks, walking dynamics.

Jesper MilaÁn [[email protected]], Lars B. Clemmensen & Niels Bonde, Geological Insti- tute, éster Voldgade 10, DK-1350 Copenhagen K, Denmark; 27th November 2003, revised 25th March 2004.

Vertebrate tracks should not be regarded as deforma- standing of the formation of undertracks and other tion structures in the sediment surface only. The related deformation structures. Avanzini (1998) exam- weight of an animal not only affects the surface it ined the walking dynamics of theropod dinosaurs by walks upon, the tracking surface sensu FornoÂs et al. sectioning Grallator footprints and studied the hori- (2002), but subjacent horizons are also subject to zontal deformation of the layers resulting from side- deformation caused by the pressure from the foot ways foot movements during the stride. transferred radially outwards during the stride (Allen The aim of this paper is to supplement the study of 1997). This leads to the formation of undertracks and Gatesy et al. (1999) by the method of Avanzini (1998) other sub-sediment deformation structures. and examine vertebrate tracks in vertical section, by Only within the last years has the formation of sectioning two slabs of ancient lake deposits, each undertracks and sub-sediment deformation structures containing a theropod track. The slabs are from the been systematically described and incorporated into Late Triassic Fleming Fjord Formation in central East the description and interpretation of the footprints. Greenland (Fig. 1). Vertical sectioning is an additional Gatesy et al. (1999) sectioned transversely or sagitally method to reveal the complex formation of under- Late Triassic theropod tracks made in deep mud to tracks, deep cuts from claw imprints and other sub- allow examination and reconstruction of the subsurface deformation structures, resulting from the sediment foot movements of the trackmaker, and dynamic interaction between the trackmaker and the FornoÂs et al. (2002) described in detail the sub- substrate. sediment deformation structures associated with Pleistocene goat tracks in aeolianites from Mallorca. Experimental work with arti®cial sediments (Jack- Stratigraphy son 2002; Milan & Bromley 2002, 2003; MilaÁn& Bromley in press), has greatly added to the under- The uppermost part of the Late Triassic Flemming

DOI 10.1080/00241160410002036 # 2004 Taylor & Francis 286 Jesper MilaÁn et al. LETHAIA 37 (2004)

Fjord Formation, the érsted Dal Member, is well- Fauna exposed in mountain slopes along Carlsberg Fjord in central East Greenland (Fig. 1) (Clemmensen et al. The vertebrate assemblages of the Fleming Fjord 1998). In this region the érsted Dal Member has a Formation are very rich and diverse (Jenkins et al. thickness of approximately 135 m and is composed of 1994) including several dinosaurs. Based on skeletal a lower unit of cyclically-bedded, structureless red, material, the dominating element in the dinosaur clay-rich mudrocks and thin greyish red or green fauna were prosauropods (Plateosaurus engelhardti)of siltrich, heterolithic mudrocks with wave-ripples (the which several well-preserved specimens have been Carlsberg Fjord Beds) and an upper unit of cyclically found, while only a single incomplete, so far inde- bedded variegated or grey mudrocks and yellowish terminable, theropod skeleton has been found grey marlstones (the Tait Bjerg Beds). Both units are (Jenkins et al. 1994). The non-dinosaurian vertebrate lacustrine and formed in a large lake system that was fauna comprises a rich ®sh fauna (hybodont sharks, situated at approximately 35° N in the central part of actinopterygians, coelacanths and lung®sh). Labyrin- Laurasia (Clemmensen et al. 1998). The climate of the thodont amphibians are common, with the plagio- area was initially dry and steppe-like but gradually saurid Gerrothorax being the most frequently changed to a more humid and temperate climate. encountered vertebrate in these deposits, and several Palaeomagnetic and cyclostratigraphic studies by larger stereospondyls such as cyclotosaurs, capitosaurs Kent & Clemmensen (1996) and Clemmensen et al. and metoposaurs. The reptilian fauna consists of (1998) suggest that the érsted Dal Member is of Late turtles (cf. Proganochelys), phytosaurs, aetosaurs Norian to Early Rhaetian age and covers a time span of (Aetosaurus ferratus and Paratypothorax) and ptero- c. 2.m.y. saurs (Eudimorphodon cromptonellus Jenkins et al. 2001). Furthermore the fauna contains a diverse mammalian assemblage including teeth of Kueh- neotherum, cf. Brachyzostrodon and triconodonts, as well as a skull with postcranial elements of Haramiya- via clemmenseni (Jenkins et al. 1997), the only skeleton known worldwide. This faunal assemblage is one of the richest Late Triassic, Norian, occurrences of continental vertebrates in the world. The vertebrate ichnofauna is dominated by tridactyl tracks of small theropod dinosaurs. The tracks are common at many levels in the érsted Dal Member, and are particularly abundant on the upper surfaces of the wave-rippled heteroliths or the grey mudrocks. One locality was so rich in tracks that it was referred to as `The Raceway' by Jenkins et al. (1994). In contrast to the numerous theropod trackways, only three trackways can be attributed to prosauropod dinosaurs (Jenkins et al. 1994; Clemmensen et al. 1998; Gatesy et al. 1999; Gatesy 2001). This contradicts the skeletal records from the area, in which the prosauropods by far outnumber the theropods. There are many smaller footprints and trackways distributed in the Fleming Fjord Formation, some of which seem to be amphibian and others possibly of crocodilian origin (Jenkins et al. 1994) and a single hitherto undescribed Chirotherium trackway (Bonde pers. obs.). The formation further comprises an extensive invertebrate ichnofauna (Bromley & Asgaard 1979).

Fig. 1. Jameson Land, East Greenland. Outcrops of Late Triassic lake sediments (Fleming Fjord Formation) are indicated in black. Methods The tracks were collected at the two localities with thick successions of mudrocks in the érsted Dal Member, Wood Bjerg and Macknight Bjerg. Modi®ed after Clemmensen et al. (1998). For this study we selected a rather diffuse theropod LETHAIA 37 (2004) Dinosaur tracks from the Late Triassic 287 track preserved on the upper surface of a red trical as the proximal pad impressions of digit IV heterolithic mudrock (track 1), and a better-de®ned extend further back than those of digits III and II. The theropod track preserved in a grey mudrock (track 2). claw impressions of digit III is always directed inward The two slabs were sliced vertically, perpendicular to towards the midline of the trackway (Lockley 1991; the long axis of the impression of digit III using a Farlow et al. 2000). stationary rock saw. The surface of each slice of rock The asymmetry of the proximal end of the foot and was subsequently polished to allow examination of the the orientation of the claw of digit III are important internal structures of the slab. The track formed in indicators of right or left feet if only single footprints reddish heterolith showed very little contrast in the are examined, as these characters can be recognized colours, and in order to enhance the sedimentary even in weathered and eroded tracks. The producers of structures, the slice was submerged for 20 minutes in a Late Triassic Grallator tracks are supposed to be small 10 % acetic acid solution, which enhanced the to medium-sized Coelyphysis-like theropods (Lockley structures signi®cantly. 1991). Each section was afterwards scanned directly on a ¯atbed scanner at 600 dpi. To protect the glass surface of the scanner a sheet of transparent plastic was placed Terminology between the rock slice and the glass. Alcohol was used as a contact medium between the rock and the plastic The animal responsible for fossil footprints is termed sheet. Digital colour and contrast enhancing was the `trackmaker', and the sediment surface it walks subsequently performed using Corel Photo-paint 11. upon is termed the `tracking surface' sensu FornoÂs et Four sections (A±D) from each of the two tracks were al. (2002) (Fig. 3). The track emplaced on the actual examined. For each track, section A is cut through the tracking surface is termed the `true track' (Lockley claw imprint of digit III, section B is from the middle 1991). The sedimentary ®lling of the track forms a of the impression of digit III, section C is cut through `natural cast' of the track (Lockley 1991). If the the middle of the impressions of digits II and IV and trackmaker's foot sinks to a considerable depth in through the basal part of the impression of digit III. the sediment, then the vertical parts of the track is Section D is cut through the proximal part of the termed `track walls' (Brown 1999). If the track walls footprint, the metatarsal area. are sloping as a result of the dynamic movement of the trackmaker's foot during the stride, then the actual Grallator footprints Grallator Hitchcock, 1858 (Fig. 2) is one of the oldest scienti®cally named tetrapod ichnogenera, ®rst described by Hitchcock (1858) as the footprints of large ¯ightless birds under the name Ornithichnites. The ichnogenus spans from the Late Triassic throughout the Jurassic, and was presumably made by several different small to medium sized theropods. Grallator tracks are mostly tridactyl, consisting of impressions digits II, III and IV. In rare specimens, traces of digit I, the posteromedially-orientated hallux (Irby 1995) are preserved. Since digit I is situated at an elevated position on the metatarsus (Christiansen 1997), impressions of the digit mostly occur in tracks deep tracks such as those described by Gatesy et al. (1999). The individual digit impressions in Grallator footprints have well-de®ned digital pads and impressions of long, slender claws, re¯ecting the phalangeal skeleton inside. Digit II, which consists of three phalanges, has two prominent pads covering the joints. The three phalangeal pads cover the four Fig. 2. Idealized Grallator track from a right foot with pedal phalanges in digit III. The claw of digit III is offset skeleton superimposed. The digital pads correspond with number towards the midline of the trackway. Digit IV consists of phalanges in digit II and III, but not in digit IV where the individual phalanges are too short. Notice the inward orientation of four or ®ve small phalangeal pads. of the claw of digit III and the pronounced asymmetry in the The proximal end of Grallator tracks is asymme- proximal part of the track. Modi®ed after Olsen et al. (1998). 288 Jesper MilaÁn et al. LETHAIA 37 (2004)

undertrack according to experimental work by MilaÁn & Bromley (2003, in press), in that the track appears broad and rounded and apparently lacks preservation of ®ner anatomical features, but the digit impressions and gross overall shape of the track. The track is approximately 24 cm long and 14 cm wide, and shows a divarication angle of 56 degrees between the impression of digits II and IV. The shallow and unde®ned nature of the track makes it dif®cult to obtain accurate measurements. When the footlength of 24 cm is used in the formula developed Fig. 3. Schematic section through the distal end of a theropod digit by Alexander (1976), that hip-height equals approxi- impression, with features mentioned in text. Tracking surface (Ts), marginal ridge (Mr), sedimentary ®ll (Sf), track wall (Tw), claw mately 4 times the foot length, we arrive at an imprint (Ci), true track (Tt), undertrack (Ut) and undertrack estimated hip-height of the trackmaker of 1 m. From formed by the claw impression (Utc). this an estimated total length of the animal, supposing that it ®ts the Coelyphysis bodyplan, would be around 4 m (Per Christiansen, personal communication impression of the track will appear longer at the 2002). surface than at the bottom of the track. In this case the No division of phalangeal pads is visible and the track in the bottom is the true track and the hole in the shape of the digit impressions appears rounded with sediment surface is termed the `overall track' (Brown short, blunt digits. The outline, however, re¯ects the 1999). On the surface around the track, a marginal rim slight turn of the claw of digit III and the asymmetry in of displaced material is formed, and depending on the the proximal end of the track, identi®es the track as a substrate consistency, marginal thrusts in the rim and right pes. radial fractures in the sediment around the track can A raised rim of displaced material is present at the form (Allen 1997). The weight of the trackmaker's foot proximal part of digit II's impression, suggesting the is transferred radially outward into the sediment trackmaker made a slight outward movement of the around and below the trackmaker's foot (Allen proximal part of the foot during progression. Lateral 1997). If the track is made in layered sediments an movements in the proximal parts of theropod feet impression of the track will be formed at horizons during walking, have previously been described by subjacent to the tracking surface; these tracks are Avanzini (1998), and in that case the movement termed `undertracks' (Lockley 1991) or `transmitted occurred inward contrary to track 1 in which the Tracks' (Thulborn 1990), for this study the term movement was outward. undertracks censu Lockley (1991) will be used. Under- The slab containing the track is a reddish heterolith tracks differ from true tracks in being less detailed 40±45 mm thick. Grain size varies from clay to ®ne- downward as demonstrated by experimental work by grained sand. The silt-sized particles are clastic grains MilaÁn & Bromley (2003, in press). If the trackmaker (predominantly quartz and a few clay peloids) while possesses sharp claws on the digits, these will, the sand-sized particles are rounded clay pelloids, and depending on the properties of the substrate leave these grains form well-de®ned laminae separated by deep claw imprints, usually to a deeper level that the more clay-rich laminae. The clay-rich laminae are red- digit impressions. In some cases the deep imprints of coloured and contain about 55% of carbonate. The the claws can cut through the layers with undertracks clay-rich matrix displays a clotted texture indicating a from the digit impressions and form their own pedogenic overprint. undertracks in the subjacent layers. The slab is divided into three units; a basal laminated (10 mm) unit, a middle unit with small- scale cross-lamination (20±25 mm), and an upper unit (10 mm) displaying a vague and indistinct lamination Material (Fig. 5A). The unidirectionally dipping cross-lamina- Track 1 tion suggests that sediment transport can be of current or combined current and wave-induced origin. All This track was collected in 1991 from the track- units display a ®ning-upward trend suggesting deposi- bearing beds in the Carlsberg Fjord Formation at tion during gradually decreasing energy. At the very Wood Bjerg. At the surface the track appears shallow top of the slab is the tracking surface, which consists of (14 mm at the deepest) (Fig. 4A). The appearance of red mudstone. In the deeper part of the track there is a the track is typical of what could be expected from an ®lling of light-coloured mudstone with intrafor- LETHAIA 37 (2004) Dinosaur tracks from the Late Triassic 289

Fig. 4. The two Grallator tracks as they appeared on the surface. The termination of each digit impression is indicated by roman letters, according to digit number. Numbered lines indicates the sections used in this study. &A. Track 1 is badly preserved and displays little more than the gross outline of the track. Of ®ner anatomical details, only the claw impression of digit III is vaguely recognizable. &B. Interpretative drawing of track 1, with the marginal rims of displaced material indicated by shaded colour. &C. Track 2 exhibits a higher state of preservation. Individual digital pads are distinguishable on digit III's impression and prominent claw marks are present at the end of all three digit impressions. &D. Interpretative drawing of track 2, showing the individual phalangeal pads. Marginal rims of displaced material is indicated by shaded colour. 290 Jesper MilaÁn et al. LETHAIA 37 (2004) mational mud clasts that clearly originated from the (Fig. 4C). The length is 21.3 cm and the width is semi-lithi®ed tracking surface. The surface of the slab, 14.9 cm, suggesting a trackmaker with a hip height of including the track is cut by several orders of around 85 cm and a total estimated body-length of polygonal desiccation structures. The underside of 3.5 m, a little smaller than the trackmaker responsible the slab is covered with numerous small Diplichnites for track 1. The divarication angle between digits II trackways preserved in positive relief and the arthro- and IV is 50 degrees, which is within the normal range pod resting trace Rusophycus, indicating a ¯uviatile for small to medium sized theropods (Farlow et al. and possibly lacustrine environment before the track- 2000). The track was emplaced in a layer of relatively bearing layer was deposited (Bromley & Asgaard thin and ®rm mud, which has caused the imprints of 1979). Clemmensen et al. (1998) suggested that most the digits to be only slightly connected as the shallow of the track-bearing beds formed during ¯ooding of interpad spaces separating the digits did not reach the marginal mud ¯ats by lake water. Most of the dinosaur tracking surface and failed to leave imprints. Impres- tracks and trackways apparently formed during sub- sions of the individual digital pads are recognizable in sequent periods of exposure and desiccation of the digits II and III, while the imprint of digit IV still ¯ood-generated deposits. contains some of the covering sediments, hindering identi®cation of individual digital pads. Impressions Track 2 of long slender claws are present at the ends of the impressions of digits II and III while the proximal part This track was collected in 1995 at MacKnight Bjerg in of digit IV's impression is partly ®lled with clay. A the transition zone between the Carlsberg Fjord Beds prominent rim of material displaced by the digits is and the Tait Bjerg Beds. The track is tridactyl with long present between the impressions of digit IV and III, slender digit impressions each terminating in narrow caused by the upward and forward movement when claw imprints. The claw imprint of digit III is offset to the foot is lifted. Low rims of displaced material are the right of the long axis of the digit impression, and present along the outline of the track. the impressions of the digital pads in the proximal end The slab containing the track is a greyish mudrock of the track have a pronounced asymmetry towards with a thickness of about 40 mm. A thin section study the right side which identi®es the track as a left pes indicates that the mudrock contains a little (10±20%)

Fig. 5. Sections showing structures in the track-bearing slabs. &A. Reddish heterolith composed of three distinct units. The basal unit is ®nely laminated, then follows a unit with small-scale cross lamination and the top unit is vaguely laminated. Total thickness of slab is 40±45 mm. &B. Grey mudrock divided into three units. The basal unit has deformed lamination, the middle unit is well laminated, but contain also layers that have been broken up. The top unit is structure less to vaguely laminated. Total thickness of the slab is 40 mm. Scale bar 1 cm. LETHAIA 37 (2004) Dinosaur tracks from the Late Triassic 291 clastic material. The clastic material is composed of At the base is a unit (about 10 mm thick) with chaotic silt-sized grains (predominantly quartz) set in a ®ne- or deformed lamination. Then follows laminated grained matrix. There is around 30% carbonate in the sediment (about 20 mm thick), with two or three matrix. The slab is composed of three units (Fig. 5B). layers that have been broken up either during subaerial

Fig. 6. Vertical sections through track 1. The sections are cut perpendicular to the length axis of digit III, see text for details. Section A is cut through the tip of digit III's impression. Arrow indicates the undertrack formed by the claw impression. Section B is cut through the middle of digit III's impression. Arrows mark the width of the undertrack formed at the horizon approximately 1 cm below the tracking surface. Section C is cut through the impressions of digits II, III and IV. The division of the digits are best recognizable in the undertracks, white arrows indicate the up-bulged material between the digits. The black arrow marks a deep outward directed fault. Section D is cut through the proximal end of the track, notice prominent disturbance from a sideways movement of the foot in the right side of the track (indicated by arrow). All sections in rear view. 292 Jesper MilaÁn et al. LETHAIA 37 (2004) exposure or by an earlier phase of trampling, and reveals more information about the con®guration of ®nally a unit (about 15 mm) of structureless to the digits than the true track, as the separation vaguely laminated sediment. At the top is the tracking between digits II, III and IV is represented by a surface of dark-coloured mudrock. The bottom of the prominent upward bulging of the layers between the track is ®lled with light-coloured muddy sediment digit impressions. This interdigital upward bulging of containing numerous intraformational clasts. These the sediment must also have been present at the true clasts originate from the semi-lithi®ed tracking surface track, but has been secondarily lost. A prominent fault and were deposited in the track simultaneously with or cut the layers, projecting downward and outward from immediately after track formation. the bottom of the right side of the surface track (see arrow on Fig. 6C). Track 1, section A Track 1, section D This section passes through the tip of the claw impression of digit III where the slender, sharp claw This section is through the proximal part of the track, has cut down through the vaguely laminated unit and close to the metatarsal joint. The track is formed in the left a narrow groove, traceable in the cross-strati®ed upper vaguely strati®ed unit of the slab (Fig. 6D). A layer (Fig. 6A). The soft parts, surrounding the base of prominent raised marginal rim of displaced surface the claw, have formed a bowl-shaped depression in the material is present at the right side of the footprint, tracking surface, from which the claw cut protrudes originating from a sideways and outwards movement downward, suggesting that the foot was at ®rst placed of the foot during the stride. While the track itself has ¯at on the ground, and the deep cut from the claw was only left a very shallow undertrack, the marginal rim is made during the subsequent kick-off where the weight also prominent in the undertrack. was transferred to the tip of the digit. Both the bowl- shaped depression and the cut from the claw are ®lled Track 2, section A sediment of a slightly lighter colour than that of the mudstone. The broad, upper part of the digit This section is through the claw imprint of digit III impression has formed an undertrack in the layers (Fig. 7A). The claw has left a clear, narrow cut down directly below it. On the tracking surface a shallow up through the upper part of the slab and the upper layer bulged area of displaced material is present on the has only been dragged a little down by the pressure right side of the digit impression, suggesting a side- from the claw. The cut from the claw protrudes 15 mm ways movement of the foot. down from the tracking surface and is ®lled with sediment of a lighter colour. A shallow v-shaped Track 1, section B undertrack is formed in the layers below the claw mark. This section was cut through the middle of the impression from digit III, and is expressed as a Track 2, section B bowl-shaped depression in the tracking surface (Fig. 6B). The vertical pressure exercised on the sediment by Section B is cut through the middle part of the the digit has caused the formation of a rim of displaced impression of digit III (Fig. 7B). The digit has not sediment on either side of the digit, most pronounced penetrated the surface layer, which has been depressed on the right side of the digit. The undertrack is evident below the digit. The material laterally displaced by the in the layer 1 cm below the tracking surface. It appears pressure of the digit, has caused an upward bulging of shallower and wider than the true track. The rim of the sediment on both sides of the digit impression. material displaced by the sideways pressure of the digit The upward bulging is more prominent at the outer can also be recognized in the undertrack, although less (left) side of the digit impression, suggesting an pronounced. outwards movement of the digit, during the stride. The bottom of the digit imprint is covered with later Track 1, section C ®lled sediment having a lighter colour. The upward bulging of the sediment is visible in the undertrack at The bottom of the track shows the initial division 20 mm depth. The structure here is shallower and between digits II and III. The separation between digits wider than at the tracking surface. III and IV are not evident at the bottom of the true track (Fig. 6C). Digit II is partly ®lled with sediment Track 2, section C and so appears very shallow and hardly recognizable at the tracking surface. The undertrack in this section This section is through the imprints of digits II, III and LETHAIA 37 (2004) Dinosaur tracks from the Late Triassic 293

Fig. 7. Vertical sections through track 2, see text for details. Section A is cut through the claw impression from digit III. Arrow marks the faint undertrack formed below the claw impression. Section B is from the middle of digit III's impression. The marginal ridge of up bulged material is most prominent on the left (outward) side of the track, indicated by black arrow. Notice the prominent undertrack at the horizon 2.3 cm below the tracking surface. White arrows indicate the width of undertrack. Section C represents a cut through the impressions of digits II, III and IV. Notice the folding of the prominent up-bulged material between the impressions of digit IV and III, marked by arrows, which indicates a sideways movement of the foot. Section D is cut through the impression of the proximal end of the track. The foot has only caused slight sideways disturbance in the thin surface layer (arrow). All sections in rear view. 294 Jesper MilaÁn et al. LETHAIA 37 (2004)

IV (Fig. 7C). The track is deepest at digit III, which is and anatomical details, like digital pads, are recogniz- 9 mm deep measured from the original tracking able in two of the digits. The state of preservation of surface. The bottoms of the digit imprints are this track is comparable to the one described by Gatesy disturbed and ®lled with sediment of a lighter colour. et al. (1999, ®g. 1b). The upward sediment bulge between the digit impres- At the surface, the differences in preservation of the sions, most pronounced between digits IV and III, and two tracks make direct comparisons dif®cult, but is asymmetric, a feature that is also clearly evident in when viewed in cross-section, a number of similarities the undertracks, although the individual digit impres- become apparent. The narrow cut from the claw of sions become less wide and shallow downward. The digit III shows that the trackmaker in both cases deformation structures around and below the track all possessed sharp laterally compressed claws. The indicate an outward movement of the foot during the apparent shortness of the digits in track 1 is shown stride. to be an artefact of erosion, as the undertracks reveals that the division of the digits occurs well before this Track 2, section D can be recognized in the track at the surface. In both tracks the undertracks are well-de®ned, although A section through the impression of the proximal pad wider and shallower than the true tracks. Undertracks of digit IV represents the metatarsal area of the track are recognizable at horizons down to 2 cm depth. (Fig. 7D). The imprint is here very shallow and has The sectioning of the two tracks further reveals that only disturbed the upper 2 mm layer of the surface. parts of the tracks still retain some of the sedimentary The pressure from this part of the foot has been too ®lling, which blurs the shape and anatomical details, light to initiate the formation of undertracks. A and in the case of track 1, combined with erosion gives shallow rim of laterally displaced material is present the true track the appearance of an undertrack. on both sides of the impression, most pronounced on The asymmetry of the two footprints with the the outside of the track. marginal rim of displaced material on the outer side of the track, in fact displays many similarities with asymmetrical deformation of tracks described in Discussion Lower Jurassic Grallator tracks from Italy (Avanzini 1998). However, in that case the asymmetry occurred At the surface, the two tracks exhibit very different in the opposite side of the track. Avanzini (1998) used states of preservation. Owing to the shallow appear- this deformation to infer information about the ance and the vaguely de®ned rounded shape of the walking dynamics of theropod dinosaurs and incor- digit impression, track 1 was initially identi®ed as an porated a slight inwards turn of the foot in his undertrack. However, the presence of sediment ®lling reconstruction of theropod foot movements during a of a slightly different colour in some of the slices and stride. The material studied by Avanzini (1998) especially the deep cut from the claw of digit III proves consisted of vertical sections of six tracks where the that the track is an eroded true track. The presence of sideways deformation structure in the proximal end of undertracks in the layers subjacent to the track cannot the track occurred in two of the tracks (Marco solely be used to argue that the track is an undertrack, Avanzini, personal communication 2003). The side- since undertracks form at several horizons subjacent ways deformation structures in the two tracks herein to the tracking surface (MilaÁn & Bromley in press), so described is similar to, but occurs on the opposite side even if the track at the apparent surface is an under- of the tracks described by Avanzini (1998). This shows track, then it would still be possible to ®nd under- a clear difference in the foot movements of the tracks at the subjacent horizons. dinosaurs responsible for the tracks in Greenland The rounded and unde®ned appearance of the track and those in Italy. Although the tracks belong to the at the surface is the result of subaerial erosion of the same ichnogenus (Grallator), the temporal gap track, and the fact that parts of the track still retain the between the Upper Triassic, (Late Norian to Early sedimentary ®lling. Whether the erosion of the track Rheatian) beds from Greenland to the Lower Jurassic occurred before burial, or has happened after the track beds from Italy, suggest different genera of trackbearing layers were re-exposed at the surface recently, makers in the two cases. is not known. The presence of desiccation cracks in the This study demonstrates that evidence obtained surface around the track witnesses a certain time of from ichnology is very important in the study of subaerial exposure of the tracking surface prior to dinosaur walking dynamics; in this case it suggests a burial. change in walking dynamics across the Triassic- Track 2 exhibits at the surface a much better de®ned Jurassic border. However, further sectioning of tracks appearance than track 1, as the digits are well-de®ned from different localities and geological stages should LETHAIA 37 (2004) Dinosaur tracks from the Late Triassic 295 be conducted to examine the full range of walking probably during the kick-off. This is in contrast to the dynamics of dinosaurs, before any ®rm conclusions pattern described from Lower Jurassic theropods. can be drawn. A track exposed in cross-section can express Sideways deformation structures in tracks can also different morphologies, of which some are hardly occur when an animal changes its direction of recognizable as tracks. This is important to bear in progression (Brown 1999) or if the tracking surface mind when studying exposures of continental depos- is sloping. Such things need to be taken into con- its, as horizons with apparent disturbance of the layers sideration when deformation structures around tracks may be the result of trampling by vertebrates. are interpreted. The marginal ridge formed on the right side of the Acknowledgements. ± We thank Hanne Lamberts, GEUS, Copen- proximal part of track 1 (Fig. 6D) shows in section a hagen and Hans Jùrgen Hansen, Geological Institute, Copenha- gen, for access to stone cutting and polishing equipment. A deep disturbed zone undercutting the rim of displaced special thank you to Richard G. Bromley for his critical reading material. A possible origin of the structure is that it is a of the early drafts of the manuscript and to Christian Meyer and fault formed during the formation of the raised pad of Mike Romano, whose reviews and helpful comments helped clarify critical points in the manuscript. Marco Avanzini provided displaced material caused by a sideways and down- helpful comments and discussions. The Carlsberg Foundation wards movement of the foot during the stride. That ®nancially supported the ®eldwork in East Greenland. microfaulting occurs in sediments in connection with track formation has been established experimentally by Allen (1997) and Jackson (2002). However, the uniform bending of the layers down into the structure (Fig. 6C) does not correspond with what would be References expected of a fault. Instead the architecture suggests Avanzini, M. 1998: Anatomy of a footprint: Bioturbation as a key to that the structure is caused by sediment ¯ow and is in understanding dinosaur walk dynamics. Ichnos 6, 129±139. Alexander, R. McN. 1976: Estimates of speeds of dinosaurs. Nature fact similar to the structures termed `Cave' and `Cave- 261, 129±130. in' by Brown (1999). Cave is the situation where a Allen, J.R.L. 1997: Subfossil mammalian tracks (Flandrian) in the slight turn of the trackmaker's foot during the stride Severn Estuary, S.W. Britain: mechanics of formation, preservation and distribution. Philosophical Transactions of the Royal causes the foot to slightly undermine the track wall, Society London, B 352, 481±518. leaving parts of the trackwall as an overhang, whereas Bromley, R.G. & Asgaard, U. 1979: Triassic freshwater ichnocoe- cave-in is the situation where the overhang afterwards nosis from Carlsberg Fjord, East Greenland. Palaeogeography, Palaeoclimatology, Palaeoecology 28, 39±80. collapses over the cave. Brown Jr., T. 1999: The Science and Art of Tracking. 219 pp. Berkley Books, New York. Christiansen, P. 1997: Hindlimbs and feet. In Currie, P.J. & Padian, K. (eds): Encyclopedia of Dinosaurs, 320±328. Academic Press. Clemmensen, L.B., Kent, D.V. & Jenkins Jr., F.A. 1998: A Late Conclusion Triassic lake system in East Greenland: facies, depositional cycles and palaeoclimate. Palaeogeography, Palaeoclimatology, Palaeo- ecology 140, 135±159. Although in its infancy, the method of studying Farlow, J.O., Gatesy, S.M., Holtz Jr., T.R., Hutchinson, J.R. & vertebrate tracks in vertical section has proved to be Robinson, J.M. 2000: Theropod Locomotion. American Zoologist indeed very useful, as information crucial to the 40, 640±663. FornoÂs, J.J., Bromley, R.G., Clemmensen, L.B. & Rodriguez-Perea, correct interpretation of both the trackmaker and A. 2002: Tracks and trackways of Myotragus balearicus Bate the substrate consistency at the time of trackmaking is (Atiodactyla, Caprinae) in Pleistocene aeolianites from Mallorca concealed below the tracking surface. (Balearic Islands, Western Mediterranean). Palaeogeography, Palaeoclimatology, Palaeoecology 180, 277±313. The initial identi®cation of track 1 as an undertrack, Gatesy, S.M. 2001: Skin impressions of Triassic theropods as based on the shallow, rounded and poorly-de®ned records of foot movement. Bulletin of the Museum of Com- appearance of the track was shown to be incorrect, as parative Zoology 156, 137±149. Gatesy, S.M., Middleton, K.M., Jenkins Jr., F.A. & Shubin, N.H. the vertical sections through the track demonstrated 1999: Three-dimensional preservation of foot movements in that the undertrack-like appearance was an artefact of Triassic theropod dinosaurs. Nature 399, 141±144. subaerial erosion, and that `real' undertracks were Hitchcock, E. 1858: Ichnology of New England, A report on the Sandstone of the Connecticut Valley Especially its Fossil Footmarks. present in the layers subjacent to the track. 220 pp. W. White, Boston (reprinted by Arno Press in the In both tracks the narrowness of the claw incisions Natural Science in America Series). observed in the sections show that the theropods Irby, G.V. 1995: Posterolateral markings on dinosaur tracks, Cameron Dinosaur Tracksite, Lower Jurassic Moenave Forma- responsible for the tracks had laterally compressed, tion, northeastern Arizona. Journal of Paleontology 69, 779±784. sharp claws. The pronounced outward rotation of the Jackson, S. 2002: How to make dinosaur tracks: interpreting foot observed in both tracks, show that Upper Triassic dinosaur footprint formation and preservation using laboratory controlled simulations. The Palaeontological Association News- theropods had a walking pattern, which included an letter 51, 81. outward movement of the proximal part of the foot, Jenkins Jr., F. A., Gatesy, S.M., Shubin, N.H. & Amaral, W.W. 1997: 296 Jesper MilaÁn et al. LETHAIA 37 (2004)

Haramiyids and Triassic mammalian evolution. Nature 385, D. & Upchurch, P. (eds): Symposium of Vertebrate Palaeontology 715±718. and Comparative Anatomy, SVPCA 50, Abstract Volume. 11±13 Jenkins Jr., F. A., Shubin, N.H., Amaral, W.W., Gatesy, S.M., September 2002, Cambridge. Schaff, C.R., Clemmensen, L.B., Downs, W.R., Davidson, A.R., MilaÁn, J. & Bromley, R.G. 2003: How to distinguish between true Bonde, N. & Osbñck, F. 1994: Late Triassic continental tracks and undertracks ± experimental work with arti®cial sub- vertebrates and depositional environments of the Fleming Fjord strates. 1st EAVP (European Association of Vertebrate Palaeontol- Formation, Jameson Land, East Greenland. Meddelelser om ogists) Meeting, 15th±19th of July, Basel, Switzerland. Abstracts of Grùnland, Geoscience 32, 1±25. papers and posters with program,1. Jenkins Jr., F. A., Shubin, N.H., Gatesy, S.M. & Padian, K. 2001: A MilaÁn, J. & Bromley, R.G. (in press): The impact of sediment primitive pterosaur (Pterosauria: Eudimorphodontidae) from consistency on track- and undertrack morphology: experiments the Greenlandic Triassic. Bulletin of the Museum of Comparative Zoology 156, 151±170. with emu tracks in layered cement. In Rainforth, E.C. & McCrea, Kent, D.V. & Clemmensen, L.B. 1996: Paleomagnetism and cycle R.T. (eds): Fossil Footprints. Indiana University Press. stratigraphy of the Triassic Fleming Fjord and Gipsdalen Olsen, P.E., Smith, J.B. & McDonald, N.G. 1998: Type material of Formations of East Greenland. Bulletin of the Geological Society the type species of the classic theropod footprint genera Denmark 42, 121±136. Eubrontes, Anchisauripus and Grallator (Early Jurassic, Hartford Lockley, M. 1991: Tracking Dinosaurs. 238 pp. Cambridge Uni- and Deer Basins, Connecticut and Massachutts, USA). Journal of versity Press, Cambridge. Vertebrate Paleontology 18, 586±601. MilaÁn, J. & Bromley, R.G. 2002: One foot ± many tracks; experi- Thulborn, T. 1990: Dinosaur Tracks. 410 pp. Chapman and Hall, ments with undertrack formation in layered cement. In Norman, London. Jesper Milàn – Vertebrate Ichnology Appendix ______

Paper 3 Theropod foot movement recorded from Late Triassic, Early Jurassic and Late Jurassic fossil footprints. Jesper Milàn, Marco Avanzini, Lars B. Clemmensen, Jose Carlos Garciá-Ramos & Laura Piñuela In Harris et al. The Triassic-Jurassic Terrestrial Transition. New Mexico Museum of Natural History and Science Bulletin, 2006, v. 37, p 352–364

Harris et al., eds., 2006, The Triassic-Jurassic Terrestrial Transition. New Mexico Museum of Natural History and Science Bulletin 37. 352 THEROPOD FOOT MOVEMENT RECORDED BY LATE TRIASSIC, EARLY JURASSIC AND LATE JURASSIC FOSSIL FOOTPRINTS

JESPER MILÀN1, MARCO AVANZINI2, LARS B. CLEMMENSEN3, JOSE CARLOS GARCIÁ-RAMOS4 AND LAURA PIÑUELA4

1Geological Institute, University of Copenhagen, Øster Voldgade 10, DK-1350, Copenhagen K, Denmark, E-mail: [email protected]; 2Museo Tridentino di Scienze Naturali, Via Calepina 14, I-38100 Trento, Italy, E-mail: [email protected]; 3Geological Institute, University of Copenhagen, Øster Voldgade 10, DK-1350, Copenhagen K, Denmark, E-mail: [email protected]; 4Museo del Jurásico de Asturias [MUJA], Colunga, Asturias, Spain - Departamento de Geología Universidad de Oviedo, C/Jesús Arias de Velasco, s/n 33005 Oviedo, Asturias, Spain, E-mail: jcgramos.mujagmail.com and [email protected]

Abstract—Vertebrate tracks reveal more information than just the shape and anatomy of a track maker’s foot. By studying the deformation of sediment, around and subjacent to tracks, important information about a track maker’s walking kinematics can be obtained. Thirteen theropod tracks spanning the Late Triassic, Early Jurassic and Late Jurassic were examined for deformation induced by foot movements during a stride. The tracks from the Late Triassic and Early Jurassic are preserved as true tracks except one that is preserved as a natural cast. The tracks show two, distinctly different types of deformation: (1) an outward twist of digit III formed during the kick-off, and (2) outward rotation of the proximal parts of the foot, causing the basal part of digit IV to dig down and outward into the sediment. Late Jurassic tracks preserved as deep natural casts show an even, outward deformation of the whole track. All the studied Late Triassic and Early Jurassic tracks and undertracks except one show the same outward deformation, but the tracks from the Late Jurassic are impressed into the sediment in a slightly different way. The base of the track under digits II and IV slopes sharply outward in the Late Jurassic footprints. This seems to demonstrate that different theropods adopted different walking strategies at different times.

INTRODUCTION cal pads should contact the substrate with all toes simultaneously, sink vertically, and then extract from the substrate vertically with digit III the A vertebrate track is a complex structure resulting from dynamic last to exit. Such un-deformed, “ideal” footprints have been experimen- contact between a track maker and the sediment upon which it treads. tally obtained by laboratory experiments with model feet, or severed Not only the actual tracking surface (sensu Fornós et al., 2002) is subject animal feet, emplaced in artificial substrates (Allen, 1997; Manning, to deformation, but also the subjacent horizons are deformed as the 2004; Milàn and Bromley, 2006), and they all show a uniform deforma- pressure from the track maker’s foot are transferred radially outward tion in the sediments in and around the tracks. When viewed in vertical into the sediment (Allen, 1997; Gatesy, 2003). In layered sediments, this sections, the undertracks formed along the subjacent horizons in the causes the formation of a stacked succession of undertracks that gradu- artificial substrates are symmetrically developed radiating outward and ally becomes wider, shallower and less detailed downward (Milàn and downward from the tracks (Allen, 1997; Milàn and Bromley, 2006). The Bromley, 2006). herein described footprints all deviate from these “ideal,” undeformed Following a simplified model proposed by Thulborn and Wade tracks in showing various degrees of lateral deformation, caused by foot (1989) and Avanzini (1998), the contact between an animal’s foot and the movements occurring during the time the track maker’s foot is in contact sediment during walk can be divided into three distinct phases. The with the substrate. A similar complex deformation pattern has been re- touch-down phase is when the foot is moved forward and emplaced on ported by Gatesy et al. (1999) and Gatesy (2001, 2003) in deep theropod the sediment surface. This is followed by the weight-bearing phase, footprints from East Greenland. Evidences of foot and toe movement are when the animal’s center of gravity passes over the animal’s foot, which recognizable with high fidelity in well preserved true track surface and, if is consequently pressed into the substrate, forming the track. Last is the the true track is obscured by overlying sediments, sectioning can reveal kick-off phase, when the weight is transferred to the distal parts of the the true contours of the surface (e.g., Loope, 1986). Generally, however, digits as the body moves forward and the foot subsequently is lifted and in the fossil record, theropod dinosaur footprints, being shallow impres- swung toward a new touch-down phase. As evidenced by Gatesy et al. sions of the plantar surface, provide little evidence of the details of limb (1999) and Gatesy (2001, 2003) however, these phases are more com- excursion. For this reason the analysis of the subsurface deformation plex and represent a continuum of interactions between foot and sub- seem to be useful in identifying dinosaur foot movements in shallow and/ strate. In fact, a ground-based reaction force (GRF) perspective (Roberts or eroded footprints. and Scales, 2002) makes it clear that there are no “distinct” phases of the Avanzini (1998) examined dinosaur foot motion by sectioning a stance phase, but a gradual shift from a low magnitude force pointing shallowly impressed Lower Jurassic theropod track vertically, and de- backward and up (decelerating the animal) to a peak magnitude force as it scribed an apparent inward twist of digit II and the proximal part of the becomes more vertical, to a forward and upward force reaccelerating the foot, relative to the midline of the trackway, made during the weight- body. The foot bears weight throughout the contact phase, not just bearing phase of the stride. A similar study of Late Triassic theropod during the middle portion. Similarly, in a dynamically stable walker, the footprints and undertracks in vertical sections revealed an outward de- body can pass lateral or medially to the point of ground contact as it formation of all the studied specimens (Milàn et al., 2004). advances throughout the stance phase. As a footprint is the result of the The aim of this study is to compare and expand on the studies of dynamic contact between the track maker and the substrate, any simul- Avanzini (1998) and Milàn et al. (2004) and describe the lateral deforma- taneous movement of the foot will be captured in the sediment and tion of footprints resulting from walking kinematics as observed in arti- subsequently be recognizable as a zone of disturbance within or around ficial (1-10) and natural (11-13) sectioned theropod footprints from the the track (Brown, 1999). Late Triassic of Jameson Land, East Greenland, Early Jurassic of Lavini To create a perfect, undisturbed footprint, a foot with symmetri- di Marco, northern Italy, and Late Jurassic of Asturias, Spain. Institu- 353 tional abbreviations: MGUH, Geological Museum, University of permost part of the Asturian Jurassic sedimentary succession. It repre- Copenhagen; MTSN, Museo Tridentino Science Naturali; MUJA, Museo sents deltaic sediments that accumulated along the coast of an inland sea del Jurásico de Asturias. separated from the open ocean by a tectonic threshold that served as protection against storms at (García-Ramos and Gutiérrez Claverol, 1995; MATERIAL García-Ramos et al., 2004). The dinosaur footprints are mainly found in Late Triassic Tracks, Jameson Land, East Greenland sand bar deposits; these deposits formed during crevasse splays along distributary channels in the fluvially-dominated deltaic complex. The Five single footprints, all assigned to ichnogenus Grallator Lastres Formation has also yielded a high number of reptile footprints Hitchcock, 1858, collected from the Upper Triassic Fleming Fjord For- including those of pterosaurs, crocodylians, turtles, and lizards (Piñuela, mation, Jameson Land, East Greenland, were studied. The specimens are 2000; Lires, 2000; Garcia Ramos et al., 2002; Avanzini et al., 2005). stored at the Geological Museum, University of Copenhagen. The Up- per Triassic Fleming Fjord Formation consists of a well-exposed 200- TERMINOLOGY 300 meter thick succession of lake deposits. The track-bearing upper The dinosaur that makes an actual footprint is termed the “track part of the formation, the Carlsberg Fjord beds of the Ørsted Dal Mem- maker” and a sediment surface on which a track is emplaced is termed the ber is 80-115 m thick and is composed of structureless red mudstones “tracking surface,” sensu Fornós et al. (2002). The track emplaced on the rhythmically broken by thin, greyish siltstones (Clemmensen et al., 1998). actual tracking surface is termed the “true track,” or just “track,” and The thin siltstone beds represent episodes where the mudflats were several consecutive tracks from the same track maker constitute a “track- flooded by lake water of a depth sufficient to allow formation of small way” (Leonardi, 1987; Lockley, 1991). The sediment infilling the track wave ripples. The flooding was followed by periods of subaerial expo- forms a “natural cast” of the track (Lockley, 1991). If the track maker’s sure and desiccation. The theropod tracks are all found in these siltstone foot sinks to a considerable depth in the sediment, then the vertical parts, beds in association with rare footprints of larger prosauropods (Jenkins between the bottom of the track and the tracking surface are termed the et al., 1994; Clemmensen et al., 1998; Gatesy et al., 1999; Gatesy, 2001). “track wall” (Brown, 1999). The weight of the track maker’s foot is The underside of the slabs containing some of the studied tracks are transferred radially outward into the sediment around and below the covered with numerous small arthropod trackways (Diplichnites) pre- track maker’s foot (Allen, 1997), causing the formation of “undertracks” served in hyporelief, and examples of the arthropod resting trace (sensu Lockley, 1991) along the subjacent horizons. Undertracks differ Rusophycus, indicating the presence of a fluvial and possibly lacustrine from true tracks in being wider, shallower, and less detailed downward environment before the track-bearing layer was deposited (Bromley and (Milàn and Bromley, 2006). Asgaard, 1979). Theropod dinosaurs were functionally tridactyl and their tracks Early Jurassic, Lavini di Marco, Northern Italy consist mainly of the impressions of digits II, III and IV. Digit I, the hallux, is situated at an elevated position on the metatarsus (Christiansen, Three single theropod tracks, two assigned to the ichnogenus 1997), and impressions of digit I mostly occur in tracks emplaced in deep Kayentapus Welles, 1971 and one to the ichnogenus Grallator Hitchcock, mud (Gatesy et al., 1999). The phalangeal skeleton inside the digits 1858, collected from the Lower Jurassic Monte Zugna Formation (Calcari contains 3, 4 and 5 phalanges respectively for digits II, III and IV. The Grigi Group), Lavini di Marco, northern Italy and currently stored in the digits are covered by fleshy digital pads, which in most cases are situated Museo Tridentino Scienze Naturali (Trento, Italy), were studied. The around the phalangeal joints, and these are recognizable as individual pad Calcari Grigi Group comprises shallow water marine deposits and formed impressions in well preserved tracks. The proximal pad impressions of on an extensive carbonate platform (Trento Platform). The lowermost digit IV in theropod tracks extend further back than those of digits III and part of the Monte Zugna Formation is composed of alternating, subtidal II, giving the tracks a typically asymmetric “heel” area. The claw impres- limestones in meter-thick beds and thin intervals of green clay without sion of digit III is, in many cases, directed inward toward the midline of evidence of tidal influence. Peritidal cycles with structures typical of the trackway (Lockley, 1991; Farlow et al., 2000). This is an important tidal deposition characterize the upper part of the formation; an abun- factor in identifying single footprints as either from a right or left foot; dant and diverse dinosaurian ichnofauna was found in the supratidal these factors can be recognizable in even badly eroded tracks. deposits (Leonardi and Mietto, 2000). The subtidal facies, representing the lower two-thirds of the formation, is mainly made up of homoge- METHODS neous micritic, peloidal-fossiliferous limestones completely reworked Eight of the tracks used for this study were sectioned vertically, by marine bioturbation. Sometimes accumulations of large bivalves (i.e., perpendicular to the length axis of the impression of digit III, using a Gervillia buchii) are present. The supratidal facies are characterized by stationary rock saw with a 3 mm-wide width. The surface of each slice sedimentary structures such as stromatolites, birdseyes, sheet cracks, was subsequently polished to enhance the view of the internal structures mud cracks, teepees, flat pebble breccias, etc. (Masetti, 2000). The algal of the sections. The sections used for the study were subsequently and palynological association of the Monte Zugna Formation suggests a scanned directly on a flatbed scanner at 600 dpi. To protect the glass Hettangian-Early Sinemurian age. The ichnotaxa identified are Grallator, surface of the scanner, a sheet of transparent plastic was placed between Eubrontes and Kayentapus, all pertaining to theropod dinosaurs (Piubelli the rock slice and the glass, and ethanol was used as a contact medium et al., 2005), and Anomoepus, probably made by primitive ornithopods between the rock and the plastic sheet. Digital color and contrast-en- (Avanzini et al., 2001a). There are several tracks that are attributed to hancing was subsequently performed using Corel Photo-paint 11. Other very early sauropods (i.e., ichnogenera Parabrontopodus and Lavinipes) sections were polished and accurate interpretative drawings made. (Avanzini et al., 2004) and some poorly preserved tracks attributed to prosauropods and possibly thyreophoreans (Leonardi and Mietto, 2000; DESCRIPTION OF TRACKS Avanzini et al., 2001b). Triassic Tracks Late Jurassic of Asturias, Northern Spain Track 1 Three single theropod footprints collected from the Upper Juras- sic Lastres Formation, Asturias, northern Spain and currently stored in This theropod track is relatively small, 14.4 cm long and 12.3 cm the Museo del Jurásico de Asturias (Colunga, Spain) are here described. wide (Fig. 1). The track is preserved as a natural cast of a left foot on the The Lastres Formation is Kimmeridgian in age and constitutes the up- underside of a reddish brown mudstone. The cast of the track is well- 354

FIGURE 2. Partial Late Triassic track of the ichnogenus Grallator (MGUH 27812) preserved as a true track in a reddish brown mudstone. The track is from a left foot and is well-preserved, with recognizable impressions of the individual digital pads and claws. Notice the prominent raised rim of displaced material on the outsides of digit II and III (arrows), resulting from an outward rotation of the digits during the stride. footprint length of 16.0 cm is estimated. A shallow but well-defined undertrack containing all three digits is present on the underside of the slab; it represents a horizon 13 mm below the tracking surface. No claw impressions are present in the undertrack, and what appears to be the middle digital pad of each digit has left the deepest impression in the FIGURE 1. A, Small, well-preserved, Late Triassic theropod track (MGUH undertrack; it is also evident that the middle digital pad in digit III was the 27811) preserved as a natural cast on the underside of reddish brown most deeply impressed in the track. A prominent raised rim of displaced mudstone. The track is from a left foot and belongs to the ichnogenus material is present along the outer sides of the impressions of digits II Grallator. B, When the track is seen in frontal view, the middle part of the cast of digit III shows a prominent outward deformation, indicated by arrow. and III, indicating an outward movement of the foot relative to the walking direction (Fig. 2). preserved, with clear impressions of the individual digital pads and claws (Fig. 1A). When studied in cranial view at a low angle, the cast of the Tracks 3 and 4 impression of digit III is asymmetrical because the bottom of the digit These theropod tracks are preserved in a reddish-brown mudrock impression is offset outward relative to the walking direction (Fig. 1B). and comprise one complete (track 3) and one partial (track 4) track (Fig. 3). Both tracks were selected for vertical sectioning because the tracks Track 2 were preserved with the original sedimentary infilling in place (Fig. 3). This theropod track is a partial imprint, consisting of the distal Furthermore, the infilling of the tracks is of a darker color than the parts of digits II, III and IV preserved on a slab of finely laminated, red sediment of the tracking surface, which makes it ideal for studying struc- and yellow-brown mudstone 130 mm thick in the part containing the tures in cross section. The complete track is a right pes, 17.3 cm long and digit impressions (Fig. 2). Although incomplete, the track can be identi- 8.7 cm wide between the terminations of digits II and IV. The incomplete fied as the imprint of a left pes because the claw of digit III is offset to the track comprises the distal ends of digits II and III of a left pes. The track right of the long axis of the digit. Only the distal digital pad and a small, is identified as a left pes imprint due to the inward orientation of the claw faint claw impression of digit II are present. The impression of digit III imprint of digit III (Lockley, 1991). Like the other track on the slab, this has three well-defined digital pads, the distal one bearing the clear claw one also preserves an infill of darker material. The orientation of the two impression. Only the medial part of the impression of digit II is present. tracks on the slab is toward each other at a high angle, measured along the Parts of neither the metatarsal pad nor the proximal end of the digits are long axis of the impression of digit III. This makes it unlikely that the present in the slab, but by extrapolating the length of the digits, a total tracks represent the right and left foot of the same animal. 355

FIGURE 3. A, Slab of reddish brown, Upper Triassic mudstone containing one complete and one partial theropod track of ichnogenus Grallator (MGUH 27813). The tracks are partially infilled with a darker colored mudstone. B, Interpretative drawing of the slab with the two tracks. The complete track is from a right foot and the incomplete track from a left foot, identified by the inward rotation of the claw of digit III. The darker colored sedimentary infilling of the complete track is encircled by dotted lines. The discontinuous lines A-E represent the studied vertical sections in Figure 4. Four sections were cut through the complete track, all perpendicular to the long axis of digit III (Fig. 3B). The first section, which runs through the tip of the claw impression of digit III, reveals that the im- FIGURE 4. Vertical sections through the tracks from Figure 3 (MGUH pression of the claw has cut down through the layers and left a v-shaped 27813). White arrows indicate direction of deformation. A, Section through impression filled with sediment of a darker color (Fig. 4A). In the next the tip of the impressions of digit III. The tip of the claw has left a v-shaped section, cut through the middle of the impression of digit III, the impres- cut down through the layers (black arrow). B, Section through the middle of sion appears as a steep-walled, clearly defined, rectangular depression in digit III. The deep impression of the digit is filled with sediment of a darker the sediment. The digit impression is clearly asymmetrically developed color. Notice the prominent outward orientation of the section through the because the bottom of the impression is offset outward relative to the digit impression. A vertical Skolithos burrow is indicated by an S. C, Sections tracking surface (Fig. 4B). In the third section, cut through the impres- through the impressions of digits II, III, and IV. The digit impressions are sions of digits II, III, and IV, the impressions of digits II and III are deep and filled with sediment of a darker color. The impression of digit IV disturbed and not so well-defined, but the impression of digit IV is shows an initial outward and downward deformation. D, Section through the clearly defined and is, toward the bottom, strongly offset outward, cre- proximal part of the track. Notice the deep downward and outward ating a zone of brecciated subsediment on the outside of the track (Fig. deformation structure created by the impression of digit IV. E, Section 4C). The last section through the complete track, cut through the proxi- through the middle of the impression of digit III from the incomplete track. mal part of the track, shows a weak impression of the proximal end of The sections show the digit impression to be outwardly deformed. digit II. The impression of digit IV is well defined and strongly developed cm below the tracking surface, indicating a pronounced outward move- downward and outward, undercutting the outer side of the track (Fig. ment of the digit during its time in contact with the ground (Fig. 5C). In 4D). The section through the partially preserved track, cut through the a section through digits II, III, and IV, the sediment between digits IV and middle of the impression of digit III, bears evidence of a slight outward III is deformed sideways in an overturned structure (Fig. 5D). A section movement that disturbed the subsurface layers on the outside of the digit through the proximal area of the track shows little more than a slight impression (Fig. 4E). sideways disturbance of the most surficial layers, which has been scraped to the side and crumbled together (Fig. 5E). For further details, see Milàn Tracks 5 and 6 et al. (2004). These two theropod tracks (here termed tracks 5 and 6) were Track 6 is the imprint of a right pes approximately 23 cm long and examined in cross section to reveal the characteristics of undertracks and 14 cm wide preserved on a slab of reddish-brown mudstone and (Fig. 6). other subsurface deformation structures (see also Milàn et al., 2004). Erosion of the surface has erased all finer morphological details of the Track 5 is a left track; it is 21.3 cm long and 14.9 cm wide and is track; only the outline and general shape of the track remain (Fig. 6A). preserved in a greyish mudstone (Fig. 5). The track is relatively shal- Four sections were made through the track. The first section was cut lowly impressed with well-defined impressions of each digit. Claw marks through the tip of the claw impression of digit III where the slender, and impressions of each digital pad are present in the impression of digit sharp ungual has left a clear, vertical cut down through the layers that is III. A prominent rim of displaced material is present along the outer side infilled with a lighter colored matrix. At the tracking surface, the cut mark of digit III (Fig. 5A). Four sections were made through the track. A starts at a bowl-shaped depression. Also at the tracking surface, a shal- section cut through the claw shows a sharp, slightly asymmetrical low, up-bulged area of displaced material is present (Fig. 6B). The sec- downcut through the layers (Fig. 5B). A section through the middle of tion cut through the middle of the impression from digit III displays the digit III shows the development of a prominent, outwardly directed digit impression as a bowl-shaped impression in the tracking surface. deformation structure recognizable down into the undertracks as far as 2 The weight of the digit has caused the formation of small raised rims of 356

FIGURE 6. Poorly preserved, Late Triassic theropod track (Grallator) FIGURE 5. Sections through a Late Triassic theropod track (Grallator) (MGUH 27815) on a slab of reddish brown mudrock. A, The track is somewhat (MGUH 27814), emplaced in a greyish mudstone. A, The track is from a eroded but can be identified as the impression of a right foot. B, Section cut left foot and has well preserved impressions of digital pads and claws. B, through the tip of the impression of the claw of digit III. The deep cut from Section through the tip of the impression of digit III. The impression is the claw is infilled with sediment of a lighter color, indicated by the arrow. preserved as a clear cut through the layers, indicated by the arrow. C, C, Section through the middle of the impression of digit III. Raised ridges of Section through the middle of the impression of digit III, which has created deformed material are present on both sides of the digit impression, but a raised rim of displaced material on the outside of the digit impression. This most prominently toward the outside of the track, indicated by arrows. D, asymmetrical deformation is also prominent in the undertracks along the Section through the impressions of digits II, III, and IV. The divisions of the subjacent horizons. Arrows indicate the sideways-deformed layers. D, Section digits are not visible in the base of the track due to erosion, but the undertracks through the impressions of digits II, III, and IV. Especially the impression of in the subjacent layers show the division of the digits clearly (vertical digit III shows an outward deformation both in the tracking surface and the arrows). The impression of digit IV shows a downward and outward undertracks, indicated by arrows. E, Section through the proximal part of deformation of the subjacent layers, indicated by arrow. E, The outward the track. The foot has created only a slight disturbance in the tracking twist of the proximal end of digit IV has created a prominent outward surface, most prominently developed toward the outside of the track, indicated deformation in the tracking surface as well as a subsediment deformation by the arrow. Figure modified after Milàn et al. (2004). below the tracking surface, indicated by arrows. Figure modified after Milàn displaced material on both sides of the digit. The rim on the outer side of et al. (2004). the digit, however, is more developed than the rim on the inner side of the digit because more pressure was exerted on the outer side of the digit The section through the impressions of digits II, III, and IV shows during its impression in the sediment. The rim is also visible in undertracks the initial division between the digits preserved as small, up-bulged areas formed approximately 1 cm below the tracking surface, again most promi- on the horizon with undertracks approximately 1 cm below the bottom nent on the outer side of the digit impression (Fig. 6C). of the track. A prominent deformation structure projects downward and 357 outward from the impression of digit IV and disturbs and folds the layers as much as 3 cm below the tracking surface (Fig. 6D). The last section, cut through the proximal part of the track, appears as a shallow, flat- bottomed depression in the tracking surface, revealing no anatomical information. However, a prominent, raised, marginal rim of displaced surface material is present at the right (outward) side of the footprint. The lateral disturbance of the layers is evident down to 3 cm below the tracking surface (Fig. 6E). For further details see Milàn et al. (2004). Lower Jurassic Tracks One dinosaur track described in a previous study by Avanzini (1998) was re-examined. Three other, very shallow tracks from the same site were examined in cross-section for the presence of subsurface deformation structures.

Track 7 The first track (Fig. 7) is a medium-sized incomplete theropod track, 19.5 cm long and 19.0 cm wide. The track, a right foot in concave epirelief, is preserved on the surface of a whitish stromatolitic bindstone. The track is incomplete, with only the impressions of the three digits well-preserved. The general features and the relatively high divarications of digits II-IV cautiously suggest attribution to the ichnogenus Kayentapus. The deformation of the stromatolitic laminite indicates a general outward movement of the foot relative to the trackway midline. In cross section C (Fig. 7), the laminae around and subjacent to the impression of digit III are compressed toward the external side of the track, creating an evident ridge. The outward translocation of the digits is most evident in cross sections D and E (Fig. 7).

Track 8 The second track (Fig. 8) is a medium-sized theropod track 22 cm long and 17 cm wide. The track is preserved as concave epirelief of a right foot on the surface of a whitish, stromatolitic bindstone. The track is complete and well-preserved, with recognizable impressions of digital pads and claws (Fig. 8). The track is tridactyl, mesaxonic, and shows a high angle of divarication between digits II and IV, a higher divarication of digits III and IV than digit II and III, a well defined covering the joint between the metatarsals and phalanges of digit IV, impressions of claws on digits II, III, and IV, and a highly projected middle digit like in other grallatorids. These features are typical of the theropod ichnogenus FIGURE 7. Early Jurassic theropod footprint (?Kayentapus isp.) (MTSN Kayentapus. The impression of digit IV shows a slight inward deforma- 5212) from the Lavini di Marco, northern Italy, tracksite sectioned at 1 cm tion in cross sections F, G, and H. In section G, the impression of digit III intervals. The track is an incompletely preserved imprint of a right foot, is apparently directed toward the inside of the print, while in the more comprising only the impressions of the digits. The sections C, D, and E cut distal section L, it seems to cut vertically down in the substrate. through the impressions of digits II, III, and IV all show an outward deformation of the digit impressions, recognizable in the undertracks as well Track 9 as on the tracking surface. Section F, cut through the distal part of digit III, shows an outward deformation of the digit impression. Arrows in sections A- The third footprint (Fig. 9) is a slender-toed theropod footprint F indicate areas of outward deformation. 19 cm long and 14 cm wide. The track is preserved as concave epirelief on the surface of a whitish stromatolitic bindstone. The track is the imprint section I, through the tip of the impression of digit III, shows a proof a left foot and is complete and well-preserved, with recognizable nounced inward deformation, contrary to the sections through the rest of digital pads and claws. The impression of digit III extends beyond the the track. axis joining digits II and IV. The divarication between digits II and III is slightly lower than that between digits III and IV. The metatarsopha- Track 10 langeal pad of digit IV is evident. All these characteristics are congruent The fourth footprint (Fig. 10) is a small theropod track 14 cm long with those of one of the morphotypes recently recognized at the Lavini and 10 cm wide. The track is the imprint of a left foot preserved as di Marco site and attributed to the ichnogenus Kayentapus (Piubelli et concave epirelief on the surface of a whitish stromatolitic bindstone. The al., 2005). track is poorly preserved but still attributable to a Grallator-like The cross sections through this track show an overall outward ichnogenus. A small depression, preserved at the base of the impression deformation of the track. Section B shows a slightly outward component of digit II is interpreted as the partial trace of digit I (Fig. 10). This track of movement of digit IV and section D, which shows the section through was previously described by Avanzini (1998) as a right footprint due to the impression of digit II, has preserved a significant outward-directed an incorrect interpretation of a depression (digit I?) at the base of digit II deformation. More evident is the deformation of the stromatolites under as the vestige of the pad covering the metatarsophalangeal joint of digit the impression of digit III. In sections E-H, the outward-directed compo- IV. For this reason, Avanzini’s (1998) interpretation of the movements of nent of external translocation is prominent and well recognizable. Only the track maker based on the deformation of the stromatolite bindstone 358

FIGURE 8. Early Jurassic theropod footprint (Kayentapus isp.) (MTSN 5211) from the Lavini di Marco, northern Italy, tracksite sectioned at 0.5 FIGURE 9. Early Jurassic theropod footprint (Kayentapus isp.) (MTSN cm intervals. The track is a right footprint and is preserved as a true track, 5213) from the Lavini di Marco, northern Italy, tracksite sectioned at 0.5 with well-defined impressions of digital pads and claws. When studied in cm intervals. The track is a left footprint and is preserved as a true track vertical sections, no apparent outward deformation is evident. Only sections still retaining some sedimentary infilling. The sections B to H represent L and G hint at a slight inward deformation of the impression of digit III, sections all through the track; all show prominent outward deformation. indicated by arrows. Section I, cut through the impression of the claw of digit III, exhibits inward deformation of the claw impression. Arrows in sections indicate areas and under the trampled surface was incorrect; we therefore redescribe it in direction of deformation. the present study. The true tracking surface is partly covered in the foot region footprint, and distally the impression of digit III becomes the dominant (post-track surface sensu Gatesy, 2003) by an infill made of dolomitized, agent of deformation in the subjacent layers, deforming the subjacent fine-grained breccia and mud (Avanzini et al., 1997) marked by the gray layers to a greater depth than either digit II or IV in the impact area of the color in Figure 10. The subsurface deformation structures, created by footprint. Deformation of the footprint is especially visible in section H the foot being emplaced on the tracking surface, are recognizable in all where the deflected lamina surrounding the toes III and IV both exhibit a vertical sections cut through the footprint. The proximal sections through pronounced, outward deformation. On the contrary, the section cut the track (A-C) show a deep vertical to sub-vertical penetration of the through the tip of the impression of digit III shows a pronounced inward foot down into the substrate, with a slight outward component in the deformation of the subjacent layers. movement. In the proximal sections of the footprint, the impressions of digits II and IV are the most prominent, indicating that these digits Late Jurassic Tracks carried the bulk of the track maker’s weight, but the pressure seems to have been exerted mostly on digit IV because it has been emplaced diago- Track 11 nally down into the sediment, making the bottom of the footprint appear The first track is a medium-sized theropod track 37 cm long and tilted slightly outward. 28 cm wide (Fig. 11). The track is preserved as a natural cast of a left foot From around the middle of the footprint, the deformation caused and so only the penetration modality of the foot in the substrate can be by the impression of digit III becomes visible in the layers below the recognized as no subsurface structures are preserved. The track is well 359

FIGURE 10. Early Jurassic theropod footprint (Grallator isp.) (MTSN 5210) from the Lavini di Marco, northern Italy, tracksite sectioned at 0.5 cm intervals. The track is a left footprint and contains the weak impression of the posteromedially directed digit I. The true track is covered by an infill FIGURE 11. Late Jurassic theropod footprint (MUJA 1080) from Asturias, made of dolomitized, fine-grained breccia and mud marked by the gray color. Sections C-H through the track show a prominent outward and Spain, preserved as a deep natural cast. A, In frontal view, the foot has been downward deformation caused by digit IV. The impressions of digits II and pressed outward and down through the sediment. The cast of the impression of digit III is the deepest, followed digits II and IV. B, When viewed from III seem not to possess any outward deformation. The tip of the impression of digit III, sections I and L, show the claw impression to be directed inward. below, the cast of the track has the typical inward twist of the of the claw The arrows in all sections indicate areas and directions of deformation. impression of digit III. Arrows indicate direction of deformation. Figure modified from Avanzini (1998). incomplete digit III seems relatively greater than 60% of the footprint length, and the proximal edge of the metatarsophalangeal pad on digit III preserved and deeply impressed to a depth of 13 cm. The track lacks is not anterior to the posterior part of the second phalangeal pad on digit discrete digital pad impressions, but this appears to be a function of IV as it is in the ichnogenus Megalosauripus (Lockley et al., 2000). The preservation and not a primary morphological feature. The impression footprint is relatively broad and wide with an elongate digit III, and of digit III spans about 60% of the total footprint length. The general seems very close to a “Eubrontes-like” morphotype (Lockley, 2000). morphology suggests a cautious attribution to the ichnogenus The cast of the track shows a pronounced outward deformation, espe- Megalosauripus (Lockley et al., 2000). The track is most deeply im- cially in the naturally-occurring cross section through the tip of digit III, pressed on the inside of the track, corresponding to digit II, and the which displays the cast of the digit protruding down from the tracking whole cast of the track is deformed slightly toward the outside of the surface at a 65° angle. The track was emplaced on a semi-lithified, 1-2 cm track. The cast of the impression of digit II shows that the digit pen- thick sandstone layer that has been dragged around the digits during their etrated the ground in a sinuous-shaped way, while digit III was initially impression in the substrate. This layer is preserved draping the inside brought down vertically, and then, at about 4 cm depth, curved outward and bottom of the cast of digit III, while the layer has been cut by a through the sediment at an angle of 80°. reverse fault on the outside of the digit (Fig. 12B). Track 12 Track 13 The second track is a relatively large theropod track 40 cm long The third track is a wide theropod track 30 cm long and 23 cm and 33 cm wide. The track is preserved as the natural cast of a left foot. wide. The track is preserved as a natural cast of a left foot. The track is The track is incomplete in that the tips of the casts of digits III and IV are complete and is identified as an undertrack because the impressions of missing, but it is otherwise well-preserved, with recognizable casts of the individual digits are undefined and broad (see Milàn and Bromley, digital pads and claws (Fig. 12). The phalangeal pad formula is 2, 3, and 2006). The footprint bears an elongate “heel” associated with a small, 4 corresponding to digits II, III and IV respectively. The length of the 360

FIGURE 12. Late Jurassic theropod track (MUJA 1261) from Asturias, Spain. A, View from the front. B, Interpretative drawing of view from front. C, View from below. D, Interpretative drawing of view from below. The track is preserved as a natural cast and is complete except for the distal part of digit III. The cast of the footprint exhibits outward deformation created by the foot being forced down and outward through the sediment. The base of the cast of the track is draped by a 1 cm thick layer of sandstone that was impressed into the sediment when the foot was emplaced. The sandstone layer was semi-consolidated when the foot was impressed into the sediment, as evidenced by the layer following the contour of the digit impression. The impact of the foot has caused the development of a fault in the sand layer on the outside of digit III. The frontal view (A, B) shows how the sand layer, indicated by shading, has been shaped and faulted by the impression of the digit. The view from below (C, D) shows the parts of the footprint that show outward deformation, indicated by arrows. When the track was emplaced, digit III was the most deeply impressed, followed by digits II and IV. Arrows indicate directions of deformation. posteromedially-directed protrusion that is identified as the cast of digit I (Fig. 13). The track lacks discrete digital pad impressions, but its general morphology suggests a tentative referral to the ichnogenus Megalosauripus. Even though the print is preserved as an undertrack, and thus lacks many anatomical details, a pronounced outward deformation is evident in the casts of digits II, III, and IV. The depths of the casts of the digits vary substantially: the cast of digit III is deepest, followed by the cast of digit II: the cast of digit IV is the shallowest. The cast of digit II penetrates the tracking surface sub-vertically and is directed slightly outward. The angle of impression, measured from the cast of FIGURE 13. Late Jurassic theropod track (MUJA 1199) from Asturias, digit III, is a constant 75°, and the cast of digit IV, though impressed in a Spain with a cast of digit I preserved. The track is preserved as a natural cast sinuous-shaped way, is on average directed outward. and represents an undertrack because of its wide and undefined appearance. DISCUSSION A, View from front. B, View from below. The track shows an outward deformation of the whole track, most prominent in the casts of digits III All five Late Triassic tracks, including undertracks described herein, and IV. The cast of digit III protrudes the deepest from the tracking surface, show the same pronounced outward deformation of both the tracks and followed by digits II and IV. Arrows indicate direction of deformation. the surrounding sediment. The tracks were randomly collected from different bedding planes and are of different dimensions, which indicate all isolated tracks whose differences in size, shape and time also exclude that they originated from animals of different sizes and times. This the possibility that these tracks were made by the same species of track excludes the possibility that the tracks pertain to the same animal (al- maker. though the tracks could be from different-sized individuals of the same Undoubtedly, the different tracks described were created by dif- species) or that the outward deformation could be a result of either an ferent theropod taxa. The Late Triassic tracks from Greenland are be- abnormal gait of an individual track maker or a special gait adopted to suit lieved to be have been made by small- to medium-sized ceratosaurian local sedimentary conditions. The Early and Late Jurassic tracks are also theropods, but the only body fossil material found at the locality is a 361 partial, small, indeterminate theropod skeleton (Jenkins et al., 1994). The Early Jurassic and Late Jurassic tracks from Italy and northern Spain could be from either ceratosaurs or more advanced tetanurans or both. One large, incomplete theropod skeleton is known from the Early Jurassic of Italy (Nicosia et al., 2005). The skeleton belongs to one of the oldest-known, large-bodied tetanurans and is still undergoing study; it has been informally named “Saltriosaurus” (nomen nudum) (Dal Sasso, 2001). From the Late Jurassic of Spain, several theropod remains are known (Martínez et al., 2000, 2001; Garcia Ramos et al., 2002; Canudo and Ruiz-Omeñaca, 2004). Most of the finds are incomplete and not diagnostic, which makes it difficult to identify all the taxa that may be represented. The identified taxa include the Ceratosauria and Tetanurae. At first glance, there seem to be little variation in the observed sideways deformation of the tracks and undertracks, but closer inspection reveals slight differences in the mode and geometry of the track and surrounding sediment deformation from the different time periods. The Late Triassic tracks from Greenland (Figs. 1-6) all show a pronounced outward deformation relative to the inferred midline of the trackway. FIGURE 14. Schematic representations of the three different types of The parts of the foot responsible for this deformation differ, however, deformation encountered in the 13 investigated tracks. Types 1 and 2 are between the studied specimens. In track 1, the main deformation is found in the Late Triassic and Early Jurassic tracks; type 3 is from the Late caused by the distal two-thirds of digit III, which also is impressed to a Jurassic tracks. Type 1 deformation comprises a pronounced outward twist greater depth than digits II and IV (Fig. 1). In track 2, most of the of digit III, and less deformation caused by digit IV. Type 2 displays outward deformation is caused by the middle and distal parts of digit III, and less deformation of digit III and a pronounced outward and downward deformation by digit II. As in the previous track, digit III is impressed to a greater caused by the proximal parts of digit IV. Type 3 shows an even, outward depth than digits II and IV (Fig. 2). The sections through these tracks deformation along the whole length of the foot. In all investigated tracks, show in more detail the different forms of deformation that occurred. the claw of digit III showed a more or less inward orientation. Track 8 falls The sections through track 3 show a prominent outward deformation of outside of these three types in showing a slight inward deformation, and track 4 was only represented by a section through the middle of digit III, and digit III, which is also the most deeply impressed digit. In this case, digit could thus be either type 1 or 2. The thickest arrows indicate the areas of IV, especially its proximal part, also caused a pronounced, outward de- strongest deformation. formation of the sediment. The whole foot was dislocated sideways during the stride (Fig. 4), contrary to the two first tracks in which the theropod tracks (Lockley, 2000). most significant deformation was caused by digit III only. Track 4 is only The nature of the outward deformation of the Late Triassic and represented by a section through digit III, but this section also shows a Early Jurassic tracks falls into two distinct groups (Fig. 14). In the group pronounced outward deformation. As the track is incomplete, no further containing tracks 1, 2, 4, 5, 7 and 9, the predominant outward deforma- deformation data from the other digits are obtainable. The two tracks tion has been caused by digit III. The other type of deformation, wit- previously described by Milàn et al. (2004) show two different types of nessed in tracks 4, 6, and 10, shows an outward deformation of the whole sideways deformation. Track 5 shows the main deformation to be caused length of digit III and a stronger, more prominent, outward and down- by digit III, with only a little deformation caused by digits II and IV (Fig. ward deformation along the outside of digit IV that, in some instances, 5). Track 6 shows a similar deformation to that of track 4 in that digit III has caused the outer edge of the digit to dig into the substrate (Figs. 4, 6 caused some sideways deformation, but the basal parts of digit IV also and 10). created a prominent outward deformation (Fig. 6). These two types of deformation result from different phases of Apparently, the observed lateral deformation of the Triassic tracks the walk. The outward twist of digit III occurs during the kick-off phase falls into two different groups: one in which the deformation is caused by (sensu Thulborn and Wade, 1989) during which the animal’s center of an outward movement of digit III during the kick-off, and one where the gravity is positioned anterior to the foot (Roberts and Scales, 2002) and whole foot is rotated slightly outward and the median and basal parts of the weight of the animal is transmitted to the distal part of the longer, the foot are displaced outward relative to the walking direction (Fig. 14). middle digit, which becomes more deeply imprinted into the substrate The tracks from the Lower Jurassic show a less clear pattern of than the outer digits. The sideways deformation of the proximal part of deformation. The incomplete track 7 exhibits primarily outward defor- digit IV occurs during the stance phase, when the animal’s center of mation caused by digits III and IV (Fig. 7), but because the track com- gravity is moved anterior to the point of force while progressing toward prises only the digit impressions, no deformation from the proximal part the kick-off phase. In this case, this must have caused an outward rota- of the foot is observable. Track 8 shows a slight inward deformation in tion of the proximal part of the foot. Within these two groups of defor- the impression of the proximal parts of digit III, contrary to what is mation, tracks 5, 8, 9, and 10 show an inward deformation of the trace observed in the other tracks that are all deformed outward and have a from the distal phalanx of digit III. This deformation pattern is inter- more normal, inward rotation at the tip of the same digit (see other preted as having formed when the distal part of the foot (in this case digit tracks). Track 9 shows the dominant outward deformation caused by III) twisted outward during the kick-off (implying rotation), causing the digit III, and less by digits II and IV. Like in track 8, the tip of digit III is digit bearing the claw to experience a slight twist inward due to drag from deformed inward, as if the tip with the claw had been twisted inward the claw in the sediment. The deformation of the Late Jurassic tracks while the middle of the digit is turned outward during the kick-off. In differs from the two types exhibited by the Late Triassic and Early track 10, the deformation is much like that of tracks 4 and 6 in that the Jurassic tracks in that their outward deformation is even and concen- outward deformation is most prominent in digit III, but also the whole trated along the whole length of digits III and IV and subordinately by length of digit IV displays a prominent outward twist. The tip of digit III digit II (Fig. 14). There is thus a clear tendency among the studied tracks shows an inward deformation in its distal part (Fig. 10). The Late Juras- from Late Triassic, Early Jurassic and Late Jurassic theropods to have an sic tracks (11-13) all show a similar, even, outward deformation, most outward component in their stride; only track 8 show no apparent out- prominent in digits III and IV (Figs. 11-13). These tracks further show ward deformation, but is instead slightly deformed inward in the middle the inward rotation of the tip of digit III, which is a common feature of part of digit III. This might merely be the result of a local, unusual foot 362

FIGURE 15. Schematics showing the different penetrations of the left foot into the sediment between the Late Triassic, Early Jurassic and Late Jurassic tracks (ideal section of the central part of the track perpendicular to the long axis of digit III). A, The base level, here defined as the line between the bottom of the impressions of digits II and IV, below the Late Triassic and Early Jurassic tracks are horizontal to sub-horizontal. In tracks with deformation type 1, the bases are horizontal; in the tracks with deformation type 2, the impression of digit IV is the deepest, making the base level slope slightly outward relative to the midline of the trackway. Base line is indicated by broken line below the foot. B, The base below the Late Jurassic tracks is sloping strongly inward relative to the midline of the trackway. This is due to digit II being significantly more deeply impressed into the sediment than digit IV. Base line is indicated with broken line below the foot. Even though the depths of the different parts of the feet are impressed into the sediment and the base lines show different angles, all the tracks (except track 8) are deformed outward relative to the midline of the trackway. movement related to differences in gait or posture during walk. A possi- etries and mechanics of the limbs between Late Triassic and Late Jurassic bility that should be taken into account is that the differences in deforma- theropods. Holistic studies of evolution of theropod tracks by Lockley tion pattern observed in the Late Jurassic tracks could be due to the foot (in press) show a trend in which the digital pads on digit II become having been impressed to a greater depth in the substrate. Future studies compressed and fuse in one large fleshy pad among the larger theropods. including more tracks from more localities, time periods, and sedimen- In non-avian theropod dinosaurs, the metatarsophalangeal joints of dig- tary conditions will shed light on this issue. its II and III must have been held off the ground, while digit IV was Theropod dinosaurs shared a common locomotory design (Farlow impressed over its entire length. This results in a typical asymmetry of et al., 2000). Despite an enormous range in body size, theropods all had the proximal portion of the footprint (Lockley, in press). Bird footprints very similar hindlimb proportions. Even so, there are distinctive varia- are much more symmetrical than those of non-avian dinosaurs. The tions on the overall theropod morphological theme along the various “heel” is formed by a thick metatarsal pad located beneath the distal end groups. Theropod hips and hindlimbs show marked morphological of the trochlea of digit III on the tarsometatarsus and the metatarsopha- changes that are consistent with functional changes during their evolu- langeal joints of digits II and IV are elevated off the ground (Lucas and tion. Holtz (1995) demonstrated that arctometatarsalians markedly in- Stettenheim, 1972). Padian and Olsen (1989) observed that the relatively creased the length of the distal hindlimb elements relative to the femur longer metatarsi of birds have to be more vertically oriented than in length compared with other theropods. In addition, these derived typical non-avian theropods to keep the animal’s center of mass posi- theropods also developed a tightly interlocked proximal metatarsus that tioned above the feet. Another complicating factor is the size and orien- more effectively transmitted locomotory forces from the foot to the tation of the femur (Padian and Olsen, 1989; Gatesy, 1991) which is lower leg than did the ancestral theropod metatarsus. Trackway data generally elongate and slender in non-avian theropods and short and suggest conservatism in some aspects of locomotory activity during stout in birds. Non-avian theropod tracks tend to have low interdigital theropod evolution (Farlow et al., 2000). The trackway data provide no angles, while bird footprints often have higher values. evidence for changes in stride lengths (at least compared with footprint Sideways deformation is not observed in studies of tracks and length) during walking or running in comparisons of Early, as opposed to trackways from extant cursorial theropods, like the rhea, Rhea americana later, Mesozoic theropods. Conservatism also seems to characterize the (Padian and Olsen 1989). Similar studies of tracks and trackways from relationship between the total leg length and the total length of digit III (a an emu, Dromaius novaehollandae, also revealed no such sideways de- very rough proxy for footprint length). So the apparent similarity of the formation, but it was observed that the emu carries significantly less walking kinematics in the footprints studied here is, as a result, not weight on digit II, which in many cases becomes less impressed into the surprising. However, while the general tendency for the tracks to show substrate (Milàn, 2006). The ostrich, Struthio camelus, has lost digit II an outward component in the foot movement seems consistent in all the and has a functionally didactyl foot consisting of digits III and IV (Davies, studied tracks (except track 8), the pattern in which the foot penetrated 2002); studies of its tracks and trackways also showed no apparent down into the substrate shows some difference from the Late Triassic outward deformation (Farlow, 1989). and Early Jurassic tracks to the Late Jurassic tracks. In the Late Triassic The subsurface deformation in theropod tracks described here and Early Jurassic tracks, the base of the footprint is of almost equal demonstrates that there was apparently a change in the walking kinemat- depth below the imprints of digits II and IV, while digit II is relatively ics from Late Triassic and Early Jurassic to Late Jurassic theropods. The more deeply impressed into the substrate than digit IV in the Late Juras- former show either an outward component in the distal or proximal part sic tracks, causing the base of the track to slope sharply toward the of the foot during the stride or an outward rotation of the basal part of the inside of the tracks (Fig. 15). This suggests an analogous shifting of the foot; the latter were created by an even, outward component of foot weight during walk, but a different geometry of penetration of the distal movement during the stride. Contrary to the outward deformations ob- part of the foot into the ground, with implications for different geom- served in the Mesozoic theropod tracks, extant cursorial birds appar- 363 ently do not incorporate an outward component in their stride. IV, which becomes deeply impressed into the substrate. The tracks from the Late Jurassic of Spain show an even, unidirectional, outward defor- CONCLUSION mation of the whole track created by the foot being pushed outward and Careful study of the deformation structures occurring in and around down into the sediment during the stride. The bases of the Late Triassic dinosaur tracks provides valuable information about the walking kine- and Early Jurassic tracks below the impressions of digits II and IV differ matics and gait of the track maker. The method of studying tracks in from the Late Jurassic tracks. In the former, the base is almost horizon- vertical section has proven successful in obtaining additional details about tal, with digit IV being only slightly more deeply impressed than digit II. the walking kinematics that rarely could be obtained from studying the In the Late Jurassic tracks, the base is sloping inward as digit II is true track at the surface, especially if these have been exposed to erosion significantly more deeply impressed than digit IV. Most of the studied or are poorly preserved. tracks show an inward rotation of the distal end of digit III created by the Taken all together or singularly, true tracks, undertracks, natural claw being rotated inward while the foot is turned outward during the casts, and subsediment deformation structures observed in vertical sec- stride. tions, provides unique information about the walking kinematics of track ACKNOWLEDGMENTS makers, and are thus important factors to incorporate in the study of fossil footprints. The research of J.M. is sponsored by a Ph.D. grant from the The studied tracks from the Late Triassic of Greenland and the Faculty of Natural Science, University of Copenhagen. Niels Bonde Early Jurassic of Italy both show an outward deformation resulting from kindly placed Late Triassic footprints at our disposition and allowed us an outward rotation of the foot during the stride. This deformation falls to cut sections through them. Jerry Harris, Martin G. Lockley, and an into two distinct types. One type is formed by digit III being twisted anonymous reviewer are thanked for their critical and constructive re- outward during the kick-off and the second type is formed by an out- views. ward translocation of the proximal parts of the foot and especially digit

REFERENCES

Allen, J.R.L., 1997, Subfossil mammalian tracks (Flandrian) in the Severn M., 2000, Theropod locomotion: American Zoologist, v. 40, p. 640- Estuary, S.W. Britain: mechanics of formation, preservation and distri- 663. bution: Philosophical Transactions of the Royal Society of London B, Fornós, J.J., Bromley, R.G., Clemmensen, L.B. and Rodriguez-Perea, A., v. 352, p. 481-518. 2002, Tracks and trackways of Myotragus balearicus Bate (Artiodac- Avanzini, M., 1998, Anatomy of a footprint: bioturbation as a key to tyla, Caprinae) in Pleistocene aeolianites from Mallorca (Balearic Is- understanding dinosaur walk dynamics: Ichnos, v. 6, p. 129-139. lands, western Mediterranean): Palaeogeography, Palaeoclimatology, Avanzini, M., García-Ramos, J.C., Lires, J., Menegon, M., Piñuela, L, and Palaeoecology, v. 180, p. 277-313. Fernández, L.A., 2005, Turtle tracks from the Late Jurassic of Asturias, García-Ramos, J.C. and Gutiérrez Claverol, M., 1995, La Cobertera Spain: Acta Palaeontologica Polonica, v. 50, p. 743-755. Mesozoico-Terciaria, in Aramburu, C., and Bastida, F., eds., Geología de Avanzini, M., Gierlinski, G. and Leonardi G., 2001a, First report of sitting Asturias: Gijón, Editorial Trea, p. 81-94. Anomoepus tracks in European Lower Jurassic (Lavini di Marco site - García-Ramos, J.C., Lires, J. and Piñuela, L., 2002, Dinosaurios: rutas por el northern Italy): Rivista Italiana di Paleontologia e Stratigrafia, v. 107, Jurasico de Asturias: Asturias, La Voz de Asturias, Lugones Siero, 204 p. p. 131-136. García-Ramos, J.C., Piñuela, L. and Lires, J., 2004, Guía del Jurasico de Avanzini, M., Leonardi, G. and Mietto, P., 2004, Lavinipes cheminii Asturias. Rutas por los yacimientos de huellas de dinosaurios: Gijon, Ed. ichnogen., ichnosp. nov., a possible sauropodomorph tracks from the Zinco Comunicación, 118 p. Lower Jurassic of the Italian Alps: Ichnos, v. 10, p. 179-193. Gatesy, S.M., 1991, Hind limb scaling in birds and other theropods: implica- Avanzini, M., Leonardi, G. and Tomasoni R., 2001b, Enigmatic dinosaur tions for terrestrial locomotion: Journal of Morphology, v. 209, p. 83- trackways from the Lower Jurassic (Pliensbachian) of the Sarca Valley, 96. northeast Italy: Ichnos, v. 8, p. 235-242. Gatesy, S.M., 2001, Skin impressions of Triassic theropods as records of Bromley, R.G. and Asgaard, U., 1979, Triassic freshwater ichnocoenosis foot movement: Bulletin of the Museum of Comparative Zoology, v. from Carlsberg Fjord, East Greenland: Palaeogeography, 156, p. 137-149. Palaeoclimatology, Palaeoecology, v. 28, p. 39-80. Gatesy, S.M., 2003, Direct and indirect track features: what sediment did a Brown, T., Jr., 1999, The science and art of tracking: New York, Berkley dinosaur touch?: Ichnos, v. 10, p. 91-98. Books, 219 p. Gatesy, S.M., Middleton, K.M., Jenkins, F.A. and Shubin, N.H., 1999, Three- Canudo J.I. and Ruiz-Omeñaca, J.I., 2004, Los restos directos de dinosaurios dimensional preservation of foot movements in Triassic theropod dino- terópodos (excluyendo aves) en España, in Pérez-Lorente, F., ed., saurs: Nature, v. 399, p. 141-144. Dinosaurios y otros reptiles Mesozoicos en España: IER, Ciencias de la Holtz, T.R., 1995, The arctometatarsalian pes, an unusual structure of Tierra, v. 26, Instituto de Estudios Riojanos, Logroño, p. 347-374. metatarsus of Cretaceous Theropoda (Dinosauria: Saurischia): Journal Christiansen, P., 1997, Hindlimbs and feet, in Currie, P.J. and Padian, K., of Vertebrate Paleontology, v. 14, p. 1100-1117. eds., Encyclopedia of dinosaurs: San Diego, Academic Press, p. 320-328. Jenkins, F.A., Jr., Shubin, N.H., Amaral, W.W., Gatesy, S.M., Schaff, C.R., Clemmensen, L.B., Kent, D.V. and Jenkins, F.A., 1998, A Late Triassic lake Clemmensen, L.B., Downs, W.R., Davidson, A.R., Bonde, N. and Osbéck, system in East Greenland: facies, depositional cycles and palaeoclimate: F., 1994, Late Triassic continental vertebrates and depositional envi- Palaeogeography, Palaeoclimatology, Palaeoecology, v. 140, p. 135- ronments of the Fleming Fjord Formation, Jameson Land, East 159. Greenland: Meddelelser om Grønland, Geoscience, v. 32, p. 1-25. Dal Sasso, C., 2001, Dinosauri italiani: Venezia, Marsilio Editore, 260 p. Leonardi, G., 1987, Glossary and manual of tetrapod paleoichnology: Brasilia, Davies, S.J.J.F., 2002, Bird families of the world: ratites and tinamous: Departmento Nacional da Producao Mineral, 117 p. Oxford, Oxford University Press, 310 p. Leonardi G. and Mietto P., eds., 2000, Dinosauri in Italia: le orme giurassiche Farlow, J.O., 1989, Ostrich footprints and trackways: implications for dino- dei Lavini di Marco (Trentino) e gli altri resti fossili italiani: Pisa/Roma, saur ichnology, in Gillette, D.D. and Lockley, M.G., eds., Dinosaur tracks Accademia Editoriale, 490 p. and traces: Cambridge, Cambridge University Press, p. 243-248. Lires, J., 2000, Icnitas de dinosaurios cuadrúpedos del Jurásico de Asturias. Farlow, J.O., Gatesy, S.M., Holtz, T.R., Hutchinson, J.R. and Robinson, J.- Morfometría, morfología e interpretación [M.Sc. thesis]: Oviedo, 364 Universidad de Oviedo, 61 p. Milàn, J., 2006, Variations in the morphology of emu (Dromaius Lockley, M., 1991, Tracking dinosaurs: Cambridge, Cambridge University novaehollandiae) tracks, due to differences in walking pattern and sub- Press, 238 p. strate consistency: ichnotaxonomic implications: Palaeontology, v. 49, Lockley, M.G., 2000, Philosophical perspectives on theropod track mor- p. 405-420. phology: blending qualities and quantities in the science of ichnology: Milàn, J. and Bromley, R.G., 2006, True tracks, undertracks and eroded Gaia, v. 15, p. 279-300. tracks, experimental work with tetrapod tracks in laboratory and field: Lockley, M.G., in press, The morphodynamics of dinosaurs, other archosaurs Palaeogeography, Palaeoclimatology, Palaeoecology, v. 231, p. 253- and their trackways: holistic insights into relationships between feet, 264. limbs and the whole body, in Bromley, R.G., Buatois, L.A., Márango, Milàn, J., Clemmensen, L.B. and Bonde, N., 2004, Vertical sections through M.G., Genise, J.F. and Melchor, R.N., eds., Sediment-organism interac- dinosaur tracks (Late Triassic lake deposits, East Greenland) – undertracks tions: a multifaceted ichnology: SEPM Special Publications. and other subsurface deformation structures revealed: Lethaia, v. 37, p. Lockley, M.G., Meyer, C. and Santos V.F., 2000, Megalosauripus and the 285-296. problematic concept of megalosaur footprints: Gaia, v. 15, p. 313-337. Nicosia, U., Avanzini, M., Barbera, C., Conti, M.A., Dalla Vecchia, F.M., Loope, D.B., 1986, Recognizing and utilizing vertebrate tracks in cross Dal Sasso, C., Gianolla, P., Leonardi, G., Loi, M., Mariotti, N., Mietto, section. Cenozoic hoofprints from Nebraska: Palaios, v. 1, p. 141-151. P., Morsilli, M., Paganoni, A., Petti, F.M., Piubelli, D., Raia, P., Renesto, Lucas, A.M. and Stettenheim, P.R., 1972, Avian anatomy, integument. Part S., Sacchi, E. and Signore, M., 2005, I vertebrati continentali del I, Agricultura Handbook 362: Washington D.C., U.S. Department of Paleozoico e Mesozoico, in Bonfiglio, L., ed., Paleontologia dei vertebrati Agriculture, 340 p. in Italia: Memorie del Museo Civico di Storia Naturale di Verona, 2° Manning, P., 2004, A new approach to the analysis and interpretation of serie, Sezione di Scienze della Terra, v. 6, p. 123-154. tracks: examples from the Dinosauria, in McIlroy, D., ed., The applica- Padian, K. and Olsen, P.E., 1989, Ratite footprints and the stance and gait tion of ichnology to palaeoenviromental and stratigraphic analysis: of Mesozoic theropods, in Gillette, D.D. and Lockley, M.G., eds., Dino- Geological Society of London, Special Publication 228, p. 93-123. saur tracks and traces: Cambridge, Cambridge University Press, p. 231- Martínez, R., García-Ramos, J.C., Piñuela, L., Lires, J., Luna, M. and Veigas, 241. D., 2001, Vértebras caudales de Sauropoda y Theropoda (Dinosauria: Piñuela, L., 2000, Icnitas de dinosaurios bípedos del Jurásico de Asturias. Saurischia) del Jurásico Superior de Asturias, España, in Diez, J.B. and Morfometría, morfología e interpretación [M.Sc. thesis]: Oviedo, Balbino, A.C. , eds., Libro de resúmenes del I Congresso Ibérico de Universidad de Oviedo, 63 p. Paleontología/XVI Jornadas de la Sociedad Española de Paleontología, Piubelli, D., Avanzini, M. and Mietto, P., 2005, The Early Jurassic ichnogenus Évora, Portugal, p. 113-114. Kayentapus at Lavini di Marco ichnosite (NE Italy). Global distribution Martínez, R., García-Ramos, J.C., Piñuela, L., Lires, J. and Luna, M., 2001, and palaeogeographic implications: Bollettino della Società Geologica Dinosaur remains from the Principality of Asturias. Journal of Verte- Italiana, v. 124, p. 259-267. brate Paleontology, v. 21, p. 78A. Robert, T.J. and Scales, J.A., 2002, Mechanical power output during running Masetti, D., 2000, La geologia dei Lavini di Marco, in Leonardi G. and accelerations in wild turkeys: Journal of Experimental Biology, v. 205, Mietto P., eds., Dinosauri in Italia: le orme giurassiche dei Lavini di p. 1485-1494. Marco (Trentino) e gli altri resti fossili italiani: Pisa/Roma, Academia Thulborn, R.A. and Wade, M., 1989, A footprint as history of movement, Editoriale, p. 65-114. in Gillette, D.D. and Lockley, M.G., eds., Dinosaur tracks and traces: Cambridge, Cambridge University Press. pp. 51-56. Jesper Milàn – Vertebrate Ichnology Appendix ______

Taxonomic disclaimer:

This is a manuscript produced only for the public examination of the doctoral thesis of which the paper is a part. Under article 8.2 of the International Code of Zoological Nomenclature the paper is not issued for permanent scientific record and is not published within the meaning of the code.

Milàn et al. – Traces of Behaviour ______

Dinosaur tracks showing evidence of individual behavior from the Lower Jurassic Navajo Sandstone, Coyotes Buttes locality, Arizona-Utah, USA.

JESPER MILÀN, DAVID B. LOOPE and RICHARD G. BROMLEY

Milàn, J., Loope, D.B. and Bromley, R.G. 200x. Dinosaur tracks showing evidence of individual behavior from the Lower Jurassic Navajo Sandstone, Coyotes Buttes locality, Arizona-Utah, USA. Acta Palaeontologica Polonica

Numerous tracks and trackways are preserved in the eolian cross-strata of the Lower Jurassic Navajo Sandstone of Northern Arizona and Southern Utah. Tracks and trackways of small theropod dinosaurs are particularly abundant at a single horizon. This paper revises the interpretation of a well preserved trackway considered to be from a sauropodomorph dinosaur. It is named Navahopus coyotensis n.isp. on the basis of morphological differences from the type ichnospecies N. falcipollex Baird, 1980. A crouching trace from a theropod dinosaur is described, having impressions of all the limbs, the ischial callosity, the tail and tracks leading to and away from the crouching site.

Keywords: Lower Jurassic, aeolian deposits, footprints, Navahopus coyotensis n. isp., sauropodomorph, crouching theropod.

Jesper Milàn [[email protected]] and Richard G. Bromley [[email protected]], Geological Institute, University of Copenhagen, Øster Voldgade 10, DK-1350 Copenhagen K, Denmark; David B. Loope [[email protected]], Department of Geociences, University of Nebraska, Lincoln, NE 68588-0340, Nebraska, USA. ______

1 Milàn et al. – Traces of Behaviour ______

Introduction tracks are abundant on the closely spaced, The Lower Jurassic Navajo Sandstone, exposed in northern Arizona and southern Utah, contains a diverse tetrapod fauna comprising tritylodonts, crocodylomorphs, basal theropods, and sauropodomorphs (Irmis 2005). In addition to this body fossil record, numerous tracks and trackways are preserved in the colorful eolian cross-strata. The Coyotes Buttes site, located on the Utah- Arizona border (Fig. 1), contains especially numerous tracks and trackways from small theropod dinosaurs, tracks and trackways of presumed crocodilian origin, and both quadropedal and bipedal trackways of larger sauropodomorph dinosaurs (Loope and Rowe 2003; Loope 2006; Irmis 2005). The exposures at Coyotes Buttes comprise deposits of the Lower Jurassic Navajo Sandstone, which is the remains of a vast inland dune sea, considered to have covered as much as 350.000 km2 at the time of deposition, and to have reached thicknesses in excess of 700 meters. Tracks and trackways of dinosaurs are common in the Navajo Sandstone, and more than 60 sites have been reported in the Navajo Sandstone and its correlative strata (Lockley 1998). The Fig. 1. Stratigraphic column of the Navajo Sandstone outcrop at Coyotes Buttes. The studied tracks are all highest density of tracks is to be found at the found at the top of the middle zone of bioturbation. The Coyotes Buttes locality is located at the border between Coyotes Buttes locality, where tracks of Utah and Arizona at 36°59´58´´N, 112°00´35´´W (WGS84). theropods, prosauropods and other reptilian

2 Milàn et al. – Traces of Behaviour ______

steeply dipping, depositional surfaces Dinosaur tracks (Loope 2006). Theropod tracks The outcrops of Navajo Sandstone at The majority of tracks found at Coyotes Coyotes Buttes and other exposures in Buttes are tridactyl theropod tracks from 5 to southern Utah and northern Arizona are 15 cm in length that in many instances can be composed of large-scale, southeast-dipping traced as trackways for several meters on eolian cross-strata (Loope and Rowe 2003). exposed surfaces (Fig. 2). Some of the The full height of the bedforms that created surfaces contain a density of tracks so high the distinct, large-scale cross-strata is that it becomes impossible to distinguish the estimated to have represented a dune height of tracks from the individual trackways. All the about 30 meters, from studies of a nearby theropod tracks are preserved as true tracks outcrop of similarly sized cross-strata (Hunter infilled with sand of a different color, which and Rubin 1983). At Coyotes Buttes, tracks of makes them stand out against the color of the vertebrates occur in a single, distinct ambient surface. The outline of the tracks stratigraphic interval up to 10 meters thick appears rounded, lacking preservation of any (Fig. 1). Loope and Rowe (2003) described anatomical details, very similar to evidence for prolonged periods of enhanced experimentally obtained tracks from dry sand summer monsoon rainfalls. During these (Milàn 2006). This mode of preservation periods animal and plant life flourished, but prevents identification of anatomical details during the drier winter season, dune migration such as number and configuration of digital continued at about the same rate as in the pads, and skin impression, but traces of long pluvial intervals. The herein described tracks sharp claws are visible in many cases. are from the middle of three trace-fossil-rich Among the numerous tracks and intervals (Fig. 1). trackways of small theropod dinosaurs, an The aim of this study is to describe the association of tracks constituting the resting dinosaur tracks found at the Coyotes Buttes trace of a small theropod dinosaur was found locality, with special emphasis on the traces on the steeply sloping lee face of a dune with of a resting theropod and on a the animal facing up-slope (Fig. 3). sauropodomorph trackway; both show unique traces of individual behavior.

3 Milàn et al. – Traces of Behaviour ______

Fig. 2. A. Long, narrow-gait trackway from a small theropod dinosaur. B. Close up of two consecutive footprints from a theropod trackway. The tracks are preserved as true tracks infilled with darker colored sand.

impressions. The sub-circular impression of The angle of slope today (after the ischial callosity is located behind the correction for tectonic tilting) is around 25 metatarsus impressions and approximately in degrees due to compaction of the sandstone, the midline of the track constellation. Further but the original angle of slope when the behind the pubic impression is the curved, animal made the trace is estimated to have partial impression of the dinosaur’s tail. In been around 32 degrees, which is the residual connection to the resting track, a single track angle of dry sand after shearing (Allen 1984). is located leading to the resting place, and Most prominent in the resting trace are the four tracks leading away from the resting two elongated, parallel metatarsus place (Fig. 3). impressions. In front of the two metatarsus impressions are two small, rounded manus

4 Milàn et al. – Traces of Behaviour ______

Fig. 3. Trace of a crouching theropod. A. The crouching track comprises subparallel impressions of the metatarsus, two small undetailed manus imprints, the imprints of the ischial callosity, the impression of the tail, and tracks from the dinosaur walking toward and away from the resting site. Upslope direction is to the right. B. Interpretative drawing of an unspecified small theropod dinosaur crouching down to produce the configuration of tracks seen in A. The tracks are reconstructed on a 32 degrees slope which is the residual angle of dry sand after shearing (Allen 1984). The animal was progressing directly up the slope and was crouching facing upslope before it continued directly up the dune face.

sloping surface representing the lee face of a Sauropodomorph tracks dune (Fig. 4). As in the case with the theropod An approximately 2.5 meter-long resting trace, the present day slope of the segment of a trackway from a larger surface is around 25 degrees, compacted from quadropedal trackmaker is found on a steeply- an original 32 degrees. The tracks show a

5 Milàn et al. – Traces of Behaviour ______distinct heteropody. The pes impressions are of a large pollex claw. Like the pes, the around 20 cm long, but the exact length is manus impressions bear evidence of the difficult to estimate as the claws have been manual claws being dragged forward through dragged forward through the substrate, the sand as the animal walked. elongating the digit impressions. The trackway pattern changes along the exposed trackway segment. The first part of the trackway shows the orientation of the feet to be angled up the slope while the direction of the trackway is at a high angle to the slope (Fig. 5). Midway through the exposed trackway segment, the trackway changes direction to directly upslope and the orientation of the feet correspond to the orientation of the trackway.

Systematic Paleontology Genus Navahopus Baird, 1980 Type species N. falcipollex Baird, 1980 Emended diagnosis: Quadropedal trackway of sauropodomorph origin. Pes tetradactyl, Fig. 4. Trackway from a sauropodomorph dinosaur clawed, having digit I shortest and either digit walking up the lee slope of a dune. Notice the elongated traces from the claws being dragged through III or IV the longest. Manus functionally the sediment. The present day slope is approximately 25 degrees, due to compaction of the sediments. The tridactyl, having short, clawed digits II and III original angle of slope was around 32 degrees which is the residual angle of slope of dry sand after shearing directed forward, and a horizontally (Allen 1984). Figure from Loope & Rowe (2003). recumbent, falciform medially directed pollax

claw. The pes is tetradactyl, with the two outer digits of subequal length, followed by two Navahopus falcipollex Baird, 1980 increasingly shorter digits. A deep gap Diagnosis: Quadropedal trackway referable separates the two outer digits from each other. to the saurischian family Plateosauridae. Pes The manus is tridactyl consisting of two short, tetradactyl, clawed, with digits in order of forward-oriented, clawed digit impressions increasing length I-II-IV-III; manus and a prominent inwardly directed impression

6 Milàn et al. – Traces of Behaviour ______functionally tridactyl with short, clawed digits II and III directed forward, and a horizontally recumbent, falciform pollex claw directed medially.

Navahopus coyotensis isp. nov. Diagnosis: Navahopus having pes digit IV of equal or greater length than digit III. Pes digits III and IV having pronouncedly greater separation than between digits I and II and II and III. Etymology. Navaho (Spanish Navajo) after the source formation and the Navajo Indian Reservation, and pus, a foot. Coyotensis, refers to the locality Coyotes Buttes, where the type specimen is found.

Discussion The vertebrate track assemblage in the ancient dune deposits exposed at Coyotes Buttes is predominantly composed of tracks of small theropod dinosaurs with foot lengths from 5 to 15 centimeters, and less common tracks from sauropodomorphs and presumed crocodilians. What is unique about the track assemblage is that even though each track is preserved in a relatively undetailed mode due to the poor preservational potential of tracks Fig. 5. Sketch of the trackway in figure 4. RM = right manus, LM = left manus, RP = right pes, in loose sand, valuable information about LP = left pes. The solid arrow indicates the direction of progression and the broken-line arrow the orientation individual trackmaker behavior and anatomy of the body during progression. Notice how the animal walked at an oblique angle up-slope in the first half of can be obtained from the locality. the trackway, and then changed to progress head-on, up the slope.

7 Milàn et al. – Traces of Behaviour ______

Direct traces of theropods in full ischial impression, and an elongated belly crouching posture with impressions of the impression (Lockley et al. 2003). metatarsi, ischial callosity and mani are The newly discovered resting trace uncommon and so far only three specimens from Coyotes Buttes resembles the Chinese have been described in the literature: two are and Connecticut Valley specimens in that it from the lower Jurassic of Connecticut Valley comprises sub parallel metatarsus and another is from the Lower Jurassic of impressions, small manus impressions, and a Sichuan, China (Lockley et al. 2003). The sub circular ischial impression. In addition to Connecticut Valley specimens were originally these features, however, the Coyotes Buttes described by Hitchcock (1848) as Sauroupus specimen contains the partial impression of barratti. The ichnotaxonomical status of these the tail, and the tracks leading to and from the specimens has subsequently changed place where the theropod was crouched down. numerous times, but at present they are There are no fine anatomical details preserved carefully assigned to Grallator. For a detailed in the Coyotes Buttes specimen, like skin ichnotaxonomical account of the Connecticut texture, number and arrangement of digital Valley specimens, see Lockley et al. (2003). pads etc. This is due to the consistency of the The Chinese specimen was found at the Wu substrate at the time of trackmaking. The Ma Cun site, Sichuan, China (Yang and Yang tracks were emplaced in eolian sands. If dry, 1987). Recently a fourth hitherto undescribed sand is a poor medium for the preservation of trace of a crouching theropod was discovered anatomical details as it has poor cohesive at the Johnson Farm Dinosaur Dinosaur Track properties and tends to flow together site, St. George, Utah (Milner et al. in press). immediately after the track has been made, The Connecticut Valley material obliterating all but the gross overall shape of comprises subparallel metatarsus impressions the footprint (Milàn 2006). and a rounded impression of the ischial The configuration of the tracks callosity (Lockley et al. 2003). One specimen constituting the resting trace allows includes the elongated impression of the reconstruction of the timing and succession in dinosaur´s belly as well, and the other which the tracks were emplaced and the specimen shows small manus impressions. possible movements exercised by the dinosaur The Chinese specimen comprises subparallel when it lay down and got up again. The first metatarsus impressions, a sub-spherical track to be emplaced is the track leading toward the resting trace. Assuming the

8 Milàn et al. – Traces of Behaviour ______bipedal theropod dinosaur settles down to rest (2003). However, a closer examination in a way similar to modern flightless birds, showed this initial description to be incorrect. like the emu (Milàn 2006), the theropod drops The morphology of the manus and pes down on the elongated metatarsi and the impressions, and especially the manus tracks hands and ischial callosity come into contact showing the prominent impression of a large, with the ground. When an emu rises up from inward directed pollex claw, suggests a resting position, it tilts the front part of the sauropodomorph trackmaker rather than a body forward and upward, before rising up on synapsid trackmaker. its feet. When a theropod with its long tail The trackway Navahopus falcipollex rises up in a similar way, then it would be Baird, 1980, described from the Navajo necessary that the tail be pressed down into Sandstone of Arizona, was identified as a the substrate. This would account for the prosauropod trackway, on the basis of curved tail impressions located behind the morphological comparisons between tracks paired impressions of the metatarsus and the and the pedal skeleton from the prosauropod ischial impression. Ammosaurus, which had been excavated in The apparent sideways walking pattern the area near the trackway (Baird, 1980). on the first part of the sauropodomorph The new trackway from Coyotes Buttes, trackway trudging up the steep slope of the Navahopus coyotensis is similar to N. dune is similar to the trackway patterns found falcipollex, in that the manus tracks consist of in the Permian tetrapod trackways from the the impressions of two short, clawed, forward eolian Coconino Sandstone (Brand & Tang facing digits and the impression of a large 1991). The sideways walking pattern of these medially directed, falciform pollex claw. The tracks was interpreted as evidence for an pes is tetradactyl having the impressions of underwater origin of the tracks, but that digits I and II of subequeal length, followed interpretation is highly controversial (Lockley by the successively shorter digits III and IV 1992; Loope 1992), and a similar sideways (Fig. 6). However, the Coyotes Buttes walking pattern has been described from goat trackway differs from N. falcipollex in having traces in Pleistocene eolianites from Mallorca a 10 cm deep gap separating the impressions (Fornós et al. 2002). of digits III and IV, while no apparent gap is The trackway from the Coyotes Buttes present between the impressions of digits I to was preliminarily identified as a synapsid III. trackway, Brasilichnium by Loope and Rowe

9 Milàn et al. – Traces of Behaviour ______

toes to allow better grip in the loose dune sand. Baird (1980) compared the N. falcipollex trackway with the pedal skeleton of the sauropodomorph dinosaur described as Ammosaurus (Galton 1971) from the Navajo Sandstone of northern Arizona, and concluded that the trackmaker responsible for N. falcipollex, was four-fifths the size of Ammosaurus.

Fig. 6. Comparisons between the manual and pedal skeleton of sauropodmorph material from the Navajo Sandstone of northern Arizona and the Navahopus coyotensis trackway from Coyotes Buttes. The skeletal material was originally referred to as Ammosaurus (Galton 1971), but has recently been revised and reinterpreted as belonging to an indeterminate sauropodomorph (Yates 2004). A. The manus of the sauropodomorph from northern Arizona is functionally tridactyl and consists of two short forward-facing digits II and III and the large pollex claw of digit I directed inward. The pes is tetradactyl with digits III and IV of subequal length, followed by the shorter digits II and I. Figure modified from Baird (1980). B, manus and pes couple from N. coyotensis. Notice Fig. 7. Sketch of N. coyotensis, manus and pes couples the close correspondence between the pedal skeleton from left and right side of the trackway. Digits III and and the tracks, here shown to the same scale. IV are separated by a deep gap, recognizable in all well-preserved tracks in the trackway.

The deep gap separating the pedal digits The taxonomic status of Ammosaurus III and IV is consistently present in all tracks material from Arizona has been revised since from both left and right side of the animal the original description by Galton (1971). (Figs. 4–7), which excludes the possibility Yates (2004) revised Ammosaurus and that it is a pathological phenomenon and Anchisaurs and demonstrated that the two suggests that it represents either a special foot were synonyms, and that the proposed morphology, with digit IV being functionally Ammosaurus material described by Galton separated from the adjacent digits, or the (1971) belonged to an indeterminate adaptation of a special walk with spread out sauropodomorph and a possible plateosaurian

10 Milàn et al. – Traces of Behaviour ______sauropodomorph. When comparing the directly up-slope. The trackway is named dimensions of the pedal skeleton of the Navahopus coyotensis n. isp. herein, based on “Ammosaurus” material from Arizona with morphological differences to the holotype N. N. coyotensis, there is an almost perfect falcipollex. agreement in size, even when additional length due to fleshy parts and horn sheaths on Acknowledgements the claws are taken into consideration (Fig. 6). The research of JM is supported by a N. coyotensis is thus a better match for an Ph.d. grant from the Faculty of Natural “Ammosaurus” trackway than is the holotype Science, University of Copenhagen. DL’s of N. falcipollex. Further, the new trackway work was supported by a research grant from shows an example of individual behaviour in the National Science Foundation (EAR02- that the tracks show the trackmaker to first 07893). Janet Gillette, Museum of Northern have walked somewhat sideways up the slope Arizona, Flagstaff, provided us with photos of of the dune, and then shifted to walk head-on the Navahopus holotype. up the slope.

Conclusion References Tetrapod tracks and trackways are abundant at several horizons in the Lower Allen, J.R.L. 1984: Sedimentary Structures: Jurassic, eolian Navajo Sandstone of Arizona Their Character and Physical Basis, v. II, and Utah. Among these tracks are several Developments in Sedimentology 30, examples of individual behavior among the Elsevier, p. 149. trackmakers. A trace of a small theropod Baird, D. 1980: A prosauropod dinosaur dinosaur crouching down on the sloping face trackway from the Navaho Sandstone of a dune, with the head orientated up slope, (Lower Jurassic) of Arizona. In Jacobs, has both the tail impression and tracks leading L.L. (ed.): Aspects of vertebrate history. to and away from the resting site. Careful Museum of Northern Arizona Press, Pp. examination of a unique quadropedal 219–230. trackway reveals it to be that of a Brand, L.R. and Tang, T. 1991: Fossil sauropodomorph dinosaur climbing the front vertebrate footprints in the Coconino face of a dune, first by progressing sideways Sandstone (Permian) of Northern Arizona: up the slope and then changing direction to go

11 Milàn et al. – Traces of Behaviour ______

Evidence for underwater origin. Geology Footmarks. Boston: William White. 220 19: 1201–1204. pp + plate I-LX. Fornós, J. J., Bromley, R.G., Clemmensen, Hunter, R.E. and Rubin, D.M. 1983: L.B. and Rodriguez-Perea, A. 2002: Interpreting cyclic crossbedding, with an Tracks and trackways of Myotragus example from the Navajo Sandstone. In balearicus Bate (Artiodactyla, Caprinae) Brookfield, M.E. & Ahlbrandt, T.S. in Pleistocene aeolianites from Mallorca (eds.): Eolian sediments and processes. (Balearic Islands, Western Elsevier, Amsterdam, p. 429–454. Mediterranean). Palaeogeography, Irmis, R.B. 2005: A review of the vertebrate Palaeoclimatology, Palaeoecology 180: fauna of the Lower Jurassic Navajo 277–313. Sandstone in Arizona. Mesa Southwest Galton, P.M. 1971: The prosauropod dinosaur Museum Bulletin 11: 55–71. Ammosaurus, the crocodile Protosuchus, Lockley, M. G. 1992: Comment and reply on and their bearing on the age of the Navaho “Fossil vertebrate footprints in the Sandstone of northeastern Arizona. Coconino sandstone (Permian) of Journal of Paleontology 45: 781–795. Northern Arizona: Evidence for Gierlinski. G. 1996: Featherlike impressions underwater origin”. Geology 20: 666–667. in a theropod resting trace from the Lower Lockley, M. G. 1998: The vertebrate track Jurassic of Massachusetts. In Morales, M. record. Nature 396: 429–432. (ed.): Continental Jurassic Symposium Lockley, M., Matsukawa, M. and Jianjun, L. Volume. Museum of Northern Arizona, 2003: Crouching theropods in taxonomic Flagstaff 60: 179–184. jungles: ichnological and Hitchcock, E. 1848: An attempt to ichnotaxonomical investigations of discriminate and describe the animals that footprints with metatarsal and ischial made the fossil footmarks of the United impressions. Ichnos 10: 169–177. States, and especially of New England. Loope, D.B. 1992: Comment on “Fossil American Academy of Arts & Sciences vertebrate footprints in the Coconino Memoir (n.s.) 3: 129–256. sandstone (Permian) of Northern Arizona: Hitchcock, E. 1858: Ichnology of New Evidence for underwater origin” Geology England, A report on the Sandstone of the 20: 667–668. Connecticut Valley Especially its Fossil

12 Milàn et al. – Traces of Behaviour ______

Loope, D.B. 2006: Dry-season tracks in dinosaur-triggered grainflows. Palaios 21: 132–142, Loope, D.B. and Rowe, C.M. 2003: Long- lived pluvial episodes during deposition of the Navaho Sandstone. The Journal of Geology 111: 223–232. Milàn, J. 2006. Variations in the morphology of emu (Dromaius novaehollandiae) tracks, reflecting differences in walking pattern and substrate consistency: ichnotaxonomical implications. Palaeontology 49: 405–420. Milner, A.R.C., Lockley, M.G. and Johnson, S.B. in press: The story of the St. George Dinosaur Discovery Site at Johnson Farm: an important new Lower Jurassic dinosaur tracksite from the Moenave Formation of Southwestern Utah. New Mexico Museum of Natural History and Science Bulletin 37. Yang, X.L. and Yang, D.H. 1987: Dinosaur footprints of Sichuan Basin. Sichuan Science and Technology Publications 1987: 1–30. Yates, A.M. 2004: Anchisaurus polyzelus (Hitchcock): the smallest known sauropod dinosaur and the evolution of gigantism among sauropodomorph dinosaurs. Postilla 230: 1–58.

13 Jesper Milàn – Vertebrate Ichnology Appendix ______

Preservation and Erosion of Theropod Tracks in Eolian Deposits: Examples from the Middle Jurassic Entrada Sandstone, Utah, U.S.A.

Jesper Mila`n and David B. Loope1

Geological Institute, University of Copenhagen, Øster Voldgade 10, DK-1350 Copenhagen K, Denmark (e-mail: [email protected])

ABSTRACT The Middle Jurassic Entrada Sandstone, exposed near the town of Escalante, southern Utah, consists of large-scale cross-bedded eolian deposits that are interbedded with horizontally laminated sand sheets and thin sets of eolian cross-strata, representing periods with a moister climate. The ﬂat-bedded units contain numerous tracks and trackways from small to large-sized theropod dinosaurs. These tracks are today exposed in several distinct erosional states, allowing detailed studies of track and undertrack formation in eolian deposits. Tracks that originally were emplaced on sloping surfaces show, in their present-day erosional state, a morphology distinct from those originally emplaced on horizontal surfaces. Further, the range of eroded track morphologies can help identify badly eroded tracks from nonbiogenic structures in similar deposits. No track begins or ends at the compression point. In other words, the track does not end at the ﬂoor, walls, horizons, and all the other visual places. Instead, the track is like the center of a concentric ring, a ring that ripples well beyond and under the existing track. (Tom Brown Jr., 1999)

Introduction Well-preserved vertebrate tracks in the rock record (Lockley 1991). When an animal walks, not only is can be an invaluable source of information about the tracking surface deformed, but also the layers foot morphology, soft tissue distribution, and skin subjacent to the tracking surface are deformed. Ex- texture of the trackmakers’ feet (Gatesy 2001). perimental work with track formation shows that However, in most instances, the tracks are less than the weight of the trackmaker is transferred radially perfectly preserved, and sometimes they can be outward in the sediment around and below the foot barely recognizable as tracks at all. Several factors (Allen 1989, 1997), forming a stacked succession of influence the preservation of tracks. The consis- undertracks (sensu Lockley 1991). Normally, it is tency of the sediment in which the track is em- possible to distinguish undertracks from true tracks placed is the most critical factor; slight variations in that they are less detailed than true tracks and in the water content of the sediment cause dra- never preserve any fine anatomical details like skin matically different track morphologies, resulting in impressions. Individual digit impressions become widely different fossil footprints from the same successively wider downward, and their ampli- trackmaker (Brand 1996; Gatesy et al. 1999, 2005; tudes are diminished in the subjacent layers (Mila`n Bromley 2001; Diedrich 2002; Gatesy 2003; Man- and Bromley 2006). The sedimentary infilling of the ning 2004; Mila`n 2006; Mila`n and Bromley 2006). track forms a natural cast of the track (Lockley The sediment surface on which the animal leaves 1991). In cases where the track is infilled gradually, its tracks is termed the tracking surface sensu For- the layers gradually drape the contours of the track no´s et al. (2002). The direct impression of the track- and can, like an undertrack, be misinterpreted as a maker’s foot in the tracking surface is the true track true track. The aim of this study is to describe and interpret Manuscript received July 24, 2006; accepted October 10, the many preservational variants of large theropod 2006. 1 Department of Geosciences, University of Nebraska, Lin- tracks from the Middle Jurassic Entrada Sandstone. coln, Nebraska 68588-0340, U.S.A.; e-mail: [email protected]. The different erosional cuts through the true tracks,

375 376 J. MILA` N AND D. B. LOOPE

Figure 1. Location map. The studied outcrop of Entrada Sandstone, 20-Mile Wash locality, is located in the southern part of Utah, southeast of the town of Escalante. Star indicates the location of the studied locality. The dashed line indicates the unpaved Hole in the Rock Road branch- ing southeast from Highway 12. the undertracks, and the natural casts allow us to demonstrate the possibility of recognizing and utilizing badly eroded tracks.

Geological Setting During the Middle Jurassic, present-day southern Utah was situated within the subtropics at a pa- leolatitude of about 18ЊN (Steiner 2003). The Es- calante Member is the uppermost unit in the Mid- dle Jurassic Entrada Sandstone (Thompson and Stokes 1970) and forms a prominent, 65–85-m- thick, white to gray, cliff-forming unit. The studied outcrops of the Entrada Sandstone are located at the northeastern margin of the Kaiparowits Pla- teau, at the 20-Mile Wash locality, about 30 km southeast of the town of Escalante (fig. 1). Dinosaur Figure 2. A, Entrada Sandstone exhibits large-scale tracks have previously been recorded from the 20- cross-bedding with varying dip directions, resulting from Mile Wash locality (e.g., Foster et al. 2000; Brei- seasonal shifts in wind regimes. B, Extensive surfaces of thaupt et al. 2004). flat-bedded deposits are exposed at the 20-Mile Wash lo- The large-scale compound crossbeds in the En- cality. The majority of the dinosaur tracks from the area are found on these surfaces. C, Flat-bedded strata in the trada Sandstone have variable dip directions (fig. 2A), Entrada Sandstone. Thin centimeter-thick sets of eolian suggesting a complex wind regime with more than cross-strata produced by small migrating dunes are in- one dominating wind direction. On the basis of the terstratified with sand sheets deposited by wind ripples. prevalence of simple sets of thick cross-strata with The section shown slopes toward the right at approxi- a persistently southeasterly dip in Lower Permian mately 30Њ. Journal of Geology THEROPOD TRACKS 377 through Lower Jurassic eolian sandstones, Loope et al. (2004) suggested that tropical westerlies, blowing across the equator from the Northern Hemisphere during December to February, dominated in the region for about 100 millions years. During the Me- sozoic, Pangaea was moving northward. The more variable dip directions observed in the Middle Jurassic Entrada Sandstone suggest that as Pangaea moved northward, summertime (June–August) windflow from the south played a more important role in the deposition of the Entrada Sandstone than in the Early Jurassic Navajo Sandstone. The abundant large dinosaur tracks in the Es- calante Member are preserved within flat-bedded eolian sand sheets and interbedded thin sets of eolian cross-strata (fig. 2B,2C; fig. 3) that accumulated above a shallow water table (Loope and Simp- son 1992). The tracks may have been formed during moist (pluvial) periods (Loope and Rowe 2003), when summer monsoon precipitation maintained a high water table in the dune field (fig. 3).

Ichnofauna

The flat-bedded deposits in the uppermost units of the Entrada Sandstone at the 20-Mile Wash locality contain a rich vertebrate ichnofauna dominated by tridactyl theropod footprints from 15 to 45 cm long. Dinosaur tracks are abundant in the Entrada–Sum- merville transition zone in Utah and can be traced in various outcrops for about 1,000 km2 as part of the “Moab Megatracksite” (Lockley and Hunt 1995; Lockley 1997). Most of the theropod tracks have well-defined digit impressions, and in the best-preserved cases, the impressions of the claws and the discrete vision of the digital pads are recognizable (fig. 4). A few of the tracks bear evidence of a semiplantigrade stance of the trackmaker, as a short metatarsus impression protrudes from the rear end of the tracks (fig. 4). The larger of the theropod tracks have been assigned to the ichnogenus Megalosauripus and the smaller ones to Theran- Figure 3. Stratigraphic section of Entrada Sandstone gospodus (Foster et al. 2000). A wide-gauge sauro- exposed at 20-Mile Wash. The large dinosaur tracks dis- pod trackway assigned to Brontopodus is extraor- cussed in this article are restricted to the thin sets of dinary in that it represents the earliest unequivocal eolian cross-strata in the upper part of the section, close evidence of sauropods from the western United to the Entrada–Summerville transition zone. States; furthermore, it has a prominent, sinuous drag mark from the tail in the midline of the track- mm-diameter traces resembling those produced by way (Foster et al. 2000). In addition to the numerous sand-swimming vertebrates (Loope 2005). dinosaur tracks, rhizoliths and insect burrows are also present in the flat-bedded strata. The large- Preservational Styles scale cross-strata of the Entrada Sandstone further contain 45-cm-diameter burrows from scratch- The dinosaur tracks in the Entrada Sandstone are digging vertebrates (Loope 2006a) and sinuous, 25- present in a wide range of preservational modes, 378 J. MILA` N AND D. B. LOOPE

Figure 4. Large theropod tracks and trackways exposed in flat-bedded eolian strata near the top of the Entrada Sandstone. A, Trackway of large, approximately 40 cm long, posteriorly elongated theropod tracks. B, Theropod track (58 cm long) with partial metatarsus impression. reflecting both differences in the gaits of the track- foot, in some cases even with the shape of the sharp makers and the paleoslope of the substrate. The claws preserved. Even the outlines of these tracks present-day subaerial erosion of the sandstone out- are well defined, but anatomical details like the crop has exposed the tracks in a variety of random size and distribution of digital pads cannot be ob- erosional cuts oriented at different angles through served because these are covered with the infilled both the true tracks and several subjacent horizons sediment. of undertracks. The sand sheets and the thin sets of cross-strata Many of the tridactyl theropod tracks are pre- produced by small dunes (that make up the gen- served as true tracks, that is, the original tracks erally flat-bedded strata between the deposits of the emplaced in the tracking surface by the dinosaurs. large dunes) consist of numerous thin layers of The true tracks have preserved impressions of finely laminated sand produced by migrating wind many anatomical details of the foot, such as the ripples. Not only did the dinosaurs deform the ac- number and arrangement of the digital pads in the tual tracking surface as they emplaced their feet, individual digits of the foot and impressions of the but they also folded and ruptured the subjacent lay- claws (fig. 4A). Several of the larger theropod tracks ers of sand to considerable depths below and around show a posterior elongation of the track, resulting the actual track. This has caused the formation of from a partial impression of the metatarsus during well-developed undertracks subjacent to many of the stride (fig. 4B). the observed tracks. The present-day erosion of the A number of the tracks and trackways are pre- outcrop has, because of differences in hardness be- served with the original sedimentary infilling still tween the flat-bedded and large-scale cross-bedded in place. The sediments infilling the tracks are, in deposits, favored the exposure of large surfaces of most cases, of a darker color than the sand consti- the flat-bedded deposits. Even though extensive tuting the original tracked surface. When eroded to surfaces of the flat-bedded deposits are exposed, the this state, the darker coloring of the infilled sedi- present-day surface does not represent a single bed- ments makes the tracks stand out clearly against ding plane; instead, it represents an erosional sur- the lighter-colored sediment of the original tracked face in a package of thin sets of cross-strata and surface (fig. 5A). Tracks preserved in this style are sandsheets eroded to different depths down through easily recognized as true tracks because they per- the layers. fectly reflect the outline and shape of the theropod The subsurface deformation of the thin sand lay- Journal of Geology THEROPOD TRACKS 379

Figure 5. Preservational styles of theropod tracks from the Entrada Sandstone. A, True track infilled with sand of a darker color. The shape and outline of the track is well defined, with clear impressions from the claws. Scale bar p 10 cm. B, True track with most of the sedimentary infilling eroded away. A series of concentric circles surrounding the track is present in the sediment. Knife handle is 10 cm. C, True track is here almost eroded away; only the imprint of the middle digit is still present. Notice the well-developed concentric deformations around the track. Knife handle is 10 cm. D, With the true track totally eroded away, the only evidence for the former presence of the track is the concentric circles developed by erosion of the undertracks. Scalebar p 10 cm. ers below the trackmaker’s foot is clearly visible pression of the middle digit, which was most deeply in the many different erosional states of the tracks. impressed into the substrate, is preserved. A large In some tracks, the sedimentary infilling is eroded set of concentric circles of deformation originating away, partly exposing the bottom of the true track from the tip of the impression of the middle digit and features such as impressions of digital pads. forms a 70-cm-long zone of deformed layers on the Interestingly, a curious feature occurs in the sedi- surface (fig. 5C). In some instances, the original ment around tracks eroded to this depth: a series tracked surface has been eroded to such an extent of apparently concentric circles are present around that the true track is completely eroded away. Sur- the track up to 20 cm away from the true track (fig. rounding the former position of the true track, a 5B). Close to the perimeter of the track, the curves prominent series of concentric circles of deformed follow the shape of the tracks; that is, the divisions layers extends for 20–30 cm outward from the po- between the digits are reflected in the concentric sition of the true track, forming an elliptical zone deformations as well. Tracks eroded to a depth of deformation approximately 70 cm in greatest di- where almost the whole true track is eroded away ameter (fig. 5D). show a distinct morphology because only the im- Large areas of the exposed flat-bedded deposits 380 J. MILA` N AND D. B. LOOPE are eroded to the same level today. This allows the consisting of a broad zone of concentric circles at same trackway to be followed for long distances in the present-day surface (fig. 6C). the same erosional state. Tracks still containing the One particular trackway consisting of four tracks darker-colored sedimentary infilling can be found is preserved on an erosional slope where the flat- as long consecutive trackways, apparently because bedded deposits have been differentially eroded the infilling is slightly more resistant to erosion down about 20 cm over a 4-m interval, thereby cut- than the surrounding sediment (fig. 6A). Another ting through several individual thin sand sheets. long trackway exists in a more advanced erosional This has caused each consecutive track in the state where the true tracks have been almost com- trackway to be present in a more eroded state than pletely eroded. Erosion has exposed the subjacent the previous one (fig. 7). The first track is preserved deformed layers, and the trackway consists of with the darker-colored infilling still in place and barely recognizable true tracks, each with a well- has further preserved an elongated metatarsus im- developed set of concentric deformations extending pression. The second track is well defined, has pre- outward (fig. 6B). The most curious trackway found served impressions of the individual digits and at the locality was an extensive trackway where all claws, and is partly infilled by darker sediment. The the true tracks are eroded totally away, but the third track is present only as an undertrack and trackway is still clearly visible and could without consists of a series of concentric circles in the sed- a problem be traced for about 40 m, because each iment. The fourth track is only vaguely recogniz- track is now represented by an eroded undertrack able as a faint zone of concentric deformations in

Figure 6. Theropod trackways in different preservation states. A, Long trackway preserved with each track still filled with the darker-colored sedimentary infilling. B, Trackway where each track is partly eroded away. Notice the prominent concentric circles of deformation surrounding the vestiges of the true tracks. C, Trackway where the true tracks are totally eroded away. Only the erosional cut through the concentric circles of the undertracks reveals the former presence of the trackway. This “undertrackway” can be traced all the way to the person in the background of the picture. Journal of Geology THEROPOD TRACKS 381 the sediment. By comparing the morphology track- sediments down the paleoslope of the deposit, by-track down through the different eroded states, forming a fan of deformation originating from the it is evident that the tracks were originally im- remains of the true track (fig. 8). pressed to about 10 cm depth (as indicated by the In areas with more extensive present-day erosion thickness of the infilled sediment) and that the un- of the flat-bedded deposits, such as on the steeply dertracks disturbed layers for a further 10 cm below sloping edges of the outcrop, several tracks can be the tracks (fig. 7C). found as oblique cuts through both the true tracks The above described tracks all originate from hor- and undertracks. The best cuts through the tracks izontally deposited sand sheet deposits; however, a were observed on an erosional surface that slopes number of localities reflect areas with sloping pa- approximately 30Њ. From these sections, the timing leorelief. Eroded tracks found on these sloping parts and formation of true tracks, undertracks, and sed- of the eolian deposits show peculiar morphology in imentary infilling can be reconstructed. The layer that they show a unidirectional deformation of the constituting the original tracking surface has been

Figure 7. Trackway segment found on a sloping erosional surface. A, Trackway consists of four tracks, each representing a different erosional layer. B, Sketch of the trackway showing the different morphology of the tracks. Track 1 has an elongated metatarsus impression and is infilled with darker-colored sediment. Track 2 is well defined with impressions of individual digits and claws and is partly infilled with darker-colored sediment. Notice the deformations of the layers surrounding the track. Tracks 3 and 4 are undertracks and are preserved only as concentric circles of deformation in the sediment. C, Hypothetical vertical section through the sloping surface with the tracks. The different sections through the tracks demonstrate that the tracks disturb a zone of 20-cm thickness in the thin-bedded eolian deposits. 382 J. MILA` N AND D. B. LOOPE

present the following scenarios for their formation and preservation. The preservational variants of the tracks emplaced on the horizontally laminated deposits (fig. 5) can all be explained by different degrees of erosion through a track emplaced and buried under the following circumstances. (1) The track is emplaced in a package of thin, horizontally laminated sand sheets (fig. 10A). (2) The weight of the trackmaker displaces the sediments surrounding the foot in a radial pattern, forming a rim of displaced material around the track. A stacked succession of successively shallower and broader undertracks is formed along the subjacent sand horizons (fig. 10B). (3) A darker-colored layer of sand covers the depositional surface (fig. 10C). (4) Present-day erosion through the layers exposes the tracks and the surrounding deformation structures, with the track and the sedimentary infilling still in place surrounded by concentric rings created by the erosion of the surrounding ring of displaced sediment (fig. 10D). If eroded to a deeper level, all of the true track and the sedimentary infilling are destroyed, and the only evidence of the track is a shallow, undefined undertrack in the middle and the concentric rings from the displaced sediment surrounding the track (fig. 10E). The formation and curious mode of preservation of the tracks from the sloping parts of the eolian deposits can be explained by the following scenario. Figure 8. Erosional cuts through tracks emplaced on a (1) The trackmaker emplaces the foot vertically on sloping tracking surface. When the tracks were formed, the weight of the trackmaker displaced a rim of sand on the downslope side of the trackmaker’s foot. In the present-day erosional cut, the rim of displaced sand is visible as a system of eccentric rings of deformation originating from the downslope side of the true track and spreading outward, indicating the original downslope side of the track. cut by the foot, and the layers have been bent downward, forming steep trackwalls. Undertracks are formed in the subjacent layers, radiating downward and outward below the track, forming a stack of successive wider and shallower undertracks subjacent to the true track (fig. 9). The sediment infilling the track is of a darker color than the sediment of the tracking surface and, when viewed in the present section through the track, appears as a subrectan- gular, structureless lens of darker sediment surrounded by the deformed, layered sediments. Figure 9. Erosional cut through a track infilled with sediment of a darker color than the sediment of the tracking surface. The view was generated by an approximately Preservational Models 30Њ cut through the layers. Notice how the coarser- grained, darker infilling of the tracks appears as a sub- The wide range of erosional states of the theropod rectangular, structureless unit surrounded by deformed, tracks from the flat-bedded deposits allows us to laminated eolian deposit. Knife handle is 10 cm long. Journal of Geology THEROPOD TRACKS 383

track). Undertracks are formed in the layers subjacent to the true track, becoming successively shallower and wider downward. The sideways deformation of the track due to the sloping of the tracking surface is also present in the undertracks (fig. 11B). (3) The tracking surface becomes covered by slightly darker-colored sediment, draping the contours of the track and adjacent deformation structures (fig. 11C). (4) Exposed to the present-day erosion, the track appears as a series of eccentric rings of deformation, originating in the eroded remains of the infilling of the true track (fig. 11D).

Discussion The tracks and undertracks of large theropods found in the Entrada Sandstone at the studied locality come in a wide range of morphologies, mainly as a result of present-day erosion that has exposed the track-bearing surfaces at different depths. It is reasonable to believe that the well- defined tridactyl footprints with preserved impressions of claws and individual digital pads are indeed true tracks (fig. 4), since experimental work with track and undertrack formation has shown that these anatomical details disappear in undertracks, while the general shape of the foot remains recognizable to a considerable depth (Manning 2004; Mila`n and Bromley 2006). The impression of the metatarsus in some of the larger theropod tracks is evidence of a semiplantigrade stance adopted by the trackmaker. Planti- grade theropod tracks are occasionally encountered Figure 10. Schematic model for the formation of the as ichnofossils and can be the result of an actual various erosional forms of the tracks found on the hor- plantigrade walk of the trackmaker (Kuban 1989). izontally deposited eolian deposits. A, Foot of the track- But partial impressions of the metatarsus can also maker is emplaced on the laminated deposits. B, Weight occur if the trackmaker walks in soft sediment and of the trackmaker displaces the sandsheet laminae out- the foot sinks to a considerable depth (Gatesy et ward around the foot, creating a circular marginal ridge of displaced material. Undertracks are formed along the al. 1999). subjacent horizons, down to at least 10 cm below the The tracks from the flat-bedded deposits appear track (cf. fig. 7). C, Track is filled with darker-colored to be emplaced in relatively damp sand, which al- sediment. D, Present-day erosion has exposed the track lowed extensive deformation to occur in the layers with the sedimentary infilling still in place. The erosion below and around the track. The laminae in and of the surrounding marginal ring of displaced material subjacent to the theropod tracks from the Entrada creates the pattern of concentric rings around the tracks Sandstone are commonly fractured (brecciated; figs. (cf. fig. 5B). E, With the true track totally eroded away, 4A,5C), which is a typical type of deformation oc- the only hint of the track is now the shallow undertrack curring in moist sand. Similar tracks from the Early and the concentric rings of the eroded marginal ridge of Jurassic Navajo Sandstone are interpreted as wet displaced material (cf. fig. 5C,5D). season tracks, in contrast to dry season tracks that preserve evidence of dry eolian grainflows triggered the sloping surface (fig. 11A). (2) The sediment dis- by the dinosaurs’ feet (Loope 2006b). Experiments placed by the trackmakers foot forms a rim on the with track formation in different horizontal sub- downslope side of the footprint (contrary to the strates, including dry sand, damp sand, and wet condition on the horizontal surface where the sed- sand, demonstrate that totally dry sand is a bad iment is uniformly displaced radially around the medium to preserve true tracks because the track- 384 J. MILA` N AND D. B. LOOPE

walls immediately collapse, leaving only the vague outline of the tracks left with little potential for fossilization. Tracks emplaced in damp sand, however, are well defined, with clear impressions of anatomical details such as digital pads, claws, and even, in some cases, skin texture. Wet sand tends to flow together and collapse some time after withdrawal of the foot, making it a bad medium for track preservation (Mila`n 2006). Even in cases where the true tracks are totally eroded away, the general trackway pattern is still recognizable because the present erosional cut through the undertracks makes the vestiges of the individual tracks appear as a series of concentric ring structures in the sand (fig. 6C). However, even though the general pattern of the trackway is still recognizable, the information about the trackmaker that can be extracted from the trackway is limited. Where the true tracks are fully eroded away and only the undertracks are present, it is still possible to calculate the trackway parameters such as stride length, pace length, and pace angulation. In this case, the measurements must be taken from a fixed point in each track, such as the middle of the deformation structures. On the other hand, attempts to estimate the progression speed from trackways consisting of undertracks should be avoided, since one of the parameters needed to calculate the progression speed is the hip height, which is a multiple of the trackmaker’s foot length (Alexander 1976; Thulborn 1990). Extreme caution should be taken when trying to estimate the foot length of an animal from trackways where only the undertracks are preserved, since the apparent dimensions of the track become larger downward at each subjacent layer because the weight of the trackmaker’s foot is transferred radially outward into the sediment (Allen 1997; Mila`n and Bromley 2006). Any application of the (apparently larger) foot length derived from an undertrack will give a higher estimated hip height and thus a relatively shorter stride length, resulting in a slower speed estimate.

Figure 11. Schematic model for the formation of the Conclusions tracks found on the sloping parts of the eolian deposits. A, Trackmaker emplaces the foot vertically on the slop- The Middle Jurassic Entrada Sandstone contains nu- ing surface. B, Trackmaker’s foot creates a rim of dis- merous dinosaur tracks, dominated by medium to placed material on the downslope part of the track. Un- large-sized theropod tracks. The tracks appear in eo- dertracks are formed in the layers subjacent to the foot. lian sand sheets interbedded with thin sets of eolian C, Surface and track are covered with sediment of a cross-strata that lie above the deposits of large-scale darker color. D, Present-day erosion of the surface has dunes. Present-day erosion has exposed the tracks exposed the tracks at various levels, with the most prom- in a wide range of preservational states, ranging from inent feature being the pattern of elliptical concentric well preserved, with the sedimentary infilling of the rings created by the erosion through the rim of displaced material from the downslope side of the track (cf. fig. 8). true tracks still in place, to deeply eroded, with the Journal of Geology THEROPOD TRACKS 385 only evidence of the animal’s passage being the cir- When studying eolian sediments, special atten- cular structures created by erosion through the un- tion should be paid to horizons with deformation dertracks. The study of the different eroded states of structures. The wide range of erosional states ex- the tracks and trackways shows that each of the hibited by the studied tracks, ranging from well- large theropod tracks has generated a zone of dis- defined, easily recognizable tracks to tracks eroded turbance up to 20 cm wide in the thin-bedded de- to a nearly unrecognizable state, demonstrate that posits, including the true track, the sedimentary in- what might look like random deformation struc- filling, and the extensive set of undertracks. tures in the sand layers might very well be the The many morphological variants of the thero- eroded remains of vertebrate tracks. pod tracks found in the Entrada Sandstone due to extensive undertrack formation and subsequent ACKNOWLEDGMENTS erosion to different levels demonstrate that great care should be taken when describing fossil foot- The research of J. Mila`n is financed by a PhD grant prints that have been exposed to subaerial erosion, from the Faculty of Natural Science, University of because the shape, dimensions, and general ap- Copenhagen. D. B. Loope’s work was funded by a pearance of the footprint become seriously altered grant from the National Science Foundation by erosion. Such erosion could take place on the (EAR02-07893). We are grateful to M. G. Lockley modern outcrop, as in the present case, or in the and R. Rogers, who provided critical and construc- ancient setting before final burial. tive reviews of the manuscript.

REFERENCES CITED

Alexander, R. M. 1976. Estimates of speeds of dinosaurs. western United States and other important vertebrate Nature 261:129–130. trackways from Grand Staircase–Escalante National Allen, J. R. L. 1989. Fossil vertebrate tracks and indenter Monument, Utah. Ichnos 7:169–181. mechanics. J. Geol. Soc. Lond. 146:600–602. Gatesy, S. M. 2001. Skin impressions of Triassic thero- ———. 1997. Subfossil mammalian tracks (Flandrian) in pods as records of foot movement. Bull. Mus. Comp. the Severen Estuary, S. W. Britain: mechanics of for- Zool. 156:137–149. mation, preservation and distribution. Philos. Trans. ———. 2003. Direct and indirect track features: what R. Soc. B 352:481–518. sediment did a dinosaur touch? Ichnos 10:91–98. Brand, L. R. 1996. Variations in salamander trackways Gatesy, S. M.; Middleton, M. K.; Jenkins, F. A.; and Shu- resulting from substrate differences. J. Paleontol. 70: bin, N. H. 1999. Three-dimensional preservation of 1004–1010. foot movements in Triassic theropod dinosaurs. Na- Breithaupt, B. H.; Matthews, N. A.; and Noble, T. A. ture 399:141–144. 2004. An integrated approach to three-dimensional Gatesy, S. M.; Shubin, N. H.; and Jenkins, F. A. 2005. data collection at dinosaur tracksites in the Rocky Anaglyph stereo imagining of dinosaur track mor- Mountain West. Ichnos 11:11–26. phology and microtopography. Palaeontol. Electronica Bromley, R. G. 2001. Tetrapod tracks deeply set in un- 8:10A. suitable substrates: recent musk oxen in ﬂuid earth Kuban, G. J. 1989. Elongate dinosaur tracks. In Gillette, (East Greenland) and Pleistocene caprines in Aeolian D. G., and Lockley, M. G., eds. Dinosaur tracks and sand (Mallorca). Bull. Geol. Soc. Den. 48:209–215. traces. Cambridge, Cambridge University Press, p. 57– Brown, T., Jr. 1999. The science and art of tracking. New 72. York, Berkley, 219 p. Lockley, M. 1991. Tracking dinosaurs. Cambridge, Cam- Diedrich, C. 2002. Vertebrate track bed stratigraphy at bridge University Press, 238 p. new megatrack sites in the Upper Wellenkalk Member ———. 1997. The paleoecological and paleoenviromental and orbicularis Member (Muschelkalk, Middle Tri- utility of dinosaur tracks. In Farlow, J. O., and Brett- assic) in carbonate tidal ﬂat environments of the west- Surman, M. K., eds. The complete dinosaur. Bloom- ern Germanic Basin. Palaeogeogr. Palaeoclimatol. Pa- ington, Indiana University Press, p. 554–578. laeoecol. 183:185–208. Lockley, M., and Hunt, A. P. 1995. Dinosaur tracks and Forno´s, J. J.; Bromley, R. G.; Clemmensen, L. B.; and other fossil footprint of the Western United States. Rodriguez-Perea, A. 2002. Tracks and trackways of New York, Columbia University Press, 338 p. Myotragus balearicus Bate (Artiodactyla, Caprinae) in Loope, D. B. 2005. Abundant traces of sand-swimming Pleistocene aeolianites from Mallorca (Balearic Is- vertebrates preserved high on the slopes of giant Ju- lands, western Mediterranean). Palaeogeogr. Palaeo- rassic dunes. Geol. Soc. Am. Abstr. Program 7:339. climatol. Palaeoecol. 180:277–313. ———. 2006a. Burrows dug by large vertebrates into rain- Foster, J. R.; Hamblin, A. H.; and Lockley, M. G. 2000. moistened, Middle Jurassic sand dunes. J. Geol. 114: The oldest evidence of a sauropod dinosaur in the 753–762. 386 J. MILA` N AND D. B. LOOPE

———. 2006b. Dry-season tracks in dinosaur-triggered maius novaehollandiae) tracks, reflecting differences grainflows. Palaios 21:132–142. in walking pattern and substrate consistency: ichno- Loope, D. B., and Rowe, C. M. 2003. Long-lived pluvial taxonomical implications. Palaeontology 49:405–420. episodes during deposition of the Navajo Sandstone. Mila`n, J., and Bromley, R. G. 2006. True tracks, under- J. Geol. 111:223–232. tracks and eroded tracks, experimental work with tet- Loope, D. B., and Simpson, E. L. 1992. Significance of rapod tracks in laboratory and field. Palaeogeogr. Pa- thin sets of eolian cross-strata. J. Sediment. Res. 62: laeoclimatol. Palaeoecol. 231:253–264. 849–859. Steiner, M. B. 2003. A cratonic Middle Jurassic paleopole: Loope, D. B.; Steiner, M. B.; Rowe, C. M.; and Lancaster, Callovian-Oxfordian stillstand (J-2 cusp), rotation of N. 2004. Tropical westerlies over Pangeaean sand seas. the Colorado Plateau, and Jurassic North American Sedimentology 51:315–322. apparent polar wander. Tectonics 22:1020. Manning, P. 2004. A new approach to the analysis and interpretation of tracks: examples from the Dinosau- Thompson, A. E., and Stokes, W. L. 1970. Stratigraphy ria. In McIlroy, D., ed. The application of ichnology of the San Rafael Group, southwest and south central to palaeoenviromental and stratigraphic analysis. Utah. Utah Geol. Mineral. Surv. Bull. 87:1–53. Geol. Soc. Lond. Spec. Publ. 228:93–123. Thulborn, T. 1990. Dinosaur tracks. London, Chapman Mila`n, J. 2006. Variations in the morphology of emu (Dro- & Hall, 410 p. Jesper Milàn – Vertebrate Ichnology Appendix ______

Paper 6 Dinosaur footprints from the Middle Jurassic Bagå Formation, Bornholm, Denmark. Jesper Milàn & Richard G. Bromley Bulletin of the Geological Society of Denmark, 2005, v. 52, p. 7–15.

Dinosaur footprints from the Middle Jurassic Bagå Formation, Bornholm, Denmark

JESPER MILÀN & RICHARD G. BROMLEY

Milàn, J. & Bromley, R.G. 2005–11–15: Dinosaur footprints from the Middle Jurassic Bagå Formation, Bornholm, Denmark. Bulletin of the Geological Society of Denmark, Vol. 52, pp. 7–15, Copenhagen. © 2005 by the Geological Society of Denmark. ISSN 0011–6297.

Dinosaur footprints have been found preserved on sandstone blocks discarded from the flooded clay pit Pyritsøen, south of Hasle, Bornholm. The sandstone belongs to the Middle Jurassic Bagå Formation, but the exact horizon is not known. Palynological studies confirm that the sandstone blocks originate from the Bagå Formation. Two specimens were collected, one showing two footprints from a sauropod dinosaur having a foot length of 68 cm, and a small pentadactyl footprint, 26 cm long, interpreted as deriving from an armoured dinosaur. These are the first dinosaur footprints recorded from Denmark.

Key-words: Dinosaur footprints, sauropod, Middle Jurassic, Bornholm, Denmark.

Jesper Milàn [[email protected]], University of Copenhagen, Geological Institute, Øster Voldgade 10, DK- 1350 Copenhagen K, Denmark. Richard G. Bromley [[email protected]], University of Copenhagen, Geo- logical Institute, Øster Voldgade 10, DK-1350 Copenhagen K, Denmark.

Terrestrial Mesozoic sediments in Denmark are ex- scarce and consists of only a series of indeterminate posed along the west and south-west coasts of the dinosaurian vertebrae (Bölau 1954). All dinosaur Baltic island of Bornholm and in small inland quar- footprints from the Höganäs Formation have been ries (Fig. 1). Only few of them are now open as quar- described as footprints from theropod dinosaurs, of rying has ceased. the ichnogenus Grallator, but at least one footprint The Lower Cretaceous sediments from the Nyker from the formation, (on display at the Geological Group at Robbedale have revealed a diverse verte- Museum, Copenhagen (MGUH 27219)), represents brate fauna comprising fish, turtles, crocodiles and the footprint of an early Ornithischian dinosaur lizards (Bonde 2004), and a multituberculate mam- (Milàn & Gierlinski 2004). mal (Lindgren et al. 2004). Remains of marine rep- The first hints that dinosaur footprints might be tiles have been found in the Lower Jurassic Hasle present on Bornholm, came when studying the de- Formation (Milàn & Bonde 2001). In the year 2000, a tailed lithological logs by Gravesen et al. (1982) that tooth from a small carnivorous dinosaur, Dromaeo- depict horizons showing peculiar steep-walled de- sauroides bornholmensis, was found in the Lower Cre- formation structures in the Tornhøj Member of the taceous Jydegård Formation, at Robbedale (Christian- Lower Cretaceous Jydegård Formation, exposed at sen & Bonde 2003), and two years later, a probable Skrædderbakken sand pit at Robbedale. Such steep- sauropod tooth was found at the same locality (Bonde walled deformation structures are in many cases & Christiansen 2003). cross sections through vertebrate footprints (Loope Over many years, especially in connection with 1986). Recent attempts by the authors to locate these the mining of brown coal from the Höganäs Forma- horizons were unsuccessful, as the quarry has been tion, numerous Late Triassic and Early Jurassic di- filled with water and the walls partly levelled. nosaur footprints and trackways have been found in The lithology and depositional environment of the southern Sweden (Bölau 1952; Pleijel 1975; Ahlberg Bagå Formation were recognized as a highly likely & Siveson 1991; Gierlinski & Ahlberg 1994). Foot- setting for the preservation of tetrapod footprints and prints of dinosaurs are abundant in the Höganäs the authors made a search for such trace fossils in Formation, whereas the record of body fossils is the summers of 2002 and 2003. Poorly preserved foot-

Milàn & Bromley: Dinosaur footprints from Bornholm · 7 Fig. 1. A. Bornholm is situated in the Baltic Sea south of Sweden (KMS G15-Ø). B. Map of Bornholm with the outcrops of Mesozoic sediments indicated with grey, based on Jensen & Hamann (1989). Asteriks marks the location of Pyritsøen, between Hasle and Rønne. C. Dinosaur footprint locality at Pyritsøen, sited near the villages of Muleby and Sorthat. Asteriks marks the position where the specimens were collected (UTM koordinates 55.15N 14.70E).

prints were indeed found, but they were not con- and transported to the Geological Museum in Co- vincingly distinguishable from slump structures or penhagen for closer study. Dinosaur footprints had results of local concretionary diagenesis. In 2004, not been found previously on Bornholm and were however, the senior author found unquestionable unfamiliar to the local geologists, and remained un- footprints on the beach beside the old clay-pit Bagå noticed. The aim of this paper is to describe the newly Graven. This pit is now flooded and has been re- found dinosaur footprints from the Bagå Formation, named Pyritsøen (Pyrite Lake; Fig. 1). The pit had their preservation and taphonomy and to place them yielded clay from the Middle Jurassic Bagå Forma- in context with other finds of dinosaurs from Scania. tion for the manufacture of hard-fired floor-tiles, bricks, pipes etc., by the Hasle Clinker Factory, which ceased production twenty years ago. The footprints occurred in the sandstone beds that Geological setting were unusable by the quarrymen. These sandstone beds were broken up and discarded by throwing The Bagå Formation (Gravesen et al. 1982) includes them onto the adjacent beach, where they have been the coal-bearing clays and sands in the Rønne-Hasle lying for over two decades. On discovery of the foot- Fault Block of southwest Bornholm. These beds had prints, the best specimens were collected by crane been traditionally named the Levka, Sorthat and Bagå

8 · Bulletin of the Geological Society of Denmark Fig. 3. The small footprint (MGUH 27754) seen in oblique frontal angle, as it was found on the beach. The footprint is preserved as a natural sandstone cast. In this view the digit impressions appear as grooves running down the front of the footprint. Scale on knife handle in centimetres.

Fig. 2. Processes leading to the preservation of dinosaur footprints in the Bagå Formation. A. Deposition of laminated clay. B. Trampling by dinosaurs causing intense disturbance of the ↓ Fig. 4. A. The small footprint (MGUH 22754) seen from laminated clay layers. C. Deposition of fluvial sand filling and above. The footprint is almost symmetrical along the length protecting the biogenic topography of the clay surface. D. Break- axis. Knife handle represents 10 cm. B. Interpretative up of the sandstone by quarrymen and dumping on the near- drawing of the small footprint, with measurements in by beach. centimetres. The footprint is 26 cm long and 19.5 cm wide. The oval front part of the footprint, showing the five digit impressions, is 13–15 cm deep, indicated by dark grey. The proximal part of the footprint, indicated by light grey, shallows proximally to a depth of 6 cm.

Milàn & Bromley: Dinosaur footprints from Bornholm · 9 (2003), deposition took place in lakes and swamps, small crevasse channels, lacustrine deltas and fluvi- atile channels. Today, quarrying has ceased and the Bagå pit and other, smaller clay pits are filled with water, hindering detailed new studies of the footprint-bearing layers (Gravesen 1996).

The footprints Two large blocks were collected; one containing a small well-defined footprint and several eroded footprint-like structures, and one bearing two large footprints, c. 70 cm long. Additionally, two blocks showing footprint-like deformation structures were collected, but the erosion of these blocks was too extensive for the structures to be confirmed as footprints. The two blocks with well-preserved footprints are curated by the Geological Museum, Copenhagen (MGUH 27754 and MGUH 27755) and are on display at the museum NaturBornholm, Aakirkeby, Born- holm. Pinched between the two large footprints, some of the original clay in which the footprints were emplaced has escaped weathering. A palynological Fig. 5. Latex mould of the small footprint. This allows the foot- study, conducted by Eva Koppelhus (personal com- print to be seen in negative relief, exactly as when it was orig- munication 2004), confirmed that the palynological inally impressed into the tracking surface. Notice how the five assemblage identified in the clay between the foot- toes have been dragged down through the sediment and formed prints, belongs to the assemblage known from the ridges in the track wall at the front of the footprint. Scale bar Bagå Formation (Koppelhus & Nielsen 1994), and 10 cm. thus confirming the origin of the blocks.

The small footprint (MGUH 27754). The small foot- beds (Gry 1969). The Bagå Formation was revised, print is preserved as a natural cast on the underside partly on the basis of well-core material, by Michelsen of a large block of sandstone (Fig. 2). The footprint is et al. (2003). In this revision, a series of beds showing 26 cm long and 19.5 cm wide (Fig. 3) and the distal evidence of marine influence was separated as the part with the digit impressions is impressed to a Sorthat Formation, thereby leaving the Bagå Forma- depth of 15 cm. The distal part shows impressions of tion entirely composed of nonmarine sediments. the digits. The proximal part deepens gradually to a Marine trace fossils from the Sorthat Formation were depth of 6 cm at the ‘heel’ of the footprint. recently studied by Bromley & Uchman (2003). The The footprint is pentadactyl, showing short but age of the Bagå Formation is Middle Jurassic, Bajo- well-defined impressions of five digits. It is close to cian–Bathonian (Gry 1969; Koppelhus & Nielsen 1994). symmetrical along the medial axis, and thus it is not The type section of the Bagå Formation is the Bagå possible to determine whether it is a left or a right Graven of the Hasle Clinker Factory at Sorthat (Grav- footprint (Fig. 4A). The impressions of the three mid- ersen et al. 1982; Michelsen et al. 2003) (Fig. 1). The dle digits are almost of equal length, and the middle formation includes thick, grey clay units, dark to digit, digit III, is only 1.0 cm and 0.8 cm longer than black coaly clays containing rootlets, coal beds, and the adjacent digits. The impressions of the two outer medium- to fine-grained, cross-bedded or laminat- digits are 4.3 and 6.0 cm behind the neighbouring ed sand beds. In the upper part, poorly sorted, mud- digit impressions (Fig. 4B). Recent erosion has round- dy and pebbly sand beds locally contain boulders of ed the tips of the digits, preventing exact measure- weathered granite. ments of their lengths. The footprint is preserved in According to Gravesen et al. (1982), Koppelhus & positive relief as a natural cast of the dinosaur´s foot. Nielsen (1994), Nielsen (1995) and Michelsen et al. A cast was made of the surface of the block in order

10 · Bulletin of the Geological Society of Denmark Fig. 6. A. The sandstone block containing the two large, sauropod footprints (MGUH 27755). The left footprint is the better preserved and the impressions of four short rounded digits are visible in the lower part of the footprint. The handle of the knife is 10 cm. B. Interpretative drawing of the block with the two footprints including measurements in centimetres. The bottom of the footprints is shaded in dark grey and the parts forming the track walls to the tracking surface are light grey. Digit numbers are indicated by Roman numbers.

to obtain a view of the footprint in negative relief, as wide, is oval to pear-shaped in outline and includes it appeared when it was impressed in the tracking the impressions of four short blunt digits. The other surface. It is apparent by doing this, that the foot had less well-preserved footprint measures 68 cm in sunk to a considerable depth, and the short toes length and 48 cm in width (Fig. 6B). The cast of the formed vertical ridges and grooves in the mud as better preserved footprint varies in depth from 25 to they were pressed into and passed through the soft 29 cm, and that of the other footprint varies from 10 substrate (Fig. 5). to 20 cm. The bottoms of the two footprints are flat and appear to lie at the same level. The digit impres- The two large footprints (MGUH 27755). Two large and sions are preserved as ridges extending from the track- well-defined footprints are preserved as sandstone ing surface to the bottom of the footprint (Fig. 7). casts on the underside of the block (Fig. 6A). The better preserved footprint, 69 cm long and 45 cm

Milàn & Bromley: Dinosaur footprints from Bornholm · 11 Fig. 7. The better preserved sauropod footprint in frontal view. The impressions of the digits, indicated by Roman numbers, are preserved as ridges formed as they were dragged through the mud. Part of the mud from the tracking surface is preserved between the two footprints, indicated by finger. Palynological examination of this mud confirms the Middle Jurassic age for the footprints (Eva Koppelhus, personal communication 2004).

Discussion subsequently deformed. Gravitational slumping at this level would cause correlative deformation of This first report of dinosaur footprints from Denmark both clay and sand. This detailed topography and is important because a number of potential alterna- deformation scenario also negates the possibility of tive explanations for the origin of the structures may the structures having been created by modern ma- be considered. Alternative scenarios for the origin rine erosion or by concretionary diagenesis. of the structures are that the structures are random Furthermore, it shows that the sedimentary infill- slump structures, or that they are products of ero- ing of each of the footprints was passive filling of a sion by the sea, as the blocks were found on the beach. pre-existing depression. If the footprints had been A number of observations, however, disqualify this. emplaced during the deposition of the sand, the fill- Although the blocks containing the footprints have ing would have been convoluted or disturbed (Na- been exposed to erosion on the beach for years, clay don 1993, 2001). The palynological examination con- from the original tracking surface is still present be- firms the presence of Middle Jurassic clay in situ between the two large footprints (Fig. 7). The clay lay- tween the footprints, dismissing the argument that ers have been disturbed and display an almost verti- the footprints are the result of modern sea-erosion. cal lamination in the parts preserved, suggesting ex- Elsewhere on the beach from which the blocks treme deformation caused by compression by the were collected, there are blocks of Bagå sandstone trackmaker´s feet. Deposition of the sand post-dates that are not connected with the top surface of the the formation of the footprints, as the clay around clay. Some of these show cross-bedding (containing the footprint is highly deformed, whereas the sand- wood clasts) that is undisturbed, whereas others stone infilling is horizontally bedded and lacks indi- show extreme deformation of the bedding. In the light cations of subsequent disturbance. Thus the config- of the collected material described here, this de- uration of the clay-sand interface, where the surface formed bedding within the sandstone probably be- layer of the clay contains large concave structures longs to other dinosaurian trampling events. directly related to deformation of the lamination, clearly demonstrates a tetrapod trample-ground in soft substrate. The preservation of the sculptured clay Interpretation surface topography demonstrates that it was overlain a short time after trampling by a unit of sand, The large footprint is 69 cm long, pear-shaped in which protected its highly sculptured topography. outline and have impressions of four short rounded The sand shows primary lamination that has not been digits, which is consistent with hind-foot prints from

12 · Bulletin of the Geological Society of Denmark dy. The sauropod forefoot (manus) is pentadactyl, with short reduced digits encapsulated by tissue to form hoof-like units, whereas only digit I, the pollex, is separate and bears a claw. In the most derived sauropods, the titanosaurs, even this claw is reduced (Bonnan 2003). The hindfeet (pes) of sauropods are pentadactyl and the digits are short and blunt. At least the inner two to three digits bear blunt but prominent claws (Christiansen 1997a, b). Well-preserved sauropod trackways from around the world all show that the impressions of the inner three digits are the most prominent in the pes prints, and that the outer two digits rarely leave impressions. The impressions of the manus are hoof-shaped, sometimes showing impressions of the inwardly directed pollex claw (Meyer et al. 1994; Santos et al. 1994) (Fig. 8). Manus prints have not been identified in the material from Bornholm. The small footprint is well preserved, showing clear impressions of five short, claw-bearing digits. However, the symmetrical configuration of the foot makes it hard to clearly assign it to any known dinosaurian ichnogenus. Furthermore, the symmetrical configuration of the digit impressions prevents it from even being identified as a right or left foot. The Fig. 8. Tracing of part of a sauropod trackway from the Middle known pentadactyl pes configurations among Mid- Jurassic Galinha tracksite near Fatima, Portugal. The manus dle Jurassic dinosaurs are all asymmetrical like the imprints are hoof-like, having only one inward-facing claw impression. The pes prints have impressions of three to four sauropod footprints, in which digits I and II are the short blunt digits. This pes morphology is consistent with the most prominent. A possibility is that the footprint is better preserved of the two sauropod footprints from Born- a manus print from an early thyreophoraen dinosaur, holm (Fig. 6). Figure modified from Santos et al. (1994). as these animals possessed pentadactyl mani, symmetrical around the middle digits (McCrea et al. 2001). However, the known manus prints from thyreophorean dinosaurs are crescent- to hoof-shaped sauropod dinosaurs. This better preserved of the two and broader than long (Romano & Whyte 2003). The footprints is the impression of a left foot. The other small footprint is longer than broad but is deeply footprint lacks any distinct digit impressions, but impressed into the substrate. When an animal walks, instead the bottom seems to be divided into a num- its feet are not merely impressed and lifted vertical- ber of ‘pads’, similar to what is observed in Middle ly. If the foot is impressed deeply into a substrate, Jurassic sauropod footprints from the Yorkshire coast parts of the forward movement of the limb will be (Romano & Whyte 2003). As distinct digit impres- captured in the footprint. This causes a prominent sions are lacking in the footprint, it is not possible to elongation of the footprint, as demonstrated with determine if it represents the impression of the right Late Triassic theropod footprints by Gatesy et al. foot, connected to the left footprint beside it, i.e. a (1999) and Pleistocene artiodactyl footprints (Fornós pair, or if it is unrelated. et al. 2002). Additional footprints, and preferably The two footprints are impressed to different trackway segments preserving both manus and pes depths relative to the tracking surface, but the bot- impressions, are needed before the trackmaker re- toms of the two impressions are at the same level. sponsible for the small footprint can be identified This apparent difference in depth is due to the une- with certainty. ven and trampled topography of the tracking surface. The depth of the two footprints thus may represent the maximum depth the softness of the mud Correlation with previous finds allowed the feet to sink to, beneath which the mud was too rigid to allow further penetration by the feet. There is a wide temporal gap between the new foot- Sauropods have an extreme degree of heteropo- prints from the Bagå Formation, the footprints from

Milàn & Bromley: Dinosaur footprints from Bornholm · 13 the Höganäs Formation of southern Sweden and the Danish Summary two teeth previously found on Bornholm. The footprints from Sweden are from the Late Triassic to Early De første forstenede dinosaur fodspor er blevet fun- Jurassic and are roughly 40 million years older than det i sandstens blokke der er blevet brudt og dum- the new footprints from Bagå (Bölau 1952; Pleijel pet på stranden under lergravning i den gamle Has- 1975; Ahlberg & Siveson 1991; Gierlinski & Ahlberg le Klinkefabriks lergrav (Bagå Graven), syd for Has- 1994; Milàn & Gierlinski 2004). The two teeth found le, Bornholm (Fig. 1). Leret der blev gravet i lergra- in Robbedale are from the Early Cretaceous (Bonde ven tilhørte den midt Jurassiske Bagå formation. I & Christiansen 2003; Christiansen & Bonde 2003), and dag er lergravningen ophørt og graven er fyldt med thus around 30 million years younger than the Bagå vand og omdøbt til Pyritsøen. footprints. The Middle Jurassic age of the new foot- To sandstens blokke med dinosaur fodspor, samt prints from Bornholm, thereby fills in the 70 million yderligere to blokke med spor-lignende strukturer years gap between the Late Triassic records of dino- er blevet indsamlet og studeret. De fundne spor er saurs from Sweden to the Lower Cretaceous records bevaret som sandstensudfyldninger af de oprindeli- from Bornholm. To date, no skeletal remains of di- ge spor og fremstår derved i positivt relief (Fig. 2). nosaurs have been found in the Bagå Formation, but På den ene blok ses et lille spor på 26 cm længde this is not surprising as the preservation of footprints (Figs. 3–5) sandsynligvis fra en pansret dinosaur. Den and bones is controlled by two widely different anden blok indeholder to store, respektivt 69 og 68 taphonomic processes. cm lange spor fra en sauropod dinosaur. Det bedst bevarede af dem har aftryk af fire korte brede tæer (Figs. 6–8). En række sedimentologiske faktorer bekræfter at Conclusions de fundne strukturer virkelig er dinosaur spor og ikke slumpstrukturer, konkretionsdannelse eller eros- Exploitation of clay of the Middle Jurassic Bagå For- ions fænomener. Underfladen af sandstens blokke- mation from the clay-pit Bagå Graven south of Hasle, ne er formet efter en ujævn mudderflade. Dette mud- Bornholm, produced waste material in the form of der er bevaret in situ imellem de to store sauropod blocks of indurated sandstone layers. Large blocks spor, og laminationen i det er her stærkt forstyrret. of these were discarded from the clay-pit onto the Sandet der udfylder sporene derimod, er horison- nearby beach. talt lagdelt, hvilket viser at udfyldningen er sket ef- Two sandstone blocks containing well-preserved ter at mudderfladen er blevet deformeret, i dette til- hind-foot prints of sauropods and a smaller footprint, fælde af dinosaurer. Palynologiske undersøgelser af interpreted as the footprint of a small armoured di- leret fundet mellem sporene yderligere bekræfter at nosaur, were found among the discarded blocks. blokkene stammer fra Bagå Formationen og at spo- Study of these loose sandstone blocks revealed de- rene er dannet i Jura tiden, og derfor heller ikke kan formed bedding, representing dinosaur footprints være et resultat af erosion efter at sandstensblokke- preserved at a clay-sand interface. The tracking sur- ne var blevet tippet af på stranden. face was the top of a clay unit, that had been considerably deformed by dinosaur trampling to a depth of 20–30 cm. This surface was rapidly overlain by fluvial sand, thus preserving the footprints in con- Acknowledgements siderable detail. The precise horizon within the Bagå Formation has not yet been identified, as the quarry We are grateful to the Geological Institute, Universi- is now flooded, but palynological examination of clay ty of Copenhagen for covering the cost of transpor- between the footprints confirms a Middle Jurassic tation of the footprints. Steen Lennart Jacobsen, Geo- age, and that the blocks originate from the Bagå For- logical Museum Copenhagen, made the latex cast of mation. It is to be expected that further dinosaur foot- the surface of the block containing the small footprints can be found in the continental Mesozoic rocks print. Eva Koppelhus (Royal Tyrrell Museum, Cana- of Bornholm and in future sedimentological studies, da) is thanked for applying her expertise in Jurassic special attention should be paid to deformed bed- palynological analysis to the clay samples. Phillip J. ding planes. Currie, (Royal Tyrrell Museum, Canada) kindly com- mented on an early draft of the manuscript. Special thanks to Finn Hansen, NaturBornholm, for arrang- ing for the footprints to be placed on permanent display there. The manuscript benefited from the criti-

14 · Bulletin of the Geological Society of Denmark cal reviews of Mike Romano, (University of Sheffield) Jensen, J.B. & Hamann, N.E. 1989: Geological mapping of Mes- and Gregers Dam, (Danish Oil and Natural Gas). ozoic deposits along the eastern margin of the Rønne Graben, offshore Bornholm, Denmark. Bulletin of the Geo- logical Society of Denmark 37, 237–260. Koppelhus, E.B. & Nielsen, L.H. 1994: Palynostratigraphy and palaeoenvironments of the Lower to Middle Jurassic Bagå References Formation of Bornholm, Denmark. Palynology 18, 139–194. Lindgren J., Rees J., Siverson M. & Cuny G. 2004: The first Ahlberg, A. & Siverson, M. 1991: Lower Jurassic dinosaur foot- Mesozoic mammal from Scandinavia. GFF 126, 325–330. prints in Helsingborg, southern Sweden. Geologiska Före- Loope, D.B. 1986: Recognizing and utilizing vertebrate tracks ningens i Stockholm Förhandlingar 113, 339–340. in cross section: Cenozoic hoofprints from Nebraska. Pal- Bonde, N. 2004: An Early Cretaceous (Ryazanian) fauna of aios 1, 141–151. “Purbeck-Wealden” type at Robbedale, Bornholm, Den- McCrea, R.T., Lockley, M.G. & Meyer, C.A. 2001: Global distri- mark. In Arratia, G. & Tintori, A. (eds): Mesozoic Fishes 3 – bution of purported ankylosaur track occurrences. In Car- systematics, palaeoenvironments and biodiversity, 507–528. penter, K. (ed.): The Armoured Dinosaurs, 413–454. Indi- Verlag Dr. Friedrich Pfeil, München, Germany. ana University Press, Bloomington. Bonde, N. & Christiansen, P. 2003: New dinosaurs from Den- Meyer, C.A., Lockley, M.G., Robinson, J.W. & Santos, W.F. 1994: mark. Comptes Rendus Palevol 2, 13–26. A comparison of well-preserved sauropod tracks from the Bonnan, M.F. 2003: The evolution of manus shape in sauropod Late Jurassic of Portugal and the western United States: dinosaurs: implications for functional morphology, forelimb evidence and implications. Gaia 10, 57–64. orientation, and phylogeny. Journal of Vertebrate Paleon- Michelsen, O., Nielsen, L.H., Johannessen, P.N., Andsbjerg, J. tology 23, 595–613. & Surlyk, F. 2003: Jurassic lithostratigraphy and stratigraphic Bölau, E. 1952: Neue Fossilfünde aus dem Rhät Schonens und development onshore and offshore Denmark. In Ineson, J.R. ihre paläogeographisch-ökologische Auswertung. Geologi- & Surlyk, F. (eds): The Jurassic of Denmark and Greenland, ska Föreningens i Stockholm Förhandlingar 74, 44–50. 147–216. Geological Survey of Denmark and Greenland Bul- Bölau, E. 1954: The first finds of dinosaurian skeletal remains letin 1. in the Rhaetic-Liassic of N. W. Scania. Geologiska Förenin- Milàn, J. & Bonde N. 2001: Svaneøgler: nye fund på Bornholm. gens i Stockholm Förhandlingar 76, 501–502. Varv 2001 (4), 3–8. Bromley, R.G. & Uchman, A. 2003: Trace fossils from the Low- Milàn, J. & Gierlinski, G. 2004: A probable thyreophorean (Di- er and Middle Jurassic marginal marine deposits of the Sort- nosauria, Ornithischia) footprint from the Upper Triassic of hat Formation, Bornholm, Denmark. Bulletin of the Geolo- southern Sweden. Bulletin of the Geological Society of Den- gical Society of Denmark 52, 185–208. mark 51, 71–75. Christiansen, P. 1997a: Locomotion in sauropod dinosaurs. Gaia Nadon, G.C. 1993: The association of anastomosed fluvial de- 14, 45–75. posits and dinosaur tracks, eggs and nests: Implications for Christiansen, P. 1997b: Hindlimbs and feet. In Currie, P.J & the interpretation of floodplain environments and a possi- Padian, K. (eds): Encyclopedia of Dinosaurs, 320–328. Aca- ble survival strategy for ornithopods. Palaios 8, 31–44. demic Press, New York. Nadon, G.C. 2001: The impact of sedimentology on vertebrate Christiansen, P. & Bonde, N. 2003: The first dinosaur from Den- track studies. In Tanke, D.H, & Carpenter, K. (eds) Mesozo- mark. Neues Jahrbuch für Geologie und Paläontologie Ab- ic Vertebrate Life, 395–407. Indiana University Press, Bloom- handlungen 227, 287–299. ington. Fornós, J.J., Bromley, R.G., Clemmensen, L.B. & Rodríguez- Nielsen, L.H. 1995: Genetic stratigraphy of the Upper Triassic Perea, A. 2002: Tracks and trackways of Myotragus baleari- – Middle Jurassic deposits of the Danish Basin and Fennos- cus Bate (Artiodactyla, Caprinae) in Pleistocene aeolianites candian Border Zone 2, 3. 162 pp. Unpublished Ph.D. the- from Mallorca (Balearic Islands, Western Mediterranean). sis, University of Copenhagen, Denmark. Palaeogeography, Palaeoclimatology, Palaeoecology 180, Pleijel, C. 1975: Nya dinosauriefotspår från Skånes Rät-Lias. 277–313. Fauna och Flora 3, 116–120. Gatesy, S.M., Middleton, K.M., Jenkins, F.A. Jr. & Shubin, N.H. Romano, M. & Whyte, M.A. 2003: Jurassic dinosaur tracks and 1999: Three-dimensional preservation of foot movements trackways of the Cleveland Basin, Yorkshire: preservation, in Triassic theropod dinosaurs. Nature 399, 141–144. diversity and distribution. Proceedings of the Yorkshire Gierlinski, G. & Ahlberg, A. 1994: Late Triassic and Early Ju- Geological Society 54, 185–215. rassic dinosaur footprints in the Höganäs Formation of Santos, V.F., Lockley, M.G., Meyer, C.A., Carvalho, J. Galopim southern Sweden. Ichnos 3, 99–105. de Carvalho, A.M. & Moratalla, J.J. 1994: A new sauropod Gravesen, P., Rolle, F. & Surlyk, F. 1982: Lithostratigraphy and tracksite from the Middle Jurassic of Portugal. Gaia 10, 5– sedimentary evolution of the Triassic, Jurassic and Lower 13. Cretaceous of Bornholm, Denmark. Danmarks Geologiske Undersøgelse Serie B 7, 51 pp. Gravesen, P. 1996. Geologisk set Bornholm, en beskrivelse af områder af national geologisk interesse. Miljø- og En- ergiministeriet Skov- og Naturstyrelsen, Geografforlaget, Brenderup, 209 pp. Gry, H. 1969: Megaspores from the Jurassic of the island of Bornholm, Denmark. Meddelelser fra Dansk Geologisk Fore- ning 19, 69–89.

Milàn & Bromley: Dinosaur footprints from Bornholm · 15 16 · Bulletin of the Geological Society of Denmark Jesper Milàn – Vertebrate Ichnology Appendix ______

Paper 7 Ichnological evidence for giant ornithopod dinosaurs in the Late Jurassic Lourinhã Formation, Portugal. Octávio Mateus & Jesper Milàn Submitted to Oryctos.

Mateus & Milàn – A giant ornithopod track ______

Ichnological evidence for giant ornithopod dinosaurs in the Late Jurassic Lourinhã Formation, Portugal

Evidence ichnologique de la présence d’un ornithopode géant dans la Formation Lourinhã du Jurassique supérieur du Portugal

1 2* Octávio MATEUS & Jesper MILÀN

1Museu da Lourinhã, Rua João Luis de Moura, 2530-157 Lourinhã, Portugal, e-mail: [email protected]; Universidade Nova de Lisboa, Faculdades de Ciências e Tecnologia, Centro de Estudos Geológicos, Monte de Caparica, Portugal. 2Geological Institute, University of Copenhagen, Øster Voldgade 10, DK-1350 Copenhagen K, Denmark, e-mail: [email protected]

Abstract The Late Jurassic Lourinhã Formation (Lusitanian Basin, Portugal) contains a diverse dinosaur fauna comprising theropods, sauropods, stegosaurs, ankylosaurs and several genera of ornithopods. The sedimentology in the area favours preservation of tracksways, and tracks from most of the dinosaurs represented by skeletal remains are present in the area. During fieldwork in the summer of 2003 a new, large, tridactyl track was found at the beach of Vale Frades, approximately 6 km north of Lourinhã, central west Portugal. The track was found together with a stegosaur track on a clay bed exposed within the tidal zone. Owing to the immediate danger of erosion, the track was collected and is now on display at Museu da Lourinhã (ML 1000). The track is 70 cm long and 69 cm wide, the toes are short and broad, with indications of short blunt claws, and there is a high angle of divarication between the outer digits. Multivariate analysis of the dimensions of the track identifies it as deriving from an ornithopod dinosaur with an estimated hip height around three metres. Although very large ornithopods are known from the Cretaceous, the largest known Jurassic ornithopod is Camptosaurus from USA, and the largest known from Portugal is the camptosaurid Draconyx loureiroi. Neither of these reached the body size suggested by the new track. So far the track described herein is the only evidence for a Jurassic ornithopod of that size.

Keywords: Portugal, Dinosaur, Footprint, Ornithopod, Late Jurassic, Giant.

1 Mateus & Milàn – A giant ornithopod track ______

Résumé La Formation Lourinhã au Portugal, d’âge Jurassique supérieur, a livré une faune variée de dinosaures, comprenant des théropodes, des sauropodes, des stégosaures, des ankylosaures et plusieurs genres d’ornithopodes. La sédimentologie de cette zone favorise la préservation d’empreintes de pas, et les empreintes de la plupart des dinosaures représentés par des restes osseux ont été retrouvées. Au cours de la campagne de fouille de l’été 2003, une grande empreinte tridactyle a été trouvée sur la plage de Vale Frades, à environ 6 km au Nord de Lourinhã. Cette empreinte a été trouvée au même endroit qu’une empreinte de stégosaure, sur un banc d’argile exposé dans la zone de battement des marées. A cause du risque immédiat représenté par l’érosion, l’empreinte a été prélevée, et est désormais exposée au musée de Lourhinã sous le numéro ML 1000. L’empreinte a une longueur de 70 cm et une largeur de 69 cm. Les doigts sont courts et larges et possédaient probablement des griffes courtes et émoussées. L’angle interdigital II-IV est important. Une analyse multivariée des dimensions des empreintes de pas indique qu'elles sont attribuables à un dinosaure ornithopode dont la hauteur au bassin était d'environ trois mètres. De très grands ornithopodes sont connus durant le Crétacé, mais le plus grand ornithopode du Jurassique est Camptosaurus, trouvé aux USA, et le plus grand ornithopode connu au Portugal est un autre Camptosauridés, Draconyx loureiroi. Aucun de ces deux derniers n’atteignait la taille suggérée par l’empreinte décrite dans le présent travail, qui demeure pour l’instant la seule indication de la présence d’un ornithopode de très grande taille dans le Jurassique.

Mots clés: Ornithopode, empreinte de pas, Jurassique supérieur

Resumo em português O Jurássico Superior da Formação da Lourinhã (Bacia Lusitânica, Portugal) tem uma fauna rica de dinossauros, incluindo terópodes, saurópodes, estegossauros, anquilossauros e ornitópodes. A sedimentologia favorece a fossilização e conservação de icnofósseis e existem pegadas da maioria dos grupos de dinossauros conhecidos a partir de restos esqueléticos da área. Uma grande pegada tridáctila foi descoberta, junto com uma pegada de estegossauro, no Verão de 2003, num argilito na zona intertidal na Praia de Vale Frades, a cerca de 6km a norte da Lourinhã, Portugal. Por perigo de erosão eminente, a pegada foi recolhida e encontra-se actualmente em exposição no Museu da Lourinhã (ML1000). A pegada, com 70 cm de comprimentos e 69 cm de

2 Mateus & Milàn – A giant ornithopod track ______largura, tem dedos curtos e largos, com marcas de garras arredondadas, e um elevado grau de afastamento entre os dedos. A análise multivariada das dimensões da pegada identifica-a como de um grande dinossauro ornitópode com uma altura até acetabular de cerca de 3 metros. Embora sejam conhecidos grandes ornitópodes no Cretácico, o maior do Jurássico é o Camptosaurus dos Estados Unidos e em Portugal é o camptossaurídeo Draconyx loureiroi. Nenhum destes atinge a dimensão sugerida por estas pegada e este icnito é a única evidência para um ornitópode Jurássico deste tamanho. ______

Introduction & Mateus, 2003) as well as dinosaur nests, The Lourinhã Formation (Hill, 1989), eggs and embryos (Antunes et al., 1998; which is exposed in the area around Lourinhã, Mateus et al., 1998). The formation further central west Portugal, is part of the Lusitanian contains numerous fragments of plants and Basin and consists of approximately 140 m of large fossilized tree logs (Pais, 1998). terrestrial sediments, deposited during the initial rifting of the Atlantic in the Kimmeridgian and Tithonian. The sediments consist mainly of thick red and green clay layers, interbedded with massive fluvial sandstone bodies and heterolithic horizons. The sandstone bodies appear as horizontally extensive and lenticular beds; some are traceable for several kilometres in the sections exposed along the coast. The sandstone lenses have been interpreted as distal alluvial fan Figure 1. Location map showing dinosaur track sites in facies originating from periods of extensive the central-west of Portugal. The herein described track faulting. was found at the beach of Vale Frades.

The Lourinhã Formation contains a very The nature of the deposits with the rich vertebrate fauna comprising fish, abundant shifts between flood-plain muds and amphibians, turtles, mammals, pterosaurs, fluvial sands create the perfect environment crocodiles, sauropods, theropods, ornithopods for preservation and fossilization of tracks. and thyreophorans (Antunes, 1998; Antunes

3 Mateus & Milàn – A giant ornithopod track ______

Figure 2. A, the large (70 cm long) track as it was found within the intertidal zone on the beach, preserved as a sandstone cast on a pedestal of clay. Scale bar 10 cm. B, stegosaurian track preserved beside the big ornithopod track. Scale on the knive-handle equals 10 cm. C, sketch of the two tracks as they appeared beside each other before excavation. Scale bar 10 cm. (fig. 2). Owing to the immediate danger that Within the last years, more than thirty the tracks would be destroyed by further well-preserved tracks of sauropods (Milàn et erosion from the sea, or by collectors, both al., 2005), theropods, thyreophorans and the tracks together with the surrounding ornithopods have been found, both preserved sediment was excavated and this is now on in situ on the sedimentary surfaces and on display at Museu da Lourinhã (ML 1000). loose blocks eroded out from the steep coastal cliffs (Antunes & Mateus, 2003). On July 3th The track 2003, a gigantic, 70 cm long, tridactyl track The track is tridactyl and almost was discovered on a clay bed exposed within symmetrical along the middle digit. The track the tidal zone at the beach of Vale Frades, is 70 cm long and 69 cm wide. The casts of approximately 6 kilometres north of Lourinhã the three toes are broad and rounded, with (fig. 1). Adjacent to the track, a smaller, 36 indications of short blunt claws. The toes cm long, hind-track of a stegosaur was found spacing is wide and the divarication angle

4 Mateus & Milàn – A giant ornithopod track ______between the outer toes is close to 90 degrees The track is preserved as a natural cast (fig. 2). of sandstone, preserved on a pedestal of clay (fig. 3). This unusual type of preservation can be explained by the following scenario. (1) The track was emplaced in the soft mud of the floodplain, and the weight of the trackmakers foot compacted the clay beneath the foot. (2) The floodplain was subsequently flooded and a layer of sand was deposited, and filling the track. (3) During fossilization, the sand filling the track hardened to sandstone, while the surrounding clay remained unconsolidated. (4) Present day wave erosion has eroded the surrounding sand layer away, except for the thicker parts, constituting the natural casts of the track, which acts like a locally thicker lens of sandstone. (5) The sandstone casts prevent erosion of the softer clay beneath, creating the pedestals. The present day surface of the track thereby does not represent the actual shape of the bottom of the track, but represents instead an erosion surface through the natural cast of the track. The bottom of the track is buried approximately five centimetres beneath the top of the clay pedestal.

Figure 3. Interpretative model for the preservation of Discussion the track. A, the track is emplaced in the soft clay on the floodplain. B, the floodplain was flooded and a The morphology of the newly found layer of sand was deposited, infilling and covering the track. C, during fossilization the sand layer hardened to track is consistent with the tracks of sandstone. D, present day erosion by the sea, eroding the sandstone layer away but leaving only the casts of ornithopod dinosaurs. As a general rule of the tracks. E, the sandstone casts prevents erosion of thumb, tracks of ornithopod dinosaurs are the softer clay layer beneath them, creating the pedestals with the sandstone casts on top. Graphics by broader than long or of equal width and Simão Mateus.

5 Mateus & Milàn – A giant ornithopod track ______length, with short, broad digits and a high distinguish the tracks of theropods from those angle of divarication between the outer digits, of ornithopods (Moratalla et al., 1988). The while tracks of theropod dinosaurs are longer parameters used in this method are track than they are wide, have long slender digit length (L), track width (W), total digit lengths impressions, and a lower divarication angle (LII–IV), basal digit lengths (BL2–4), basal between the outer digits (Lockley, 1991). digit widths (WBII–IV), middle digit widths However, a wide range of morphological (WMII–IV) and “heel” to interdigital point variation exists between the “typical” tracks (hypex) lengths (K, M). These measurements of ornithopod and theropod dinosaurs, and the were used on the track described herein (fig. identification of single tracks can be difficult 4), and the results of the analysis were if the track does not fall within the typical compared with the values of Moratalla et al. morphology. (1988) (tabl. 1). The results of the analysis show that the values for all parameters except one fall within the values expected for tracks of ornithopod dinosaurs. The values L/K and L/M are close to 2, which is the threshold value for discriminating ornithopod from theropod tracks, but the other parameters all fall well within the expected values for ornithopod tracks (fig. 4, tabl. 1). Only the relation LIII / WBIII, the total length of the middle digit to the basal width of the middle

digit, gives a value that falls within the range Figure 4. Sketch of the track with the measurements used in table 1. L: track length. W: track width. LII–IV: of expected theropod values (fig. 4, tabl. 1). whole digit length. BL2–4: basal digit lengths. WBII– IV: basal digit widths. WMII–IV: middle digit widths. In this case, however, the dimensions of the K, M: “heel”to interdigital (hypex) lengths. As the track was a single find, the assignment of digits II and middle digit could be a result of the IV are coincidental. Adapted from Moratalla et al. 1988. taphonomy of the track. The surface of the track is not the true track but, but represents The proportions between various an erosional cut through the sandstone cast of measurements of the length of the track and the track, approximately five cm above the size of the individual digit impressions of the true track. track can be used in multivariate analysis to

6 Mateus & Milàn – A giant ornithopod track ______

Track Treshold values and probability that the track is New track parameters either theropod or ornithopod L / W 80.0% Theropod >1.25> Ornithopod 88.2% 1.01 L / K 70.5% Theropod >2.00> Ornithopod 88.0% 1.96 L / M 65.0% Theropod >2.00> Ornithopod 90.7% 1.98

BL2 / WMII 76.1% Theropod >2.00> Ornithopod 97.7% 1.28

BL3 / WMIII 72.7% Theropod >2.20> Ornithopod 97.7% 1.91 BL4 / WMIV 76.1% Theropod >2.00> Ornithopod 97.4% 1.6

LII / WBII 84.6% Theropod >3.75> Ornithopod 90.2% 2.87

LIII / WBIII 70.6% Theropod >4.00> Ornithopod 91.5% 4.09

LIV / WBIV 73.7% Theropod >3.75> Ornithopod 93.4% 3.24

Table 1. Track parameters used in discrimination between theropod and ornithopod tracks (see fig. 4). The track parameters, threshold values and percentages of probability that the track is from a theropod or ornithopod is adopted from Moratalla et al. (1988). The values for the new big track is shown in the last column, and all except the value for LIII / WBIII fall within the values for tracks assigned to ornithopod dinosaurs.

The narrowing of the basal part of the will either support or weaken the true track in this case can be explained by identification of the track. partial collapse of the track subsequent to the An undescribed large-sized theropod lifting of the foot. Convergence of the track from the Lourinhã Formation, being 79 trackwalls is a common feature of tracks cm long but only 60 cm wide, with a low imprinted in soft sediments, and will give the angle of divarication between the digits, appearances of narrower digit impressions strongly differ in morphology to the (Milàn, in press, Milàn & Bromley, in press). ornithopod track here described. Several A narrowing of the digit impressions would smaller ornithopod tracks from the formation give a value closer to, or within the range for all have dimensions typical of ornithopods tracks of theropods by the method of (Antunes & Mateus, 2003). Moratalla et al. (1988). New methods for No manus impressions were found discriminating between theropod or associated with the track. Several trackways ornithopod tracks using cluster analysis and from around the world show that ornithopods ichnophenetics (Braddy & Elvidge, in prep), occasionally used quadropedal locomotion (Moratalla et al., 1992, 1994). Ornithopods

7 Mateus & Milàn – A giant ornithopod track ______bore only a relative small portion of their Jurassic ornithopod. However, recent work weight on their forelimbs which were based on computer models of theropods and therefore less deeply impressed in the bipedal ornithopods have shown that, for sediment than the hind limbs, which resulted large sized ornithopods, the estimate of in the manus tracks commonly being small or Alexander (1976) that the hip-height equals absent in ornithopod trackways. In this case, 4*foot-length, is the most precise (Henderson, however, the smaller sized manus track, if 2003). This makes it most likely that the present at all, may have been destroyed by the ornithopod trackmaker responsible for the extensive erosion from the sea. herein described track stood, with a hip-height The preservation of tracks and skeletal of just below three metres. material is the result of two widely different Although very large ornithopods are taphonomic processes. It is common that known from the Cretaceous, and ornithopod tracks provide additional data about faunal tracks up to 80 cm have been reported elements not represented by skeletal material (Peterson, 1924), the largest known Jurassic (e.g. Milàn & Bromley, 2005, Romano & ornithopod is Camptosaurus from USA, Whyte, 2003), or animals are indicated that which could reach 8.13 meters in length are significantly larger than any skeletal (Erickson, 1988: 10), albeit the largest known material has represented (Morales & Bulkley, from Portugal is the camptosaurid Draconyx 1996). loureiroi (Mateus & Antunes, 1993). The The hip-height for a bipedal dinosaur Draconyx holotype ML357 femur is 29 cm can be estimated by as a multiplicaton of the long as preserved and would have been about foot-length. Alexander (1976), proposed the 80 cm if complete and has a diaphyseal relation hip-height = 4*foot-length. Applying perimeter of 23 cm, which gives an estimated this formula to the newly found track gives a hip height of about two meters when hipheight of the trackmaker of 2.8 metres. compared with the values presented by Thulborn (1990) suggested that the hip-height Henderson (2003), and a body mass of about of ornithopod track-makers with a foot-length 835 kg according to the formula body mass = in excess of 25 cm should be calculated as 0.16 * femur perimeter2.73 presented by 5.9*foot-length. In this case the hip-height of Anderson et al., (1985). Neither of these the dinosaur responsible for the herein reached the body size suggested by the new described track becomes 4.13 metres, which track, and so far, the herein described track is is an unrealistically large value for a Late

8 Mateus & Milàn – A giant ornithopod track ______the only evidence of a Jurassic ornithopod (Mammalia) from Pai Mogo, Portugal. dinosaur of that size. Memórias da Academia de Sciencias, 37: 125–153. Acknowledgements The research of JM is supported by a ______& MATEUS, O. 2003. Dinosaurs of Ph.D. grant from the Faculty of Natural Portugal. Comptes Rendus Palevol, 2: 77–95. Sciences, University of Copenhagen and of OM supported by a Ph.D fellowship ______, TAQUET, P. & RIBEIRO, V. 1998. (BD/21616/99) of the Portuguese Fundação Upper Jurassic dinosaur and crocodile eggs para Ciência de Tecnologia. We thank Anne from Pai Mogo nesting site (Lourinhã- Schulp for the perusing and Simão Mateus for Portugal). Memórias da Academia de the illustrations. Richard G. Bromley, Sciencias, 37: 83–100. Geological Institute, University of Copenhagen kindly read the manuscript and BONAPARTE, J. & MATEUS, O. 1999. A improved the English and Gilles Cuny, new diplodocid, Dinheirosaurus Geological Museum, University of lourinhanensis gen. et sp. nov., from the Late Copenhagen provided a French translation of Jurassic beds of Portugal. Revista del Museo the abstract. Argentino de Ciencias Naturales, 5: 13–29.

References ERICKSON, B. R. 1988. Notes on the postcranium of Camptosaurus. Scientific ALEXANDER, R. McN. 1976. Estimates of Publications of the Science Museum of speeds of dinosaurs. Nature, 261: 129!130. Minnesota n.s. 6: 13 p.

ANDERSON, J. F., HALL-MARTIN, A. & HENDERSON, D. M. 2003. Footprints, RUSSELL, D. A. 1985. Long-bone Trackways, and Hip Heights of Bipedal circumference and weight in mammals, birds Dinosaurs – Testing Hip Height Predictions and dinosaurs. Journal of Zoolology Series A, with Computer Models. Ichnos, 10: 99–114. 207: 53–61. HILL, G. 1989. Distal alluvial fan sediments ANTUNES, M. T. 1998. A new Upper from the Upper Jurassic of Portugal: Controls Jurassic Paulchoffatiid multituberculate on their cyclicity and channel formation.

9 Mateus & Milàn – A giant ornithopod track ______

Journal of the Geological Society, London, Jurassic of Lourinhã, Portugal. Annales 146: 539–555. Paleontol, 87: 61–73.

LEINFELDER, R. R. 1987. Multifactorial control of sedimentation patterns in an ocean MILÀN, J. (in press 2006) Variations in the marginal basin: the Lusitanian Basin morphology of emu (Dromaius (Portugal) during the Kimmeridgian and novaehollandiae) tracks, due to differences in Tithonian. Geologische Rundschau, 76: 599– walking pattern and substrate consistency: 631. ichnotaxonomical implications. Palaeontology 49(2) LOCKLEY, M. 1991. Tracking Dinosaurs.

Cambridge University Press, Cambridge. 238 ______& BROMLEY, R. G. 2005. Dinosaur pp. footprints from the Middle Jurassic Bagå

Formation, Bornholm, Denmark. Bulletin of MANUPPELLA, G. 1998. Upper Jurassic the Geological Society of Denmark, 52: 7–15. theropod dinosaur embryos from Lourinhã

(Portugal). Memórias da Academia de ______BROMLEY, R. G. (in press). True Sciencias, 37: 101–110. tracks, undertracks and eroded tracks,

experimental work with tetrapod tracks in MATEUS, I., MATEUS, H., ANTUNES, M. laboratory and field. Palaeogeography, T., MATEUS, O., TAQUET, P., RIBEIRO, Palaeoclimatology, Palaeoecology. V. &

______, CHRISTIANSEN, P. & MATEUS, O. 1998. Lourinhanosaurus MATEUS, O. 2005. A three- antunesi, a new Upper Jurassic allosaurid dimensionally preserved sauropod (Dinosauria- Theropoda) from Lourinhã, manus impression from the Upper Portugal. Memórias da Academia de Jurassic of Portugal: Implications for Sciencias, 37: 111–124. sauropod manus shape and locomotor

mechanics. Kaupia, Darmstädter ______& ANTUNES, M. T. 2001. Beiträge zur Naturkunde, 14: 47–52. Draconyx loureiroi, a new Camptosauridae

(Dinosauria: Ornithopoda) from the Late

10 Mateus & Milàn – A giant ornithopod track ______

MORALES, M. & BULKLEY, S. 1996. PETERSON, W. 1924. Dinosaur Tracks in Paleoichnological evidence for a theropod the Roofs of Coal Mines. Natural History, 24: dinosaur larger than Dilophosaurus in the 388–397. Lower Jurassic Kayenta Formation; pp. 143– 146. In Morales, M. (ed.), Continental ROMANO, M. & WHYTE, M. A. 2003. Jurassic Symposium Volume. Museum of Jurassic dinosaur tracks and trackways of the Northern Arizona Press, Flagstaff. Cleveland Basin, Yorkshire: preservation, diversity and distribution. Proceedings of the MORATALLA, J. J., SANZ, J. L. & Yorkshire Geological Society, 54: 185–215. JIMENEZ, S. 1988. Multivariate analysis on Lower Cretaceous dinosaur footprints: THULBORN, T. 1990. Dinosaur tracks. Discrimination between ornithopods and Chapman and Hall, London. 410pp. theropods. Geobios, 21: 395–408.

______, SANZ, J. L., JIMENEZ, S. & LOCKLEY, M. G. 1992. A Quadropedal ornithopod trackway from the Lower Cretaceous of La Rioja (Spain): Inference on gait and hand structure. Journal of Vertebrate Paleontology, 12: 150–157.

______, SANZ, J. L., JIMENEZ, S. 1994. Dinosaur tracks from the Lower Cretaceous of Regumiel de la Sierre (province of Burgos, Spain): inferences on a new quadrupedal ornithopod trackway. Ichnos, 3: 89–97.

PAIS, J. 1998. Jurassic plant macroremains from Portugal. Memórias da Academia de Sciencias, 37: 25–48.

11 Jesper Milàn – Vertebrate Ichnology Appendix ______

NORWEGIAN JOURNAL OF GEOLOGY Tracking polar dinosaurs, Svalbard 397

Tracking polar dinosaurs - new finds from the Lower Cretaceous of Svalbard

Jørn H. Hurum, Jesper Milàn, Øyvind Hammer, Ivar Midtkandal, Hans Amundsen & Bjørn Sæther

Jørn H. Hurum, Jesper Milàn, Øyvind Hammer, Ivar Midtkandal, Hans Amundsen & Bjørn Sæther. Norwegian Journal of Geology, Vol. 86, pp. 397-402. Trondheim 2006, ISSN 029-196X.

A new discovery of ornithopod dinosaur tracks from Svalbard is described. The Lower Cretaceous (Barremian) section at Isfjorden consists of sandstones and interbeds consistent with an alluvial flood plane. The newly discovered tracks are situated on two different horizons stratigraphi- cally below the original horizon found in 1960. Footprint evidence from Festningen and Kvalvågen suggests that during the Early Cretaceous there was a diverse dinosaur fauna on Svalbard and that both theropods and ornithopods were present at the time.

Jørn Harald Hurum, Natural History Museum, University of Oslo, P.O.Box 1172 Blindern, NO-0318 Oslo, Norway ([email protected]). Jesper Milàn, Geological Institute, Øster Voldgade 10, DK-1350, Copenhagen K., Denmark. ([email protected]). Øyvind Hammer, Natural history Museum, University of Oslo, P.O.Box 1172 Blindern, NO-0318 Oslo, Norway ([email protected]). Ivar Midtkandal, Department of Geosciences, P.O Box 1047 Blindern, NO-0316 Oslo, Norway ([email protected]). Hans Amundsen, PGP, University of Oslo, P.O Box 1048 Blindern, NO-0316 Oslo, Norway ([email protected]). Bjørn Sæther, Statoil ASA, Research and Technology, Postuttak, N-7005 Trondheim, Norway ([email protected])

Introduction related, ornithopod dinosaur (Sarjeant et al. 1998). Unfortunately all of the tracks discovered in 1960 have In 1960, the discovery of ornithopod dinosaur footprints since been lost due to erosion by the sea. from Svalbard was received as a sensation because they The aim of this study is to describe the newly found constituted the first indisputable evidence that dinosaurs tracks, their ecology and the importance of such finds at had inhabited the polar latitudes (Lapparent 1960, 1962). polar latitudes Later, in 1976, tracks of a medium-sized theropod were discovered at a second locality in the same formation at Kvalvågen, eastern Spitsbergen (Edwards et al. 1978). These two records (Fig. 1a), together with a find from Methods Kempendajay, Siberia, were isolated discoveries until the late 1980s and polar environments were not considered In-situ laser scanning of further importance for dinosaur research. However, The tracks were found on the hanging wall of a crevice this has since changed, and in the last 20 years, more than some 30 cm in width, making observation and mapping ten arctic areas have yielded both skeletal remains and difficult. We constructed a simple laser profiler with a footprints of dinosaurs, ranging from the Late Jurassic to horizontal laser line that was projected obliquely onto Late Cretaceous in North America, Siberia and Svalbard the bedding plane and tracks. The equipment consisted (Rich et al. 2002; Fiorillo 2004, 2006), demonstrating that of a laser with a cylindrical lens and a digital camera with the finds of dinosaurs at polar latitudes were not isolated remote control, both mounted on an aluminum frame. phenomena. The frame was suspended on ropes and could be raised and lowered inside the crevice from the top as far as hid- During the 2001 SVALEX excursion to the Festingen den ice would allow. The frame was hoisted upwards locality at Isfjorden, and later with the SVALSIM group at 2.5cm intervals and the bedding plane and tracks mapping in April 2002 of the Festningen sandstone, Hel- scanned at a horizontal width of 30-40 cm. Because of vetiafjellet Formation (Early Cretaceous, Barremian), the limited width, the apparatus had to be moved later- several new dinosaur tracks were discovered on the same ally 14 times to cover the full length of the crevice. Each escarpment as the 13 footprints found in 1960 (see Lap- of the 14 strips was aligned manually using the computer. parent 1960, 1962). Based on superficial resemblance, the A total of 3,850 laser profiles were scanned and later latter were identified as belonging to Iguanodon (Lappar- converted to 3-D coordinates and triangulated using inent 1960, 1962; Heintz 1963; Steel et al. 1978) but have house software. Finally, the resulting polygonal 3D mesh since been considered to belong to a different, although was visualized on a Silicon Graphics workstation. The 398 J. H. Hurum et al. NORWEGIAN JOURNAL OF GEOLOGY result was mixed, with clear images in some areas, while along the coasts. Inland, relief was low, with areas of relative others were blurred or missed because of shadows due to stability at the fringes of rivers and streams, where levees high-amplitude topography. Additional footprints were and bars provided ideal sites for the growth of large plants also identified without laser scanning. including trees. Elsewhere, changing conditions along the coastal plain inhibited plant growth and large tidal flats were bare, save for vegetation along the sides of narrow Casting of tracks channels (Nemec 1992). It is estimated that Spitsbergen was located at about 65º N at the time the sediments of the Hel- Three tracks were covered with silicone to make moulds vetiafjellet Formation were being deposited, and that there in April and May 2002. A problem with the setting time was some connection to Greenland to the south (Fig. 1). of the silicone was severe while making the first mould in April 2002 when the crew experienced air temperatures down to minus 40 degrees Celsius (with the calculated The Festningen locality wind factor). During the second fieldwork in 0 degrees Celsius we had no problems with the silicone. The Festningen profile is known for its almost vertically inclined strata where a nearly continuous succession of sediments from Permian into Cretaceous are exposed along 5 km of beach cliffs facing north at the mouth of Geological setting Isfjorden. Due to the vertical orientation of the strata, The Festningen Sandstone Member (Parker 1967; Dall- only a 5 – 10 m wide section of the strata can be seen mann 1999) forms the lowest stratigraphic unit of the at any stratigraphic level. The Helvetiafjellet Formation Helvetiafjellet Formation (Barremian - Aptian) with type is exposed at the eastern limit of the profile, facing east locality at the vertical sandstone cliffs on Festningsodden into Grønfjorden. At the type locality, the Festningen (festning = fortress, odden = spike of land). The Helve- Sandstone Member is composed of two channel sand- tiafjellet Formation represents a wide range of depositional stone bodies, separated by a fluvial erosion surface and environments and occurs over much of Spitsbergen. It was fine grained heterolithic sediments (Fig. 2). Immediately deposited following a significant regional fall in relative sea above the Festningen Sandstone Member, two separate level that exposed the shelf deposits of the Janusfjellet Sub- horizons of dinosaur tracks are contained within a 3 m group (Dypvik 1985). At this time, strong tides were active silt- and sandstone unit (Fig. 2, ~20m). Above this strati-

Fig 1. :A. Map of Svalbard showing the two Early Cretaceous tracksites. 1 Ornithopod tracks (Lapparent 1960). 2. Theropod tracks (Edwards et al. 1978). B. Svalbard at 125 Ma. Before the opening of the Atlantic, Svalbard was located north of Greenland, with a probable connection to large landmasses to the south. From Torsvik et al. (2001). NORWEGIAN JOURNAL OF GEOLOGY Tracking polar dinosaurs, Svalbard 399

Results The results from the in-situ laser scanning (Fig. 3) showed at least five ornithopod tracks in the crevice. Several other tracks were seen in the crevice outside the scanned area and in two other stratigraphic levels at the locality (Fig. 2). All the tracks are from ornithopod dinosaurs, measured around 60 cm in length and are of equal width, with impressions of relatively short blunt digits (Fig. 4).

Fig. 3. Digital model of the tracks in the crevasse at Festningen. Com- puter reconstruction of part of the dinosaur trackway site at Festnin- gen, Svalbard, based on laser scanning in 2002. The vertical surface covered, seen from the fissure side, is approx. 3x3.5 m. Higher relief is given brighter shade. Flat areas represent missing data and openings in the wall. Features seen in the scan include an almost complete footprint pointing to the lower left (1), an almost complete footprint pointing upwards (2) it is the footprint cast and pictured in figure 4 and 5b , a smaller, less clear footprint pointing upwards (3), an unclear footprint possibly pointing upwards (4), and a large, blurred footprint with unclear orientation (5). Several other footprints are blurred or distorted, and a number of footprints were also observed outside the scanned area. Fig 2. Sedimentological log of the Helvetiafjellet Formation at Festnin- gen, showing the stratigraphic position of the dinosaur tracks (modi- Fig. 4. Schematic dra- fied from Midtkandal 2002). wing of footprint with measurements. Drawn from a cast of the best graphic level, nearly 100 m of heterolithic coastal and preserved track in the shallow marine sediments are stacked, forming the mid- crevice PMO 210.570. dle and upper Helvetiafjellet Formation at Festningen. A PMO - Paleontological single dinosaur footprint is recorded in the upper Hel- Museum Oslo. The ori- vetiafjellet Formation at Festningen (Fig. 2, 55 m). The entation of the footprint crevice containing the tracks is the result of selective ero- is shown as the number 2 in figure 3a and a cast is sion by the sea of a coal rich mudstone bed interbedded figured in 5b. between the sandstone beds. 400 J. H. Hurum et al. NORWEGIAN JOURNAL OF GEOLOGY

Description of the tracks The tracks at the Festningen locality are found in three distinct preservational types: 1. preserved as concave low-relief impressions in sandstones (concave epirelief) (Fig. 5a). 2. preserved as natural sandstone-casts in the eroded coal bed (convex hyporelief) (Fig. 5b). 3. preserved as isolated sandstone casts in organic rich shales (full relief) (Fig.5c).

In type 1, the tracks are rounded in appearance and only the gross overall shape is preserved (Fig. 5a). This morphology can either be the result of a true track exposed to erosion, or that the exposed track is an undertrack originally formed along a subjacent horizon, due to the pressure of the dinosaur’s foot being transferred down into the subjacent layers (Lockley 1991). Experiments using artificial substrates show how the shape of a tridactyl track can change from being well defined with sharply defined digit impressions, to rounded, shallow-relief depressions as undertracks along successive subjacent horizons (Manning 2004; Milàn & Bromley 2006).

In type 2 (Fig. 5b), the tracks are thought to have been formed when the dinosaur walked in soft organic rich swamp mud that was subsequently flooded and the tracks were infilled by sand. Later erosion of what subsequently became a coal-seam left the sand infillings as natural casts (Lockley 1991) on the underside of the overlying sandstone bed. This situation is very common and numerous such dinosaur tracks have been exposed during coal mining (eg. Peterson 1924; Parker & Balsley 1989; Parker & Rowley 1989; Milàn & Gierlinski 2004). This is the most common track type in the locality, and all the scanned tracks are of this kind of preservation. The exposed part of the sandstone contains at least nine single tracks and several distorted multiple tracks. They are all of similar size (length: 63 cm Fig 5.: A1. The first tracks described from Festningen, which are now lost to erosion, were preserved as shallow imprints in the sandstone and width 62 cm.). A tennisball-sized outgrowth on the side beds. Photograph of a cast of one of the best tracks, PMO X621. A2, of the middle toe is present in at least two of the tracks. This schematic series of possible formation of the tracks. Upper illustration: is probably a pathological phenomenon (see Fig. 4). the track is imprinted in sandstone layers separated by thin mud layers. The weight of the animal causes the formation of undertracks in the In type 3, the isolated sandstone casts in organic rich layers subjacent to the foot. Middle illustration: the track is covered by shale are preserved as a combination of both under- other sediment and is lithified. Lower illustration: present day erosion of the cliffs has exposed the horizon with the shallow-relief undertrack. track formation and erosion (Fig. 5c). First, the tracks B1. Tracks in the crevice are found as natural casts on the underside of were emplaced in soft layered mud and, depending on a sandstone bed, due to erosion of the coal seam. Photograph of cast the weight of the animal, could extend into underlying

PMO 210.570. B2 schematic series of possible formation of the tracks. beds. Subsequent infilling by sand shows this extension Upper illustration: the track is emplaced in an organic-rich swamp. and how the sides of the former mud layer are dragged Middle illustration: after removal of the foot, the track is covered by downwards into the undertracks. sand. Lower illustration: present day erosion has removed the softer coal seams and exposed the natural sandstone casts of the tracks. C1. Photograph of original track preserved as isolated sandstone casts in organic rich shale. C2 interpretative schematic series explaining the unusual type of preservation. Upper illustration: the track is imprinted Interpretation of the palaeo-environment in soft, layered organic-rich clay. Undertracks are formed along several subjacent horizons. Middle illustration: subsequently to lifting of The lower portion of the Helvetiafjellet Formation at Fest- the foot, the tracks become filled with sand. Lower illustration: present ningen and in other areas is interpreted to represent a low day erosion level has exposed the track as an isolated sandstone cast. relief coastal plain to shallow marine environment (Steel et The clay layers around the sandstone cast seem to form a bowl shaped al. 1978; Midtkandal 2002). A high influx of sediment from structure below the sandstone cast. This is due to differentiated erosion of the clay layers the track was emplaced in. the hinterland was reworked by tides and redistributed at NORWEGIAN JOURNAL OF GEOLOGY Tracking polar dinosaurs, Svalbard 401

Fig 6: Palaeo-environment of the Helvetiafjellet Formation at Festningen.

and along the coast. Colville River, Alaska, have yielded The result was a highly a diverse dinosaurian ichnofauna irregular coastline with a comprising tracks from ornithopods, complex distribution of subenvi- theropods and unspecified tetrapods as ronments along the coastline and on the well as abundant skeletal material of cera- coastal plain. A possible palaeo-environment topsians, theropods and several genera of scenario is illustrated in Fig. 6. The entire area ornithopods (Rich et al. 2002). In Yukon Terri- was susceptible to minor changes in relative sea level, tory, Canada, tracks and trackways of ornithopods, which could relatively quickly submerge or dry out large theropods and unspecified tracks from quadrupedal swaths of land. Anomalous events such as floods or storm dinosaurs have been reported (Weishampel 1992; Rich et surges deposited sand sheets in places where sediment oth- al. 1997; Rich et al. 2002). This shows that a diverse dino- erwise would not reach. Areas left undisturbed for longer saur fauna inhabited the northern hemisphere during the periods of time would have the potential to sustain a signif- Late Mesozoic. icant vegetation cover. All the dinosaur tracks at the locality are preserved in the sandstones and interbedded coal- So far no dinosaur footprints has been reported from rich mudstone beds. The interbedded coal-rich mudstone within the modern polar latitudes of the southern hemi- beds and the upward thickening sandstone beds seen at sphere, but skeletal remains of theropods, sauropods and Festningen reflect a subaerial environment with sediment thyreophoreans have been found in New Zealand (Molnar input occurring in pulses between relatively long periods & Wiffen 1994), and skeletal remains of theropods, orni- of time in which peat accumulation is favorable. thopods and thyreophoreans are known from Antarctica and southern Australia (Rich et al. 2002).

Dinosaur remains are now known from more than ten Arctic Discussion areas in North America, Siberia and Svalbard in sediments of Late Jurassic to Late Cretaceous age. At this time the area The dinosaurian ichnofauna from Svalbard bears evi- was landlocked around the North Pole except for the strait dence of the presence of medium sized theropods with between Alaska and the northeastern parts of Asia (Allen et al. a foot length around 30 cm, and larger ornithopods with 1993; Briggs 1995), and oceanographic conditions may have a foot length in excess of 60 cm. The new findings show resembled those of the present Antarctic Sea. Polar currents that the presence of dinosaur tracks is not limited to a circulating around a polar ice cap would have lowered the sea single stratigraphic horizon in the Festningen area. The temperature and the adjoining areas, with the interiors of the distribution of the ornithopod tracks of equal size found northern continents experiencing severe winters. If this is the in the crevice (Fig. 3 and 4) suggests either a herd of sim- case, then the dinosaurs of the Alaskan Peninsula, northern ilar sized animals or repeated trampling from the same Canada, Siberia and Svalbard would have to cope with snow. individual, maybe during browsing for food. Lockley (1991) suggested that in North America some dino- Dinosaur tracks in modern polar regions are only known saur species could make long distance migration down the from a few localities on the northern hemisphere. In shores of interior seaways as the arctic caribou do today. This addition to the Cretaceous tracks from Svalbard, so far, idea he supported by finds of numerous north-south head- five Cretaceous and one Jurassic tracksite are known ing shoreparallel trackways along the palaeocoast deposits from Alaska (Fiorillo 2006). The Late Jurassic tracks of the interior seaway together with records of the same were found in the Black Lake area of the western Alaska dinosaur species found at both ends of the hypothetical Peninsula, and consist of a hitherto undescribed slab of migration paths (Dodson 1997). Fiorillo & Gangloff (2001) rock containing several tracks from a medium sized the- found these statements unlikely on the basis of growth stud- ropod (Fiorillo 2006). The Upper Cretaceous deposits in ies of hadrosaurs and migration patterns and growth studies 402 J. H. Hurum et al. NORWEGIAN JOURNAL OF GEOLOGY on extant caribou. They concluded that dinosaurs probably faunas. Journal of the Paleontological Society of Korea 22, 15-27. remained year-round in the arctic in the Late Cretaceous. Fiorillo, A.R. & Gangloff, R.A. 2001. The caribou migration model for Arctic hadrosaurs (Ornithischia: Dinosauria): a reassessment. His- torical Biology 15, 323-334. Protection against the cold is a problem that theropod Heintz, N. 1963: Dinosaur-footprints and polar wandering. Norsk dinosaurs solved by developing down and feathers (Norell Polarinstitutt, Årbok 1962, 35-43. & Xing 2005 and references therein), but so far no ornitho- Holtz, T.R., Chapman, R.E. & Lamanna, M.C. 2004. Mesozoic biogeo- pods have been found with feathers. However, featherlike graphy of Dinosauria. In Weishampel, D.B., Dodson, P. & Osmól- structures are known from other ornithischians, as bristle ska H. (Eds.); The Dinosauria. 2nd edition. 627-642. University of like integumentary structures have been found on the tail of California Press, Berkeley. a Psittacosaurus, a small horned dinosaur (Mayr et al. 2002). Lapparent, A.F. de 1960: Decouverte de traces de pas de dinosauriens dans le Crétacé de Spitsberg. Comptes rendus de l’Académie des sci- This is of interest since both theropods and ornithopods ences 251, 1399-1400. inhabited Svalbard in the Early Cretaceous. Small dinosaurs Lapparent, A.F. de. 1962: Footprints of Dinosaur in the Lower Cretaceous may have been able to hibernate but this is unlikely for large of Vestspitsbergen - Svalbard. Norsk Polarinstitutt, Årbok 1960, 14-21. ornithopod genera. Unfortunately no skeletal remains of Lockley, M. 1991: Tracking Dinosaurs. Cambridge University Press. 238 pp. dinosaurs are known from Svalbard, which has an impor- Manning, P. 2004: A new approach to the analysis and interpretation tant position midway between North America and Siberia. of tracks: examples from the dinosaurian. In McIlroy, D. (Ed.); The application of ichnology to palaeoenviromental and stratigraphic analysis. Geological Society, London, Special Publication 228, 93-123 Mayr, G., Peters, D.S., Plodowski, G. & Vogel, O. 2002: Bristle-like integumentary structures at the tail of the horned dinosaur Psittaco- Conclusion saurus. Naturwissenschaften 89, 361-365. Midtkandal, I. 2002: Depositional environment, sandstone architecture The newly found ornithopod tracks from Festningen, and sequence stratigraphy of the lower Cretaceous Helvetiafjellet For- Svalbard, demonstrate that previous finds of dinosaur foot- mation, Spitsbergen. Unpublished thesis, University of Oslo, 162 pp. prints from the area were not isolated phenomena, and that Milàn, J. & Bromley, R.G. 2006: True tracks, undertracks and eroded dinosaur tracks occur at several stratigraphical levels within tracks, experimental work with tetrapod tracks in laboratory and the Formation, proving a spatial presence of dinosaurs in field. Palaeogeography, Palaeoclimatology, Palaeoecology 231, 253-264. the area. In the early part of the Early Cretaceous a “greater Milàn, J. & Gierlinski, G. 2004: A probable thyreophoran (Dinosau- Wealden” dinosaur fauna (Holz et al. 2004) lived through- ria, Ornithischia) footprint from the Upper Triassic of southern Sweden. Bulletin of the Geological Society of Denmark 51, 71-75. out most of Laurasia. It consisted of iguanodontians, basal Molnar, R.E. & Wiffen, J. 1994. A Late Cretaceous polar dinosaur fauna euornithopods, nodosaurid ankylosaurians, brachiosau- from New Zealand. Cretaceous Research 15, 689-706. rids and various theropods. Together with the other finds Nemec, W. 1992: Depositional controls on plant growth and peat accu- of dinosaurs from areas that were located in polar latitudes mulation in a braidplain delta environment: Helvetiafjellet Forma- in the Mesozoic, the tracks from Svalbard represent impor- tion (Barremian – Aptian), Svalbard. In McCabe, P.J. & Parrish, J.T. tant information about the distribution and composition (eds.): Controls on the Distribution and Quality of Cretaceous Coals. of the “greater Wealden” polar dinosaur faunas. Geological Society of America, Special Paper 267, 209 – 226. Norell, M. & Xing, X. 2005: Feathered dinosaurs. Annual Reviews Earth Planetary Sciences 33, 277–99 Parker, J.R. 1967: The Jurassic and Cretaceous sequence in Spitsbergen. Acknowledgement: Thanks to Prof. David L. Bruton, Natural History Geological Magazine 104, 487-505. Museum, University of Oslo for improving the manuscript and to Ric- Parker, L.R. & Balsley, J.K. 1989: Coal mines as localities for study- hard Bromley for his review of the initial manuscript. Thanks to Sindre ing dinosaur trace fossils. In Gillette, D.G. & M.G. Lockley (Eds.), Flatås for excellent logistics and safety. Dinosaur tracks and traces, 353-360. University Press, Cambridge. Parker, L.R. & Rowley, R.L. 1989: Dinosaur footprints from a coal mine in East Central Utah. In Gillette, D.G. & M.G. Lockley (Eds.); Dino- References saur tracks and traces, 361-366. University Press, Cambridge. Allen, J.R.L., Hoskins, B.J., Sellwood, B.W., Spicer, R.A. & Valdes, P.J. Peterson, W. 1924: Dinosaur Tracks in the Roofs of Coal Mines. Natu- (Eds.), 1993: Paleoclimates and their modeling with special refe- ral History 24, 388-397. rence to the Mesozoic Era. Philosophical Transactions of the Royal Rich, T.H., Gangloff, R. A. & Hammer, W.R. 1997: Polar dinosaurs. In Society of London, Series B, 341, 1297 pp. Currie, P. & Padian, K. (Eds.); Encyclopedia of Dinosaurs, 562-573. Briggs, J.C. 1995: Global Paleogeography. Developments in Palaeonto- Academic Press. New York. logy and Stratigraphy, 14. Elsevier Science BV, Amsterdam. 472 pp. Rich, T.H., Vickers-Rich, P. & Gangloff, R.A. 2002: Polar dinosaurs. Dallmann, W.K. (Ed.), 1999. Lithostratigraphic Lexicon of Svalbard, Science 295, 979-980. Norsk Polarinsitutt, 318 pp. Sarjeant, W.A.S., Delair, J.B. & Lockley, M.G. 1998: The footprints of Dodson, P. 1997: Paleoecology. In Currie, P. & Padian, K. (Eds.), Ency- wIguanodon: a history and taxonomic study. Ichnos 6, 183-202. clopedia of Dinosaurs, 515-519. Academic Press. New York. Steel, R.J., Gjelberg, J.G. & Haarr, G. 1978: Helvetiafjellet Formation Dypvik, H. 1985: Jurassic and Cretaceous black shales of the Janusfjel- (Barremian) at Festningen, Spitsbergen a field guide. Norsk. Polar- let Formation, Svalbard, Norway. Sedimentary Geology 41, 235-248. instutt, Årbok 1977, 111 128. Edwards, M.B., Edwards, R. & Colbert, E.H. 1978: Carnosaurian foot- Torsvik, T.H., Van der Voo, R., Meert, J.G., Mosar, J. & Walderhaug, H.J. prints in the Lower Cretaceous of eastern Spitsbergen. Journal of 2001: Reconstructions of the continents around the North Atlantic at Paleontology 52, 940-941. about the 60th parallel. Earth and Planetary Science Letters 187, 55– 69. Fiorillo, A.R. 2004: The dinosaurs of Arctic Alaska. Scientific American 291, 85-91. Weishampel, D.B. 1992: Dinosaurian Distribution. In Weishampel, Fiorillo, A.R. 2006. Review of the dinosaur record of Alaska. With com- D.B., Dodson, P. & Osmolska, H. (Eds.); The Dinosauria, 63-139. ments regarding Korean dinosaurs as comparable high-latitude fossil University of California Press, Berkeley, Los Angeles. Jesper Milàn – Vertebrate Ichnology Appendix ______

Milàn et al. – Vertebrate ichnofauna from Rhodes ______

A diverse vertebrate ichnofauna from a Quaternary eolian oolite, Rhodes, Greece

Jesper Milàn1, Richard G. Bromley1, Jürgen Titschack2 & Georgios Theodorou3

1Geological Institute, University of Copenhagen, Oester Voldgade 10, DK-1350 Copenhagen K, Denmark. 2Institute of Palaeontology, University of Erlangen, Loewenichstrasse 28, 91054 Erlangen, Germany. 3Department of Historical Geology and Palaeontology, Athens University, Athens, Greece.

Abstract: In coastal areas of the SW part of the island of Rhodes, Greece, eolian oolitic sediments represent the latest depositional phase, and are presumed to have a Late Pleistocene to Early Holocene age. The ooids have sand-sized nuclei, and are lightly cemented by vadose meniscus cement. In a road cutting, supplemented by minor sections in small, ancient stone quarries nearby, the sedimentary architecture and trace fossils are visible. The dunes have ramp morphology and contain three horizons of paleosols that divide the eolian sediments into three units. The paleosols contain rhizoliths and poorly preserved invertebrate bioturbation. Vertical sections allowed representative measurements of 79 tracks and limited horizontal surfaces supplied 4 measured tracks. Five size-classes of tracks are distinguished. The mode of preservation of the tracks is poor, probably on account of the oolitic nature of the substrate. The three smallest size- classes are probably of artiodactyls. The largest class probably was produced by proboscidians. A bedding-plane view of one track indicates that the next-largest class may be the work of camels. If this is the case, and the bedding-plane specimen is convincing, it is the first record of Pleistocene or Early Holocene camels on Rhodes. The combination of size groups of tracks differs in the three units, demonstrating differences in faunal composition in the different periods of deposition of the oolite. ______

1 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

INTRODUCTION We refer to the unit herein as the In the coastal regions of the southwestern part of the island of Rhodes, Greece, are patches of carbonate eolianitic sediments resting on the Early Oligocene Kattavia Flysch. The sediments rest on the deeply weathered erosion surface, locally preserving a deeply weathered paleosol. The upper surface of the eolianite unit is the present day erosion surface and shows advanced weathering features that indicate subsoil karstification. Deforestation and soil loss has now exposed this surface to the atmosphere. Figure 1. Location map of the island of Rhodes. The newly discovered Kattavia Owing to the heavy weathering of the eolianite locality, marked by a star, is exposed in a road cut near the SW coast. carbonate sediments, natural surfaces show few remains of sedimentary Kattavia eolianite. The fieldwork was structures, even in vertical sections completed in March 2005. Data were along small watercourses. Possibly for collected chiefly from the road cutting these reasons, the unit does not appear to but also from a few small quarries for have been mentioned in the published building stone close by the cutting. literature. When the road around the Mammal tracks, invertebrate burrows island was widened about 20 years ago, and plant rooting structures are a road cutting was made through the abundant. Owing to the generally poor eolianite west of Kattavia (Fig. 1), quality of preservation of invertebrate revealing reasonably well preserved trace fossils, particular attention was sedimentary structures. The ichnological paid to the mammal tracks. significance of this section (Fig. 2) went Fossil and subfossil animal tracks are unnoticed until February 2004, when most easily recognized when exposed on preliminary fieldwork was undertaken. bedding planes representing the original tracking surface.

2 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Figure 2. Schematic section of the Kattavia eolianite. A) Photo of the section viewed from south-east. The section dips on average at 4 degrees toward NW. The highest point of the section is located 41 m above present day sea level (mas), and the lowest part exposed is 17 mas. The contact between the eoliante and the Kattavia Flysch is exposed in the lower right corner of the photo. The persons on the photo are spaced at 10 m intervals. B) Sketch of the total section. The exposed eolianite is divided into three units separated by well-developed soil horizons, indicated by dotted lines. Recorded footprints are represented by black dots.

3 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

But as tracks are three-dimensional their lateral continuity clearly indicate structures deforming the original an eolian origin. And in this case, the sedimentary fabric, commonly to presence of numerous vertebrate tracks considerable depth below the original and plant roots further proves the tracking surface, it is possible to eolian origin of the deposit. recognize tracks in vertical sections or Coastal eolian dune sands in the random erosional cuts through bedding Mediterranean region are well described (Loope 1986; Lea 1996; Allen 1997; from the Valencia region and from the Fornós et al. 2002). Balearic Islands in Spain, from Lebanon Small surfaces of the original bedding and from the northern and central coastal planes were exposed in the bottom of the plain of Israel, from the north-western ditch running beside the road section and coastal plain of Egypt, from Libya and on loose slabs of carbonate grainstone, Tunisia, as well as from the lying on the opposite side of the road. Mediterranean coast of Morocco Four measurable tracks were found in (Brooke, 2001 and references therein; bedding-plane view. In contrast, 79 Clemmensen et al. 1997, 2001). On the tracks were measured in vertical section. island of Rhodes, eolianites were mentioned by Hansen (2001) as common GEOLOGY features on a wave-cut platform level, approximately 25-35 m above present Recent work has demonstrated that sea level. They represent the youngest eolian carbonate deposits in some deposits, overlying Mesozoic and cases can be virtually indistinguishable Tertiary basement rocks as well as from sub-tidal marine deposits if marine Plio-Pleistocene sediments. macro scale sedimentary structures are Hansen (2001) distinguished throughout not observed or weakly developed the island different dune shapes: coastal (Guern and Davaud, 2005). Even dunes, echo dunes and dunes having though the sedimentary structures in ramp morphology. The sediment the Kattavia eolianite are not that well composition is variable and consists of preserved, steep slopes of the reworked marine bioclasts (molluscs, red sedimentary foreset sheets, as well as algae, bryozoans, etc.) and lithoclasts

4 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

(calcite or quartz grains, dependent on significantly in its composition from the the source areas of the hinterland). eolianites known from the rest of the Commonly the dune sands are island owing to the presence of ooids as subdivided by soil horizons rich in the dominant sediment component. rhizoliths. The stratigraphic position of The ooids enclose large sand-sized the eolianites overlying Late Pleistocene cores consisting of fragments of red marine sediments at Cape Plimiri algae, bryozoans, serpulids, (observations by J. T.) and Pleistocene foraminferans, quartz grains and sediments in Lindos Bay (Hanken et al., lithoclasts (Fig. 3). The cores and 1996) suggest a Late Pleistocene to individual layers of the ooids have been Early Holocene age for these eolianites. subject to variable degrees of The road cutting exposure studied dissolution; cementation is weak and herein has a length of 210 m and a dominated by miniscus cements, clearly maximum height of 8 m (Fig. 2A). No due to exposure to a vadose meteoric wave-cut platform level as mentioned by diagenetic environment. A detailed Hansen (2001) is developed at this part description of the sedimentology of the of the island, owing possibly to the Kattavia eoliante is in preparation by J. lesser resistance of the Kattavia Flysch Titschack and colleagues. to erosive processes than many other basement rocks on the island and the lack at this locality of Plio-Pleistocene carbonate deposits. However, the Kattavia eolianite occurs at a comparable altitude, between 17 and 41 m above present day sea level, as the dune sands described by Hansen (2001), Figure 3. Thin section of the ooid grainstone and represents a ramp-shaped dune. of the Kattavia eolianite. The ooids show large nuclei composed of quartz grains, lithoclasts The Kattavia eolianite can be and bioclasts. The ooidal coating is thin and subdivided into three units by tracing laminae are partially dissolved. Cementation is weak to moderate and of meniscus type. Plain reddish soil horizons that are rich in polarized light. rhizolith traces and shells of land snails

(Fig. 2B). The sediment differs

5 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Ooids occur as a dominant sediment development, Brooke (2001) favored component in many tropical carbonate highstand condition for their formation. settings. The occurrence of ooids on However, there are multiple reports of Rhodes represents one of the few known dune formation during glacial periods examples where ooids have formed in a (McKee and Ward, 1983), also in the non-tropical environment. Other Mediterranean region (Frechen et al., occurrences are known from southern 2001). Tunisia and Libya (Fabricius and As a preliminary working model, we Berdau, 1970), from the southern suggest therefore, that the dunes of the Peloponnes, Greece (Richter, 1976) and Kattavia eoliante were formed during from Turkey (but the Turkish occurrence glacial intervals and that the soil may have been transported there in formation occurred during interglacials. historic times by humans, see El- Further, we envision the following Sammak and Tucker, 2002). scenario: (1) The ooids were formed off The age of the Kattavia eolianite is the coast of Rhodes during the warm difficult to estimate, but we consider it interglacial periods. (2) During sea-level to be Late Pleistocene to Holocene in lowstands of the cold glacial periods, the age on the basis of the following criteria. ooids became aerially exposed on the (1) The presence of metastable carbonate shelf around the island and were phases such as aragonitic bioclasts; (2) reworked and transported inland as some of the snail shells retain their color eolian dune sands. (3) During the banding; (3) the weak cementation and subsequent warm interglacial, sea level 4), the comparable altitude above sea rose again and flooding of the island level to the east coast eolianites on shelf cut off the source of ooids. The Rhodes. However, McKee and Ward wetter and warmer climate favored soil (1983) pointed out that preservation and formation in the dune sands. However, degree of lithification are variable in this preliminary model for the formation both Holocene and Pleistocene dune and timing of the Kattavia eolianites is sands. While McKee and Ward (1983) so far only speculative and must be considered drops in sea level, attributed proved or disproved by precise dating of to the start of glacial phases, as a the eolianite. favorable condition for eolianite

6 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

VERTEBRATE TRACE FOSSILS In soft or loose tracking substrates, the trackmaker’s foot can plunge to a Terminology considerable depth, creating vertical or inclined walls from the bottom of the The original sedimentary surface in true track to the tracking surface. The which the tracks were emplaced is vertical parts of the track are termed termed the “tracking surface” sensu “trackwalls” (Brown, 1999) or “shaft” Fornós et al. (2002), and the animal (Allen, 1997). In cases where the responsible for the track is the trackwalls are inclined, the track at the trackmaker. The tracks formed in the surface appears wider than the true track tracking surface are termed “true tracks” at the bottom of the track and is termed and represent the direct impression of the “overall track” (Brown, 1999). When the trackmaker’s foot (Lockley, 1991). tracks are emplaced in dry, loose The weight of the trackmaker is sediments, the trackwalls collapse after transferred via the foot radially outward removal of the foot, destroying the shape into the sediment, causing deformation of the true track and in extreme cases not only of the tracking surface, but also leaving only a bowl-shaped depression of subjacent layers (Allen, 1989, 1997). on the sediment surface. In this case the The deformation structures formed in the overall track is the only information layers subjacent to the true track are available about the size of the termed “undertracks” (Lockley, 1991). trackmakers foot (Fig. 4B), and it is Undertracks preserve less detail than important to bear in mind that the overall true tracks, and they become track can appear significantly larger than successively shallower and wider, and the true track (Milàn, in press). If the preserve fewer anatomical features track subsequently is covered by several downward, layer by layer (Milàn and thinner layers of sediment, the layers Bromley, in press). The radial pressure will drape the contours of the track and of the foot further creates a “marginal can form a shallowing upward stacked ridge” of displaced sediment around the sequence of “ghost tracks” sensu Fornós tracks (Fig. 4A). et al. (2002). If observed on horizontal surfaces, these ghost tracks may be misinterpreted as undertracks (Fig. 4C).

7 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Figure 4. Schematic representation of track formation in dry sediments. A) The impression of the trackmaker’s foot forms the true track, TT, in the tracking surface, TS, while a stacked succession of downward-shallowing undertracks, UT, deforms the subjacent layers. Vertical or near-vertical trackwalls, TW, connect the tracking surface with the true track. A marginal ridge, MR, of laterally displaced sediment surrounds the track. B) After withdrawal of the foot, the trackwalls collapse, and the track appears as a bowl-shaped depression with an overall track, OT, width significantly larger than the true track. C) Subsequent sedimentary covering creates a stacked succession of ghost tracks, GT.

Tracks Exposed in Vertical Section tracks is layered, causing the formation of several distinct ghost tracks above the In the road cutting, 74 footprints true track (Fig. 6A,B). exposed in cross section were measured, and an additional 5 footprints seen in Tracks on Horizontal Surfaces cross section were measured in a small quarry on the other side of the road. The Since the eolianite is exposed as a footprints ranged in size from 4 to 37 cm road cut, morphologic variation of the in cross section (Fig. 5). All tracks tracks as expressed in horizontal view is appear as steep-walled, flat-bottomed limited. Only a few surfaces depressions in the sediment surfaces. representing the original bedding planes With very few exceptions they lack any are exposed, but in the bottom of the preserved morphologic details of the ditch along the road, tracks in several feet, such as division into digits or surfaces were exposed in horizontal hooves. view. Two of these were complete The sediment layers around and enough to record and, in addition, a slab below the flat-bottomed tracks have containing two footprints was found in a been deformed by the impression of the loose block on the other side of the road. feet, clearly leaving distinct undertracks below all the recorded specimens (Fig. 6). In several specimens the filling of the

8 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Figure 5. Size/frequency histogram of the 79 tracks recorded in vertical section. A) The histogram shows five distinct size groups: (1), small tracks from 4 to 8 cm in diameter, (2) tracks about 9 cm in diameter, (3) tracks from 12 to 22 cm in diameter, (4) tracks from 25 to 28 cm in diameter and (5) large tracks with a diameter between 32 and 37 cm. B) The five tracks recorded in the small quarry, fall into size groups 1 and 4. C) The tracks from unit 2 fall into size groups 3 and 4, and perhaps one track in group 2. D) The 61 tracks recorded from unit 3 include all 5 size groups.

9 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

10 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Figure 6. Five different sizes of tracks as they appear in the section, figures are not to same scale. A) Small track 4 cm wide. The smallest tracks appear mostly as v-shaped pinch structures protruding down from the tracking surface. B) The cross section through this track is 10 cm wide. Notice the well- developed undertracks (UT) in the subjacent horizons as well as the ghost tracks (GT) in the overlying layers created by gradual filling of the tracks. A prominent raised margin is seen on the right side of the track (arrows). C) Track 16 cm wide. The track is poorly preserved owing to little color contrast between the individual laminae. However, the shape of the track is defined by the undertracks (UT), indicated by arrows. D) Large 25 cm wide, bowl-shaped track. Notice the well-developed succession of downward- shallowing undertracks (UT) below the track. No ghost tracks are formed above the track. E) One of the largest tracks recorded from the section, with a diameter of 35 cm. This and other large tracks are interpreted as proboscidean (elephant) tracks. The track appears more box-shaped than the smaller ones. Undertracks (UT) are present to a considerable depth below the track.

The first is a small, circular track 7 that it is indeed a true track altered by cm in diameter, partly filled by erosion. The upper part reveals the sediment. The exposed parts of the track remains of a weak division into two show no internal structures or hooves (Fig. 7B). indications of the trackmaker’s anatomy A slab of grainstone found among owing to secondary collapse and filling loose blocks at the roadside opposite the of the footprint. A prominent pressure cutting contained two tracks, one of pad (sensu Fornós et al. 2002) of which was well preserved. This track is sediment displaced by the trackmaker’s 16 cm long, almost circular in outline foot is developed on the downslope side and consists of two broad crescent- of the footprint, and several sets of shaped impressions weakly divided concentric fractures in the sediment are anteriorly and posteriorly by a small visible out to six cm down the wedge of sediment. The bottom of the paleoslope of the dune the track was track is flat, with a slightly raised area in emplaced in. Radiating fractures in the the middle. The morphology of the two sediment around the track are formed in crescent shaped hoof imprints is the lower right side of the track (Fig. consistent with the foot morphology of 7A). an even-toed trackmaker. Surrounding A less well preserved and heavily the track is a prominently raised rim of weathered track is approximately 18 cm displaced sediment. The track walls are in diameter. The eroded state of the track collapsed, giving the characteristic low- makes it hard to identify whether it is a angle trackwalls of tracks emplaced in true track, an undertrack or a ghost track. dry sand (Milàn in press). Beside the But the vague remains of a raised margin track, a less well preserved track is around the track makes it most likely present, having a diameter of 12 cm. No

11 Milàn et al. – Vertebrate ichnofauna from Rhodes ______anatomical details have been preserved, around the track (Fig. 7C). but a partly raised margin is present

Fig. 7.—Four tracks were found exposed in horizontal view. Photos are not to scale. A) Track 7 cm wide. The track itself is partly filled with sediment and shows few morphologic details. The sediment surrounding the track was damp at the time of track formation, allowing the preservation of a series of concentric rings of displaced sediment and radiating fractures to be preserved around the track (arrows). B) Heavily weathered track 18 cm wide. The track shows no anatomical details, except for a vague division into two broad hooves, owing to the erosion of the grainstone. C) Slab containing two tracks. The first (middle of the picture) is 16 cm in diameter, divided into two distinct weakly divided crescent-shaped hoof impressions, which forms an almost flat bottom in the track. A prominent marginal ridge of displaced sediment is present surrounding the track. This track is interpreted as a camel track. The other track (below) is 12 cm wide, preserving no anatomical details. A partly preserved marginal ridge is present around the track.

12 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Figure 8. Root structures and possible invertebrate traces in the lowest paleosols. A) General total bioturbation of the paleosol. B-D) Various configurations of bush-sized roots preserved as rhizoconcretions.

Bioturbation and Root Traces The paleosols, however, clearly have an ichnofabric indicating total bioturbation The eolianitic grainstone commonly (Fig. 8A). Larger plant rooting structures shows a mottled or lumpy texture that are locally well-preserved as may have resulted from animal burrows rhizoconcretions (Fig. 8B-D) but, despite and plant roots. However, the degree of the good preservation of snail body weathering precludes any detailed study. fossils, smaller rooting structures and

13 Milàn et al. – Vertebrate ichnofauna from Rhodes ______bioturbation were too poorly preserved the mode of progress and the presence, for study on the weathered road cut absence and degree of erosion prior to surface. burial of the track; but especially to the consistency of the substrate. Eolian DISCUSSION sediments, especially if dry, are far from an optimal medium for track preservation. Among the tracks exposed in cross A track formed in totally dry sand will section, those that were considered collapse immediately after lifting of the “measurable” were those in which the foot, leaving only a hole vaguely true diameter could be estimated. resembling the outline of the foot, while Tangential sections were avoided as these even slightly damp sand can preserve gave no clear idea of the true diameter of well-defined tracks (Milàn in press). The the track. problem is intensified by the dominance The width and depth of the 79 of ooids in the sediment, rendering it measurable tracks were determined and highly unstable. Clear preservation of the plotted in a set of size/frequency shaft is probably only possible in various histograms, based on diameter, to show states of dampness of the substrate. the distribution of track sizes. The It must be admitted that, in weathered histograms show groupings of tracks vertical section some inaccuracy is indicating five distinct size-classes: (1) unavoidable in measurement of small tracks from 4 to 7 cm in diameter; “diameter”. It is usual that the orientation (2) a distinctive group having a diameter of the track is unknown and the width and of 9 cm; (3) tracks 14-21 cm across; (4) a length of the track are usually not the group of larger tracks 25-28 cm; and (5) same. Nevertheless, the fact that some very large tracks ranging from 32 to distinctive size-classes have indeed 37 cm (Fig. 5A). emerged from the measurement in hand In contrast, an attempt to plot the indicates that the measurements are diameter against depth of the tracks sufficiently accurate. showed no sign of the size-classes, As all the tracks, except for three, are indicating that diameter and depth of exposed in vertical cross section, it is not tracks were unrelated. Depth of a track is possible to identify the trackmakers with related to the weight of the trackmaker; any degree of accuracy, based on

14 Milàn et al. – Vertebrate ichnofauna from Rhodes ______morphologic features in the tracks. along the exposure. The tracks in unit However, the size and morphology of 2 range in width from 10 to 28 cm, and the bilobed track (Fig. 6C) bears of these, 7 of the tracks are of resemblance to the morphology of relatively similar size, between 14 and purported Pleistocene camel tracks from 17 cm. Three tracks are smaller, New Mexico (Lucas et al. 2002) and the having smaller cross sections between cameloid ichnogenus Lamaichnium 10 and 13 cm and the last three range macropodium, described from the from 23 to 28 cm in width (Fig. 5C). Miocene of California (Sarjeant and Unit 3 contains 61 measurable Reynalds 1999). Tracks and feet of a tracks, the majority of tracks recorded. living camel was studied for comparison The diameters of the tracks range from with the fossil track, and both the foot 4 to 37 cm, and as a prominent morphology and morphology of tracks difference from unit 2, a large number emplaced in dry coarse sand closely of the tracks (16) have a small cross resemble the specimen from the Kattavia section ranging between 4 and 7 cm. Aeolinite (Fig. 9). Five tracks peak out prominently with Although the main section is a diameter of 9 cm. Twenty-eight divided into three units by soil tracks fall between 10 and 23 cm with horizons (Fig. 2B), unit 1 is only the majority between 15 and 20 cm in poorly exposed along a short distance diameter. The next grouping consists at the bottom of the section. The five of 11 tracks between 25 and 28 cm and tracks recorded from the small quarry last is the grouping of 5 large tracks on the other side of the road may from 32 to 37 cm in diameter (Fig. belong to unit 1, but as the geometry 5D). of the dunes is complicated, this The distribution of tracks in units 2 cannot be proved at this stage. Of the and 3 show two different patterns. The measurable 5 tracks from the quarry, tracks in unit 2 are randomly the three are of similar size, 25 to 27 distributed along the profile, while the cm in diameter, and the two others tracks in unit 3 seem to cluster mainly respectively 5 and 8 cm (Fig. 5B). Unit in two groups in the section (Fig. 2B). 2 yielded 13 measurable tracks that seem to be randomly distributed all

15 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Figure 9. Foot and footprint of a living camel from the Zoological Gardens, Copenhagen. A) The camel foot consists of two broad, weakly divided digital pads, each terminating in short triangular claws. The foot has a sub-circular outline. B) Track from the same foot emplaced in the coarse sand within the camel paddock. Even if the footprint is partly collapsed in the dry sand, it retains the distinctive shape of the two weakly divided digital pads and the almost flat bottom of the track, compare with figure 7C.

As the vertical distribution of the group having sizes between 14 and 17 tracks in the two groups is up to 4 m, cm and a larger group having sizes the distribution cannot be explained as between 23 and 28 cm, the three a single trampling event, but perhaps smallest tracks between 10 and 13 cm this stacked succession of tracks represent perhaps a third, smaller represents a preferred migration route group of trackmakers. The or favorable topography, which was composition of track-sizes in unit 3 is maintained even after new sands were more varied than in the quarry and in deposited. unit 2. Especially striking are the large When comparing the size-groups of number of small tracks less than 10 cm tracks in the quarry, and the two track- in diameter and the group of large bearing units, the tracks in the quarry tracks up to 37 cm in cross section. indicate two different size-groups of Assuming from the preliminary trackmakers, one having sizes about 25 sedimentologic investigations that the to 27 cm and smaller trackmakers three eolian units represent three having sizes around 5 to 8 cm. Unit 2 different time periods, a difference in bears evidence of two, perhaps three faunal composition should be distinct sizes of trackmakers, one expected. The track fauna of unit 3

16 Milàn et al. – Vertebrate ichnofauna from Rhodes ______differs from that of unit 2 in that the basin south of Apolakkia (Fig. 1) have size variation in the tracks is much yielded an extensive mammal fauna larger. In unit 2 the tracks range from (van de Weerd et al.,1982), while a 10 to 28 cm in cross section, while the balanced mainland fauna is known size range in unit 3 is from 4 to 37 cm from Late Miocene and Pliocene in cross section. In particular, the localities of Rhodes (Desio, 1931; whole group of small tracks having a Boni, 1943; De Bruijn et al., 1970; cross section less than 10 cm is Meulenkamp et al., 1972), and altogether missing in unit 2 as are also according to Sondaar and Dermitzakis the very large tracks above 28 cm in (1982) also from Lower Pleistocene. cross section. On the basis of the The Damatria airport fauna includes a recorded sizes of the tracks in the balanced fauna with Equus and different units, it is evident that the Leptobos. This fauna should be placed size distribution of trackmakers, and in Pleistocene (Dermitzakis and thus animal diversity, was highest in Sondaar, 1978, p. 823). The Damatria the time period when unit 3 was formation is overlain by the Kritika deposited. formation which is considered Upper The camel track (Fig. 7C) is 16 cm Pliocene, in the sense of the marine in diameter and tracks near that size Mediterranean stratigraphy are the dominant size in units 2 and 3 (Meulenkamp et al., 1972; Dermitzakis and are in total the most abundant and Sondaar, 1978). The mammal track size (Fig. 5A). Interestingly, no fauna from the Apolakkia Formation, track in “camel” size was observed in described by van de Weerd et al. the quarry, but the exposure in the (1982), contains five genera of quarry is limited in extent and the 5 insectivores, one lagomorph species measurable tracks recorded there are and four genera of rodents. Large probably not representative for the carnivores are represented by hyaenas whole fauna of that unit. and canids, and in addition, Richter No skeletal remains have been (1997) found a canine from a found associated with the footprints in sabertooth cat. Ungulates are the eolianite, but exposures of the represented by Hipparion, Cervus, Pliocene Apolakkia Formation in the Gazella and indeterminate species of

17 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Rhinoceratidea and Proboscidea (van described from eolian sediment in de Weerd et al., 1982). The largest Mallorca (Fornós et al. 2002), and the tracks, having a diameter from 32 to presence of cervids and Gazella (Kuss 37 cm, are considered to be of 1975; Symeonidis et al., 1974; van de proboscidean origin, and abundant Weerd et al., 1982) in the paleofauna finds of skeletal material of Pliocene makes them possible trackmakers but, and Pleistocene proboscideans from the above mentioned, possibly demonstrates their presence on the only Cervus and the dwarf elephants island (Kuss 1975; Symeonidis et al., have been documented for the 1974; van de Weerd et al., 1982) and Pleistocene of Rhodes. The balanced later finds of proboscidean material late Ruscinian fauna of Damatria allowed the identification of Anancus airport became extinct during arvernensis (Theodorou et al., 2000). Pleistocene when Rhodes became an The endemic elephant Palaeoloxodon island ((Dermitzakis and Sondaar antiquus maidriensis has been 1978). discovered in 1974 at the Ladiko cave. No camel remains have hitherto been At present existing findings point to an described from Rhodes, but camel average size of elephant of about 180 remains are known from the Upper cm (Theodorou, 1983) but the Miocene of Turkey (van der Made et al., available elephant material from 2002), therefore, the possibility that Rhodes is not adequate to estimate the another even-toed animal could be size variation within the elephants responsible for the tracks must be from Rhodes. This means that we can considered. Cervids have been found expect findings of both larger and both from the Pleistocene fissure fillings smaller elephants than these already at the Kritinia coast, and from the available, and thus expand the range of Apolakkia Formation deposited during sizes of tracks expected to be of the Pliocene (van de Weerd et al. 1982), proboscidean origin. and as both camels and deer (in the The small to medium sized tracks, broad sense) are artiodactyls, their 4–10 cm, occurring in the quarry and footprints all reflect the even-toed in unit 3 are comparable in size and morphology of two hoofs. However, a shape to the Pleistocene caprine tracks number of differences distinguish

18 Milàn et al. – Vertebrate ichnofauna from Rhodes ______cameloid footprints from deer footprints. direction, would show the bottom of the The hoof of a deer is covered with horn, footprint to be W-shaped, and not flat- forming two distinct crescent shaped bottomed with sloping trackwalls as seen impressions facing each other, with a in both the fossil and the recent camel prominent gap between the digits, which tracks (Fig. 9a,b). Furthermore, the are preserved in the footprints as a raised shape of the herein described track is ridge. Furthermore, the outline of a deer consistent with the shape of cameloid footprint is normally oval to pear tracks described by Sarjeant and shaped, and the individual hoof Reynalds (1999) and Lucas et al. 2002. impressions have pointed tips. In All considered, this makes it most likely contrast, the foot sole of a camel consists that a camel is the trackmaker of the of two broad fleshy digital pads, which herein described footprint. Future finds are almost grown together, and thus of either skeletal material, or more tracks leaves footprints without a central raised and trackways might either strengthen or ridge (Fig. 9). Only reindeer, Rangifer weaken this identification. tarandus, has a footprint with a nearly A possibility for the occurrence of circular outline, but their hoof camel tracks on Rhodes is that they impressions are widely crescent shaped belong to camels introduced by humans. and divided by a broad central ridge Camels have never been found in an (Rezendes, 1999). The bottom of the unbalanced endemic fauna. Thus, they herein described footprint is flat without could represent a relict of an older, indications of a prominent central ridge. balanced, mainland population, or much The division of the two digital pads are more probably, these tracks might belong only indicated in the ends of the to camels introduced on Rhodes by man. footprint (Fig. 9). However, the last possibility would give a In a deer, even where the footprint very young age for the deposits, which was emplaced in dry sand, which would does not correspond with the herein collapse after removal of the foot, a described scenario for the formation of raised ridge formed between the digit the eolianite. Further studies by J. impressions would still be visible. A Titschack and colleagues may shed light vertical section through a deer footprint, on this problem when a definitive age of cut perpendicular to the walking the Kattavia eolinite is determined.

19 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

The abundance of cross sections sized tracks exposed in cross section, through tracks of similar size to the represents the first evidence of camels identified camel track, indicates that on Rhodes. camels were indeed present as part of the Pleistocene mammal fauna on Rhodes. ACKNOWLEDGEMENTS

CONCLUSION We are grateful to the Zoological Gardens, Copenhagen and especially to The new Kattavia eolianite is Michael O. Jørgensen and Claus suggested to be of Late Pleistocene to Petersen, for their hospitality and Early Holecone age and consists of three assistance with obtaining photos of distinct units separated by soil horizons camel tracks and feet. Lothar Vallon rich in rhizoliths and gastropod shells. (Stuttgart), Huriye Demircan (Ankara), The different units are considered to Inken Mueller-Töwe (Mainz) and Tina have been deposited during glacial Kock Kjeldahl (Copenhagen) provided periods and the soil horizons formed valuable aid in the field. We thank the during the warmer interglacials. reviewers, David B. Loope and Christain The eolianite contains ichnological A. Meyer, whose constructive reviews evidence of a diverse vertebrate fauna, helped clarify the scope of the paper. comprising footprints ranging from 4 to This paper is published with the 37 cm in diametric cross section. The permission of the Director of the size range is largest in unit 3, suggesting Institute of Geology and Mineral a larger diversity in the vertebrate fauna Exploration, Athens, permit 846/10.2.73. during that period. The largest of the tracks are REFERENCES interpreted to have been made by proboscideans, the medium sized Allen, J.R.L., 1989, Fossil vertebrate endemic Pleistocene elephants of the tracks and indenter mechanics: island. One well-preserved footprint, Journal of the Geological Society, exposed on a horizontal surface, is London, v. 146, p. 600–602. identified as a track from a camel, and Allen, J.R.L., 1997, Subfossil this, together with abundant similar- mammalian tracks (Flandrian) in

20 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

the Severn Estuary, S.W. Britain: Desio, A., 1931, Isole Italiane mechanics of formation, del’egeo: Memorie Descrittive preservation and distribution: della Carta Geologica d’Italia, v. Philosophical Transactions of the 24, p. 1–546. Royal Society of London, B, v. Dermitzakis, M., and Sondaar, P.Y., 352, p. 481–518. 1979, The importance of fossil Boni, A., 1943, Fauna di Hipparion a mammals in reconstructing Rodi: Palaeontographia Italica, v. paleogeography with special 41, p. 23–36. reference to the Pleistocene of Brooke, B., 2001, The distribution of Aegean Archipelago: Annales carbonate eolianite: Earth-Science Géologiques des pays Helléniques, v. Reviews, v. 55, p. 135–164. 29, p. 808–840. Brown Jr., T., 1999, The science and El-Sammak, A. and Tucker, M., 2002, art of tracking: New York, Berkley Ooids from Turkey and Egypt in Books, 219 p. the eastern Mediterranean and a Clemmensen, L.B., Fornós, J.J. and love-story of Antony and Rodríguez-Perea, A., 1997, Cleopatra: Facies, v. 46, p. 217– Morphology and architecture of a 228. late Pleistocene cliff-front dune, Fabricius, F.H. and Berdau, D., 1970, Mallorca, Western Mediterranean: Early Holocene oöids in modern Terra Nova v. 9, p. 251–254. littoral sands reworked from a Clemmensen, L.B., Lisborg, T., coastal terrace, southern Tunesia: Fornós, J.J. and Bromley, R.G., Science, v. 169, p. 757–760. 2001, Cliff-front aeolian and Fornós, J.J., Bromley, R.G., colluvial deposits, Mallorca, Clemmensen, L.B. and Rodriguez- Western Mediterranean: a record Perea, A., 2002, Tracks and of climatic and environmental trackways of Myotragus balearicus change during the last glacial Bate (Artiodactyla, Caprinae) in period: Bulletin of the Geological Pleistocene aeolianites from Society of Denmark, v. 48, p. 217– Mallorca (Balearic Islands, 232. Western Mediterranean): Palaeogeography,

21 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Palaeoclimatology, Palaeoecology, der Naturforschende Gesellschaft, v. 180, p. 277–313. Freiburg i Br. v. 65, p. 25–79. Frechen, M., Dermann, B., Boenigk, Lea, P.D., 1996, Vertebrate tracks in W. and Ronen, A., 2001, Pleistocene eolian sand-sheet Luminescence chronology of deposits of Alaska: Quarternary eolianites from the section at Givat Research, v. 45, p. 226–240. Olga - Coastal Plain of Israel: Lockley, M. 1991, Tracking Quarternary Science Reviews, v. Dinosaurs: Cambridge, Cambridge 20, p. 805–809. University Press, 238 p. Guern, P., and Davaud, E., 2005, Loope, D.B., 1986, Recognizing and Recognition of ancient carbonate utilizing vertebrate tracks in cross wind deposits: lessons from a section: Cenozoic hoofprints from modern analogue, Chrissi Island, Nebraska: Palaios, v. 1, p. 141– Crete: Sedimentology, v. 52, p. 151. 915–226. Lucas, S.G., Morgan, G.S., Hawley, Hanken, N.-M., Bromley, R.G. and J.H., Love, D.W. and Myers, R.G., Miller, J., 1996, Plio-Pleistocene 2002, Mammal footprints from the sedimentation in coastal grabens, Upper Pleistocene of the Tularosa north-east Rhodes, Greece: Basin, Doña Ana County, New Geological Journal, v. 31, p. 393– Mexico: New Mexico Geological 418. Society Guidebook, 53rd Field Hansen, K., 2001, Sedimentology of Conference, Geology of White Pliocene-Pleistocene temperate Sands, p. 285–288. water carbonates from northeast McKee, E.D. and Ward, W.C., 1983, Rhodes, Greece: Unpublished Eolian environment, In Scholle, Ph.D. Thesis, University of P.A., Bebout, D.G. and Moore, Copenhagen, 311 p. C.H., eds., Carbonate Depositional Kuss, S.E., 1975, Die Pleistozänen Environments: AAPG Memoir, p. Hirsche der ostmediterranean 131–170. Inseln Kreta, Kasos, Karpathos und Made, J. van der., Morales, J., Sen, S. Rhodos (Griechenland): Bericht and Aslan, F., 2002, The first camel from the Upper Miocene of

22 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Turkey and the dispersal of the Rhodes]: Varv, v. 1997(3), p. 63– camels into the Old World: 75. Comptes Rendus Palevol, v. 1, p. Richter, D.K., 1976, Gravitativer 117–122. Meniskuszement in einem Milàn, J., (in press), Variations in the holozänen Oolith bei Neapolis morphology of emu (Dromaius (Süd-Peloponnes, Griechenland): novaehollandiae) tracks, due to Neues Jahrbuch für Geologie und differences in walking pattern and Paläontologie – Abhandlungen, v. substrate consistency: 151, p. 192–223. ichnotaxonomical implications: Sarjeant, W.A.S. and Reynolds, R.E., Palaeontology. 1999, Camel and horse footprints Milàn, J. and Bromley, R.G., (in from the Miocene of California: press), True tracks, undertracks San Bernardino County Museum and eroded tracks, experimental Association Quarterly, v. 46, p. work with tetrapod tracks in 31–32. laboratory and field: Sondaar, P.Y., and Dermitzakis, M., Palaeogeography, 1982, Radiation of land vertebrate Palaeoclimatology, Palaeoecology. migrations to paleogeography and Muelenkamp, L.E., De Mulder, E.F.J., tectonics: International Symposium and Van De Weerd, A., 1972, on the Hellenic Arc and Trench Sedimentary History and (HEAT), April 8-10, Proceedings paleogeography of the late II, p. 283–308. Cenozoic of the Island of Rhodos: Symeonidis, N., Bachmayer, F., and Zeitschrift der Deutschen Zapfe, H., 1974, Endeckung von

Geologischen Gesellschaft, v. 123, Zwergelefanten auf Insel Rhodos p. 541–553. (Ausgrabungen 1973): Annalen Rezendes, P., 1999, Tracking and the des Naturhistorischen Museums in art of seeing, 2nd edition: New Wien, v. 78, p. 193–202. York, Harper Perennial, 336 p. Theodorou, G., 1983. The Fossil dwarf Richter, B., 1997, Sabelkatte og et elephants of Charkadio Cave, Tilos tandfund fra Rhodos [Sabertooth Island, Dodecanese. Ph.d. thesis in cats and a find of a tooth from

23 Milàn et al. – Vertebrate ichnofauna from Rhodes ______

Greek, p. 1-232, (Offset edition), Géologiques des pays Helléniques, Athens, Greece. v. 38C, p. 133–156. Theodorou, G., Spjeldnaes, N., Weerd, A. van de., Reumer, J.W.F. Hanken, N.-M., Lauritzen, S.E., and Vos, J. de., 1982, Pliocene Velitzelos, E., Athanassiou, A. and mammals from the Apolakkia Roussiakis, S., 2000, Description Formation (Rhodes, Greece): and taphonomic investigations of Proceedings of the Koninklijke Neogene Proboscidea from Nederlandse akademie van Rhodos, Greece: Annales Wetenschappen, Series B, v. 85, p. 89–112.

24 Jesper Milàn – Vertebrate Ichnology Appendix ______

Milàn et al. – A Late Holocene tracksite from Denmark ______

A LATE HOLOCENE TRACKSITE IN THE LODBJERG DUNE SYSTEM, NORTHWEST

JYLLAND, DENMARK

1 1 1 1 JESPER MILÀN , LARS B. CLEMMENSEN , BJØRN BUCHARDT and NANNA NOE-NYGAARD

1Department of Geography and Geology, Geology Section, University of Copenhagen, Oester Voldgade 10, DK-1350 Copenhagen K, Denmark. e-mail: [email protected], [email protected], [email protected] and [email protected]

Abstract - An exposed surface of a lake/bog deposits, dated to 1265 ± 145 cal. yr. BC (Early Bronce Age) contains a track assemblage comprising tracks from at least 6 different kinds of animals, including cattle, sheep or goat, red deer, horse and domesticated dog or wolf. The trackbearing lake/bog deposit is found within the Holocene Lodbjerg dune system on the North Sea coast of northwestern Jylland, Denmark. A number of tracks are further found preserved in cross section within the overlying dunes. The fauna composition indicated by the track assemblage are comparable with skeletal remains found from archaeological excavations of nearby settlements. The inland part of the eolian system is formed by stabilized parabolic dunes and a vegetated eolian sand plain. Seawards the dune system is bounded by a prominent coastal cliff with partly fixed cliff-top dunes. The dunefield is characterized as transgressive and evolved episodically since about 2200 BC. The aeolian system, which overlies a Weichselian till, is composed of alternating eolian sand units and peaty paleosols recording alternating periods of eolian activity and dunefield stabilization. The paleosols are typically of large lateral distribution and locally develop into thick peat lake/bog deposits. ______

1 Milàn et al. – A Late Holocene tracksite from Denmark ______

INTRODUCTION

The eolian Lodbjerg dune system is located in Thy, at the North Sea coast of the northwestern part of Jylland, Denmark (Fig. 1) (Clemmensen et al., 2001a). The dune system is bounded by a coastal cliff towards the sea lined with partly fixed cliff-top dunes. The inland part of the eolian system is formed by stabilized parabolic dunes and a vegetated eolian sand plain. Occasionally the wind breaks through the stabilized cliff-top dunes, causing the formation of inland migrating dunes. The dunefield can be characterized as transgressive and evolved episodically since about 2200 BC (Clemmensen et al., 2001a; Figure 1. Location map. The Lodbjerg dune system is located in the northwestern part of Jylland, Denmark, Pedersen and Clemmensen, 2005). The just north of the small town of Agger. The star marks the location of the tracksite approximately 8 kilometers coastal cliff is retreating and displays high north of Agger. Modified from Clemmensen et al. quality exposures of the eolian system, which (2001a). attains thicknesses of 10-15 m below the Reports of Cenozoic and especially present sand plain. The eolian system, which Holocene tracks and trackways are scarse in overlies a Weichselian till, is composed of the litterature, in contrast to Mesozoic, and alternating aeolian sand units and peaty especially dinosaur ichnology which has paleosols. The paleosols record periods of experienced an increase in interest within the dunefield stabilization and are typically of latest 30 years. Mesolithic footprints of bare- large lateral distribution (Pedersen and footed humans are described from estuarine Clemmensen, 2005). Locally the paleosol clay from the Severn Estuary, Wales develops into relatively thick peat deposits of (Aldhouse-Green et al., 1993). Early shallow lake origin. Holocene trackways from artiodactyls, birds and humans have been described from shore- near settlements in Argentina, where the footprints were found in association with

2 Milàn et al. – A Late Holocene tracksite from Denmark ______bones of sea mammals, artiodactyls and birds (Politis and Bayón, 1995). A large number of trackways of red deer, roe deer, aurochs and cranes and humans are described from intertidal silts and sands dating from the Neolithic to Bronze Age from the Formby Point at the Mersey estuary, northwest England (Roberts et al., 1996; Huddart et al., 1999). Tracks of humans and cattle walking together along the coastline together are known from lithified Roman age beachrock deposits from the Greek island of Rhodes

(Bromley et al. in press), and recently, Kim et al. (2004), have renewed the focus on Figure 2. The trackbearing surface is part of a laterally extensive peat horizon with a topographic relief Quaternary ichnology, and especially hominid ranging from present day sea level to approximately five meters above present day sea level. The tracks ichnology. were found on the uppermost surface of bog deposits in Animal tracks and traces from the the topographically lower parts of the peat horizon, close to present day sea level. The tracks were Lodbjerg dune system have previously been restricted to the low-lying parts of the bog deposits. The picture is taken from north and show the peat observed and preliminarily described both horizon rises to more than 5 meters above sea level in the background of the picture. Human for scale. within the eolian sand deposits and in the peaty paleosols (Clemmensen et al., 2001a, STUDY AREA Milàn et al. 2006). In this study we describe in detail an extensive track fauna from the The newfound tracksite is situated on upper surface of a peaty paleosol/lake bog the beach in the northern part of the Lodbjerg deposit, and draw comparisons between the eolian system, and fieldwork was carried out trackfauna and the skeletal fauna known from in November 2005. The stratigraphy at the excavations of nearby Bronze Age settlements site comprises a lowermost unit of (Liversage and Robinson 1992-93). Further Weichselian till draped by a well-developed we describe a number of single tracks peaty paleosol which is overlain by a encountered preserved in cross section within relatively thick eolian sand unit, with several the dune deposits. internal paleosols.

3 Milàn et al. – A Late Holocene tracksite from Denmark ______

Figure 3. Principal sketch of the coastal cliffs at the tracksite. The lake/bog deposit is located in a topographical low in the glacial till. The two exposed sites with tracks are located approximately 600 meters apart.

The top of the till and the overlying the low-lying parts of the paleosol, and no palaeosol dips gently northward and in this tracks were observed on higher-lying more direction the paleosol gradually increases in peaty areas of the surface (Fig. 3). This is thickness and develops into a lake/bog deposit interpreted as a taphonomic feature, as the up to 50 cm thick (Figs. 2, 3). Several similar nature of the substrate only allowed lake/bog basins are found southward from the trackformation in the softer, low-lying, shore- locality. These basins are likewise formed in near areas. topographical lows in the surface of the The top of the lake/bog deposit, glacial till, and in some of the basins the constituting the trackbearing surface, is thickness of the lake/bog deposits locally radiocarbon dated to 1265 ± 145 cal. yr. BC, reaches thicknesses in excess of three meters, which corresponds to the Early Bronce Age in but are most commonly developed to less than Denmark. Two samples from the bottom of one meter (Fig. 4). the lake/bog deposit were radiocarbon dated Two surfaces of the lake/bog deposit to determine the time span represented by the were exposed at beach level at the studied lake/bog deposition. One sample taken from locality, hereafter referred to as Site 1 and the roots penetrating down into the underlying Site 2 (Fig. 3). The surface was exposed as till gave an age of 3500 ± 140 cal. yr. BC and wave erosion had removed the overlying, a sample from the bottom of the char peat unconsolidated eolian sand (Fig. 2). The lake gave an age of 3410 ± 220 cal. yr. BC. Thus, bog deposits are semi-consolidated and the lake basins existed for approximately resistant to wave erosion, and the exposed 2200 years before it became buried by eolian surface has preserved a high density of animal sand. tracks (Figs. 5, 6). The tracks are restricted to

4 Milàn et al. – A Late Holocene tracksite from Denmark ______

Figure 4. A. The general stratigraphy of the lake/bog deposits measured in a fully developed sequence approximately 500 meters south of the tracksite. The lake basin is developed directly on top of the glacial till, and roots are penetrating up to 40 cm down into the underlying till. The packages of gyttja and peat deposits is here developed to a thickness of around 90 cm. The upper 37 cm of gyttja and char peat contains abundant eolian sand. The top of the gyttja is overlain by several meters of eolian sand. B. At tracksite 1, the package of peat and gyttja is only developed to a thickness of on average 30 cm. Directly on top of the till is an 8–15 cm thick layer of brown peat with wood fragments and roots penetrating up to 40 cm down into the underlying till. A thin layer of eolian sand separates the brown peat from the overlying layer of humified sandy peat. The tracks are all found in the very surface of the peat layer and are directly infilled by the overlying eolian sand. Scale on knife is 10 cm.

5 Milàn et al. – A Late Holocene tracksite from Denmark ______

The ages of the basal part of the overlying oligotrof, vegetated swampy wetlands and eolian sand has been determined to be ponds (Bech, 1997). Evidence from pollen between 700 and 200 years BC (Clemmensen show that the dominating vegetation during et al, 2001b). the Bronce age was oak (Querus), birch The onset of large-scale aeolian sand (Betula) and alder (Alnus) and abundant herbs movement in the Late Bronze Age to Early (Liversage and Robinson, 1995). This Iron Age was probably related to a marked landscape model is reflected by the studied climatic shift and related increase in exposures at the study locality at Lodbjerg. At storminess in the northwestern part of Jylland the studied site, the landscape morphology is taking place about 800-700 BC (Clemmensen largely controlled by the undulating et al, 2001b). This climatic event has also topography of the underlying Weichlian till been observed at Bjerre in northern Thy, with local highs and lows, ranging from where several Bronze Age settlements present day sea level to around six meters became covered by eolian sand at about 700 above present day sea level. Elsewhere the BC (Clemmensen et al., 2001b). In addition to topography of the landscape is an undulating the tracks on the surface of the lake/bog eolian system with partly deflated dunes deposit, a number of tracks are found exposed stabilized by vegetation and soil formation. in cross sections along several horizons The peat content of the soil horizon varies within the overlying eolian sand deposits. laterally and is concentrated in the When the Locality was re-visited a year topographic low areas of the horizon, while later in January 2007, the erosion from the the soil at the topographic highs have a much sea, had removed the seaward half of site 1 lower peat content. and completely removed destroyed site 2. The The present day rate of coastal erosion remaining parts of site 1 was so heavily worn is on average 2 meters each year (Liversage by wave swash, that all that remained of the and Robinson, 1992-93, Clemmensen et al., tracks was indistinct holes in the peat surface. 2001a). Assuming that the rate of coastal erosion has been constant since the Bronze The Bronze Age landscape Age, the study site would have been located The Bronce Age landscape of approximately 6 kilometers from the northwestern Jylland around what is now the paleoshore at the time the tracks were made. Lodbjerg area, was that of an open low-relief dune landscape interspersed with low-lying,

6 Milàn et al. – A Late Holocene tracksite from Denmark ______

VERTEBRATE TRACKS to a considerable depth, creating vertical or inclined walls from the bottom of the true Track terminology track to the tracking surface. The vertical The animal responsible for the track is parts of the track are termed “trackwalls” termed the “trackmaker”, and the surface the (Brown, 1999) or “shaft” (Allen, 1997). In animal emplaces its tracks within is the cases where the trackwalls are inclined, the “tracking surface” (sensu Fornós et al., track at the level of the surrounding tracking 2002). The tracks formed by the trackmaker surface appears wider than the true track at in the tracking surface are the “track” or the the bottom of the track and is termed the “true track” and represent the direct “overall track” (Brown, 1999). If the impression of the trackmaker’s foot tracking surface with the true tracks (Lockley, 1991). The weight of the subsequently becomes covered by several trackmaker´s foot is transferred radially thinner layers of sediment, the layers will outward into the sediment from the true track drape the contours of the true track and can causing deformation of not only the tracking form a shallowing upward stacked sequence surface, but also of the layers subjacent to of “ghost tracks” (sensu Fornós et al., 2002). the trackmakers foot below the tracking If encountered exposed on a horizontal surface (Allen, 1989, 1997). The surfaces, ghost tracks may be misinterpreted deformation structures formed in the layers as undertracks. subjacent to the true track are termed “undertracks” (Lockley, 1991). Undertracks Site 1 can be distinguished from true tracks in that The main exposure of the track-bearing they preserve less detail than the true tracks, peat horizon is exposed at beach level at the and they become successively shallower and base of the dune succession. The semi- wider, and preserve fewer anatomical consolidated peat horizon is more resistant to features downward from layer to layer erosion from the sea, than the unlithified dune (Milàn and Bromley, 2006). The radial sands, and thus a large surface had been pressure of the foot further creates a uncovered due to removal of the overlying “marginal ridge” of horizontally displaced eolian sands by wave erosion (Fig. 5). The sediment around the tracks. In cases where majority of the exposed horizon had been the track is emplaced in soft or loose exposed to extensive erosion from the sea, substrates, the trackmaker’s foot can plunge which has altered the appearance of the tracks

7 Milàn et al. – A Late Holocene tracksite from Denmark ______to shallow rounded, undiagnostic depressions Two tracks each comprising in the surface. The parts of the horizon, near four short, forward facing digit impressions the base of the cliffs, however were relatively and a single triangular “heel” pads were found unharmed by waves and exposed a high among the tracks. Each digit impression density of well-preserved tracks, up to 5 to 6 terminates in the impression of a short sharp tracks/m2 in the most heavily trampled areas. claw (Figs. 9f, 10). An area of 5x3 meters with the highest density of well-preserved tracks was mapped Site 2 directly on large sheets of transparent plastic, Half a kilometer further to the north and later scaled down to a map (Fig. 5c). The along the beach, another small exposure of high number of tracks and the large number the hardened peat horizon was exposed, of badly preserved tracks made it impossible measuring approximately 10x3 meters (Fig. to identify longer trackway segments from the 6a). The exposed surface was composed of same animals. similar dark lake/bog deposits of at least half The track assemblage comprises a meters thickness. Again the surface of the predominantly bilobed tracks from peat horizon was heavily trampled, and where artiodactyls, and ranges in size from 5 to 14 the tracks were not modified by wave swash, cm in length. The outline of the tracks is sub- the tracks could be identified as artiodactyls circular and the tracks consist of two crescent- tracks. The trackfauna on this exposure is not shaped hoof impressions separated by a raised as diverse in regard to size and shape as on area. Morphologically the tracks consist of the main exposure, and the tracks are all from different types of hoof impressions and varies artiodactyls with hoof lengths between 8 and from impressions of straight parallel hoof 14 cm. The shape of the tracks varies between impressions (Fig. 6b,c), to tracks comprising parallel, straight, blunt hoof impressions, to curved semilunate impressions facing each diverging, curved, pointy hoof impressions other, giving the tracks a ovate to sub-circular (Fig. 6b-d). circumference (Fig. 9b). A group of smaller In total 16 well-preserved tracks, 12 tracks 5 – 7 cm in length often occur as paired from Site 1 and 4 from Site 2, representing or partly overprinting manus-pes couples the full morphological specter of tracks (Fig. 9d-e). Among the bilobed tracks occur encountered at the tracksite, were selected and occasional tracks consisting of a single hoof- casted in Plaster of Paris. shaped imprint (Figs. 9c, 11).

8 Milàn et al. – A Late Holocene tracksite from Denmark ______

Figure 5. The main trackbearing peat horizon (Site 1) is exposed at beach level due to wave erosion of the overlying unconsolidated eolian sand. A. Overview photo of the largest of the exposed trackbearing peat horizons. The peat horizon continues out into the water. The photo is taken from the top of the coastal dunes and the broken line square indicates the area (15 m2) that was mapped in detail and where the tracks were unaltered by wave erosion. B. Part of the track-bearing surface showing a high density of well-preserved tracks. The tracks in the left side of the picture are hoof-shaped and presumably from an unshod horse, and the tracks in the right side consist each of two crescent-shaped impressions and are interpreted as cattle tracks. C. Map of the best preserved part of the tracksite. The mapped area measures 3x5 meters and has a track density of up to six tracks per square meter. The broken-lined rectangle indicates the photo from B.

9 Milàn et al. – A Late Holocene tracksite from Denmark ______

Figure 6. A. The small exposure of the trackbearing horizon (Site 2) was found approximately 500 meters north of the main track bearing exposure. B – D. Tracks from the small exposure.

Tracks in eolian sand undisturbed bedding plane (Fig. 7a-c). One of In addition to the extensive track fauna the tracks is shallow and displays a central encountered at the described peat horizon, a division into two lobes (Fig. 7a). The other number of tracks were at several instances cross sections encountered are approximately found exposed in cross section at various as deep as they are wide and the trackwalls stratigraphical levels within the overlying are steeply inclining towards the bottom of eolian dune and sand sheet deposits. The the tracks. A marginal ridge of displaced morphology of the tracks in cross section is material is recognizable in one of the that of a steep-walled, flat-bottomed specimens (Fig. 7b). depression, originating from an otherwise

10 Milàn et al. – A Late Holocene tracksite from Denmark ______

The sand layers subjacent to the impressions are deformed and bend below the tracks forming a stacked succession of undertracks below the track (Fig. 7b-c). The subsequent infilling of the tracks has occurred gradually and several generations of ghost tracks are evident in the sand layers in and above the true tracks.

INTERPRETATION OF TRACK FAUNA

The tracks found in the peat horizons at Site 1 & 2 are well-preserved and are preserved as true tracks, i.e. the direct impression of the trackmakers foot (Lockley, 1991). The tracks were emplaced directly in the soft peaty surface and subsequently infilled with eolian sands. There is no evidence of the tracks having been exposed to erosion before burial, as erosion would have altered and blurred the shape of the tracks and giving them a more undefined appearance

(Milàn and Bromley, 2006), and the infilling Figur 7. Tracks exposed in cross section from various stratigraphical levels in the eolian sands overlying the of eolian sand demonstrates that the tracks are Bronce Age peat horizon. A. Large shallow track consisting of two lobes divided by a central ridge. This indeed true tracks emplaced directly on the size and morphology is consistent with cross sections through the tracks from cattle (Loope 1986). Scale in peat surface, and not undertracks. The tracks centimeters. B. Cross section through track showing gradual infilling. Notice the vertical trackwalls, flat found preserved as cross sections within the bottom and the gradual infilling which has created a dune deposits show a characteristic stacked succession of gradually shallowing upward ghost tracks. Scale bar on knife handle 10 cm. morphology of eolian tracks in cross section, C. Cross section through two consecutive tracks spaced 15 cm apart. Scale bar on knife handle 10 cm. with steeply sloping trackwalls, the formation

of undertracks and several generations of

11 Milàn et al. – A Late Holocene tracksite from Denmark ______ghost tracks from the gradual infilling of the walking perpendicular to the dip direction, tracks (e.g. Loope, 1986; Allen, 1997). with a few animals walking down slope The eolian sand movements and burying towards the low lying parts of the lake/bog of soil horizons during the Bronce Age was basin. not restricted to the west coast of Denmark. A similar situation is described from Djursland, on the east coast of Jylland where intensive farming and deforestation resulted in extensive eolian sand movements burying the fields. Excavations at these sites have revealed soil horizons containing thousands of tracks and trackways of cattle as well as traces of Bronce Age agricultural tools, directly buried beneath the eolian sands (Broholm,

1946; Jensen, 2002). The quality and preservation of these tracks are similar to Figur 8. Diagram showing the progression directions of 42 identified tracks from site 1. The predominant freshly exposed tracks from Lodbjerg. directions are north- and southward, which is perpendicular to the northward dipping slope of the Within the mapped area at site 1 which lake/bog basin. had the highest concentration of well- preserved tracks, it was possible to determine Types of tracks the direction of progression in 42 out of the The tracks found on the peat horizons 57 mapped tracks. 15 of the tracks were too all fall within five distinct types. The most indistinctly preserved to allow determination common tracks are bilobed tracks of even- of direction, and only appeared as elongated toed artiodactyls, and among them there are at depressions in the surface. When the least three different types based on differences directions are plotted in a diagram (Fig. 8), in size and hoof-morphology. The largest of the majority of the tracks are heading either the tracks are 10-14 cm in length and comes east- or westward with a minor amount in two different variations. One variation has heading north- and southward. The hoofs that are proximally broadest and trackbearing part of the peat horizon is distally tapering toward the pointy tip (Fig. dipping gently northward, so the majority of 9a). The other type has banana to crescent the identified tracks are made by animals shaped hoofs of more equal width (Fig. 9b).

12 Milàn et al. – A Late Holocene tracksite from Denmark ______

Figur 9. The five main types of tracks from the trackbearing horizon, all tracks are shown to same scale. A. Large bilobed track. The hoof impressions are tear shaped and terminates distally in pointy tips. This type of track is identified as a red deer track. B. Large bilobed track, consisting of two crescent-shaped hoof impressions. Identified as tracks from domesticated cattle. C. Track consisting of a single hoof-shaped impression. Identified as the track from an unshod horse. D. Small bilobed track with parallel straight hoofs, identified as a track from either a sheep or goat. E. Partly overprinting manus-pes couple from either sheep or goat. F. Track from a dog or wolf.

The tracks with the pointy tips are in into the substrate to a depth of 5 cm. Any size and shape similar to tracks of the red deer dewclaws located near the main hoofs would Cervus elaphus (Bouchner, 1982). Tracks of have left impressions in the substrate behind red deer resemble tracks of wild boar (Sus the track. Based on this morphological scrofa) in size and shape, and can easily be evidence we interpret the large-sized bilobed confused. However, wild boar tracks mostly tracks with the pointy tips to be tracks of red comprise impressions of the dewclaws while deer. on the red deer, the dewclaws are located high The second type of large sized, bilobed on the metatarsus and rarely comes into tracks with curved hoofs of a more consistent contact with the ground (Bouchner, 1982). equal width and rounded tips, are identified as The tracks from this locality are impressed tracks from bovids, in this case most likely

13 Milàn et al. – A Late Holocene tracksite from Denmark ______domesticated cattle (Bos taurus), as the tips (Fig. 9d). These tracks are often found youngest aurochs (Bos primigenius) from the paired as manus/pes couples and in many northen Jylland dates back to the Mesolithic instances the two tracks are partly (Noe-Nygard et al., 2005). overprinting each other (Fig. 9e). There are several possible candidates to the origin of these tracks, but the configuration of the relative straight hoofs fits the shape of tracks from domesticated sheeps or goats (Bouchner, 1982), and bones from sheeps and goat are frequently found at excavation sites in the area (Bech, 1997). Roe deer, whose tracks are of comparable size, has more rounded hoofs producing a more heart-shaped track (Bouchner, 1982). The tracks consisting of a single hoof- shaped imprint (Fig. 9c, 11a) is from an unshod horse. In the best preserve of the horse tracks the complete hoof morphology of the horse is preserved. This is especially seen in a cast of the best preserved specimen (Fig. 11b). The last type is the tracks consisting of four forward facing, clawed digit impressions and a triangular metatarsal pad (Figs. 9f, 10). These tracks are interpreted as carnivore tracks, and due to size and shape identified as Figur 10. A. Plaster cast of one of the canid tracks B. Interpretative drawing based on the plaster cast. The tracks of canids, presumably domesticated track consists of four forward facing digits and a single dogs, but they could also be tracks of wolves. triangular metatarsal pad. The tracks preserved in cross section The third group of bilobed tracks is within the dune deposits are unidentifiable, as smaller with hoof lengths from 5 to 7 cm and they do not reveal any anatomical details, has parallel slightly curved hoofs with pointed except for one example showing a central

14 Milàn et al. – A Late Holocene tracksite from Denmark ______ridge, as evidence of a bilobed Footprints exposed in cross section are footmorphology (Fig. 7a). difficult to assign to any trackmaker, as a random section through a track can exhibit widely different morphologies according to the orientation of the section. Only very general features such as size or number of digits exposed in the cross section can be used to identify tracks in cross section (Loope, 1986; Milàn et al., in press). The cross sections from the present study are from 9 to 14 cm, measured at the bottom, the largest one being the one with evidence of a bilobed foot. Tracks of this size could be made by red deer, cattle or horse.

Archaeological finds from the area Archaeological finds from the Lodbjerg area have shown that humans lived in the study area in a number of periods since around 3200 BC (Liversage et al., 1987, Liversage and Robinson, 1992-93). Around 2200 BC people began to clear the forest in the area, leaving the area exposed to wind erosion and eolian sand movement during stormy periods. However, several times during periods of decreased storminess eolian sand movement diminished and vegetation cover of the dune system was re-established Figur 11. A. Track consisting of a single hoof-shaped impression. The track is from an unshod horse. B. and the area was re-populated. Plaster cast of the horse track. In the Late Bronze Age (1100-500 BC)

there was a considerable amount of settlement

in the Thy area. Settlements were permanent

15 Milàn et al. – A Late Holocene tracksite from Denmark ______year-round and the Bronze Age farmers along the North Sea coast of Denmark and probably kept livestock including herds of Holland. This fauna composition is very cattle, and sheep or goats as indicated by similar to the trackfauna from Lodbjerg, in numerous the finds of ruminant teeth and that tracks of large cattle are predominant, bones in midden deposits connected to the followed by smaller tracks from sheep or settlements, and by the occurrence of tracks of goats, and then single tracks of horse and these animals in the eolian deposits along the dogs. coastline (Liversage and Robinson, 1992-93;

Clemmensen et al., 2001a). CONCLUSION Although there is scant evidence of The Late Holocene Lodbjerg tracksite Early Bronce Age settlements in immediate contains a trackfauna comprising tracks of proximity of the Lodbjerg area, elsewhere in cattle, sheep or goats, red deer, horse and the region the constructions of settlements canids preserved on a buried Bronze Age peat began early in the Early Bronze Age and horizon overlain by eolian sand. Further, continued into the Late Bronce age, i.e. about tracks of animal are found preserved in cross 1500 to 500 BC (Bech, 1997). Excavations of section within the overlying eolian deposits Bronze Age settlements from a similar comprising the Lodbjerg dune system, landscape in Bjerre, close to Hanstholm demonstrating that animals inhibited the area located approximately 25 km north of also in the periods of dune activity. Lodbjerg (Fig. 1), revealed skeletal remains Site 1 of the exposed trackbearing bog of the Bronze Age livestock (Bech, 1997). deposit has a trackfauna consisting of tracks The domesticated fauna from Bjerre was from cattle, sheep or goats, horse and dogs, predominantly composed of cattle, with goats while the smaller site 2 contains tracks of and sheep constituting a much smaller part, sheep or goats and red deer. The track and few indications of horse and dogs. assemblage is in agreement with similar aged Interestingly no traces of pigs were present in skeletal remains of livestock excavated in the the fauna (Bech, 1997). A similar fauna is nearby area. Increased utilization and known from excavations in the Dutsch West- inclusion of ichnological data and methods, Friesland, which at that time consisted of low will provide valuable additional information lying near-marine wetlands (Bech, 1997), about fauna compositions, palaeoecology and suggesting that this was the common-most sedimentary conditions in the Lodbjerg dune fauna composition among Bronze Age settlers

16 Milàn et al. – A Late Holocene tracksite from Denmark ______system, and thus help create a more complete Philosophical Transactions of the Royal palaeoecological reconstruction of the area. Society of London, B, v. 352, p. 481–518. Bech, J.-H., 1997, Bronze age settlements on

ACKNOWLEDGEMENTS raised sea-beds at Bjerre, Thy, NW- The research of JM is sponsored by a Jutland. Internationale Archäologie, v. 38, Ph.d. grant from the Faculty of Sciences, p. 3–15. University of Copenhagen, Denmark. Bouchner, M., 1982, Der kosmos- Quaternary Sciences Radiocarbon Dating spurenführer. Artia, Praha., 272 p. Laboratory, Lund, Sweden provided 14C Broholn, H.C., 1946, Danmarks Broncealder dating of the track site. The Carlsberg 3. Copenhagen., 278 p. founded project ”The Mesolithic-Neolothic Bromley, R.G., Uchman, A., Milàn, J. and transition from hunter-gatherers to farmers - a Hansen, K.S., in press, Rheotactic result of changes in climate, sea-level, Macaronichnus, and human and cattle terrestrial environment, food resources, social trackways in Holocene beachrock, structure or cultural influence”, provided Greece: reconstruction of palaeoshoreline means for fieldwork and 14C datings. And orientation. Ichnos. finally Agger Badehotel provided excellent Brown Jr., T. 1999. The science and art of food and shelter from the harsh Danish tracking. Berkley., 219 p. autumn weather during the field work. Clemmensen, L.B., Pye, K., Murray, A. and Heinemeier, J., 2001a, Sedimentology,

REFERENCES stratigraphy and landscape evolution of a Aldhouse-Green, S.H.R., Whittle, A.W.R., Holocene coastal dune system, Lodbjerg, Allen, J.R.L., Caseldine, A.E., Culver, NW Jutland, Denmark. Sedimentology, v. S.J., Day, M.H., Lundquist, J. and Upton, 48, p. 3–27. D., 1993, Prehistoric footprints from the Clemmensen, L.B., Murray, A.S., Bech, J.-H. Severn Estuary at Uskmouth and Magor and Clausen, A., 2001b. Large-scale Pill, Gwent, Wales. Archaeologia aeolian sand movement on the west coast Cambrensis, v. 141, p. 14–55. of Jutland, Denmark in late Subboreal to Allen, J.R.L., 1997, Subfossil mammalian early Subatlantic time - a record of tracks (Flandrian) in the Severen Estuary, climate change or cultural impact? GFF, S.W. Britain: mechanics of formation, v. 123, p. 193–203. preservation and distribution.

17 Milàn et al. – A Late Holocene tracksite from Denmark ______

Fornós, J.J., Bromley, R.G., Clemmensen, southern Thy. Journal of Danish L.B., and Rodriguez-Perea, A., 2002, Archeology, v. 11, p. 39–56. Tracks and trackways of Myotragus Lockley, M. 1991. Tracking Dinosaurs. balearicus Bate (Artiodactyla, Caprinae) Cambridge University Press. Cambridge, in Pleistocene aeolianites from Mallorca 238 p. (Balearic Islands, Western Loope, D.B., 1986, Recognizing and utilizing Mediterranean). Palaeogeography, vertebrate tracks in cross section: Palaeoclimatology, Palaeoecology, v. 180, Cenozoic hoofprints from Nebraska. p. 277–313. Palaios, v. 1, p. 141–151. Huddart, D., Roberts, G. and Gonzalez, S., Milàn, J. and Bromley, R.G., 2006, True 1999, Holocene human and animal tracks, undertracks and eroded tracks, footprints and their relationships with experimental work with tetrapod coastal environmental change, Formby tracks in laboratory and field. Point, NW England. Quaternary Palaeogeography, Palaeoclimatology, International, v. 55, p. 29–41. Palaeoecology, v. 231, p. 253–264. Jensen, J., 2002, Danmarks Oldtid, Milàn, J., Avanzini, M., Clemmensen, L.B, Broncealder. Gyldendal, Copenhagen, 612 Garciá-Ramos, J.C. and Piñuela, L., 2006, p. Theropod foot movement recorded from Kim, J.Y., Kim, K.-S., Park, S.I. and Shin, Late Triassic, Early Jurassic and Late M.-K., 2004, Proceedings of the Jurassic fossil footprints. New Mexico international symposium on the Museum of Natural History and Science Quaternary footprints of hominids and Bulletin, v. 37, p. 352–364. other vertebrates, Jeju Island, Korea. 175 Milàn, J., Clemmensen, L.B., Buchardt, B. pp. and Noe-Nygaard, N., 2006, Tracking the Liversage, D., Munro, M.A.R., Courty, M.-A. Bronze Age fauna: preliminary and Nørberg, P., 1987, Studies of a buried investigations of a new Late Holocene Iron Age field. Acta Archaeologia, v. 56, tracksite, Lodbjerg dune system, p. 55–84. northwest Jylland, Denmark. Liversage, D. and Robinson, D., 1992-93, Hantkeniana, v. 5, p. 42–45. Prehistoric settlement and landscape Milàn, J., Bromley, R.G., Titschack, J. and development in the sandhill belt of Theodorou, G., In press. A diverse vertebrate ichnofauna from a Quaternary

18 Milàn et al. – A Late Holocene tracksite from Denmark ______

eolian oolite from Rhodes, Greece. In: Bromley, R.G., Buatois, L.A., Márango, M.G., Genise, J.F. and Melchor, R.N., (eds.), Sediment-organism interactions: A multifaceted ichnology. SEPM Special Publications. Milàn, J., and Bromley, R.G., In press, Do shod humans leave true tracks?. Ichnos. Noe-Nygaard, N., Price, T.D., and Hede, S.U., 2005, Diet of aurochs and early cattle in southern Scandinavia: evidence from 15N and 13C stable isotopes. Journal of Archaeological Science, v. 32, p. 855– 871. Pedersen, K., and Clemmensen, L.B., 2005, Unveiling past aeolian landscapes: A ground-penetrating radar survey of a Holocene coastal dunefeld system, Thy, Denmark. Sedimentary Geology, v. 177, p. 57–86. Politis, G.G. and Bayón, C., 1995, Early Holocene human footprints and sea mammals in the tidal zone of the Argentinian foreshore. Past, v. 20, p. 5–6. Roberts, G., Gonzalez, S. and Huddart, D., 1996, Intertidal Holocene footprints and their archaeological significance. Antiquity, v. 70, p. 647–651.

19 Jesper Milàn – Vertebrate Ichnology Appendix ______

Taxonomic disclaimer:

Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______Rheotactic Macaronichnus, and human and cattle trackways in Holocene beachrock, Greece: reconstruction of paleoshoreline orientation

Richard G. Bromley1 Alfred Uchman2 Jesper Milàn1 & Klaus Steen Hansen3

1Geological Institute, University of Copenhagen, Copenhagen, Denmark 2Institute of Geological Sciences, Jagiellonian University, Kraków, Poland 3Carl Johans Gade 13,3tv, DK-2100 Copenhagen, Denmark

Macaronichnus segregatis occurs in an upper foreshore conglomeratic to sandy beachrock dating from Roman times, on the east coast of Rhodes, Greece. In some laminae, M. segregatis is strongly oriented perpendicularly to the Recent and Holocene seashore. In other laminae, this trace fossil displays a winding to spiral course. These forms are introduced as new ichnosubspecies, M. segregatis lineiformis, M. segregatis maeandriformis and M. segregatis spiriformis, respectively. It is suggested that these middle-tier trace fossils were produced under conditions of different porewater flow. During higher energy periods, predominantly in winter, water movement perpendicular to the coastline caused M. segregatis to be oriented rheotactically. During periods of more or less stationary pore-water, predominantly in summer, M. segregatis assumed a spiral form. The winding form represents intermediate conditions. Closely beneath the Macaronichnus horizons a large burrow fill was found, probably referable to Psilonichnus isp. In the same beachrock are undertracks attributed to artiodactyls, associated with tracks and a trackway of human origin. The anatomy of the tracks indicates the direction of travel of the trackmakers, eastward and westward along-shore in all cases but one. The morphology of the tracks has been influenced by the gravitational effect of beach slope. Orientation, morphology and structure of trace fossils provide clear evidence that the sandstone and gravel unit is a beachrock and reveal the precise orientation of the paleoshoreline.

Keywords beachrock, paleoshoreline, trace fossils, Macaronichnus, human footprints, artiodactyl footprints, Holocene, Greece.

1 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

INTRODUCTION determine the orientation of a Defining the position and course of a paleoshoreline. In this paper, we present an paleoshoreline is commonly an important example where both position and problem in facies analysis. Determination orientation of a paleoshoreline can be of these parameters is a common task in oil determined on the basis of trace fossils in a prospection; coastal orientation is not clear geologic setting. The example comes always obvious from standard analyses, from a Holocene foreshore beachrock on but in some cases, significant evidence is the island of Rhodes, Greece, where provided by ichnology. For instance, parallel-oriented and randomly oriented Radwański (1969, 1977) reconstructed the Macaronichnus segregatis occur (localities course of a Dalmatinian-type Miocene 1-3 in Fig. 1). Additionally, a probable paleoshoreline in Poland on the basis of crustacean burrow and terrestrial mammal bioerosion. But while the position of the trace fossils, including human footprints, backshore is well defined by the were found at locality 1 (Fig. 1) in Psilonichnus ichnofacies (e.g., Pemberton February 2001. et al., 2001), it is not always easy to

Figure 1. Location map. The island of Rhodes is situated in the eastern Mediterranean Sea, southwest of Turkey.

2 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

During subsequent winters, gradual western end of Pefkos village, SW of erosion of the site was observed. A re- Lindos (Fig. 1). examination was undertaken in March 2005 that revealed several features that were not previously exposed. A combination of the two sets of data is presented here. Only two thirds of the originally mapped area was still present in 2005. The remaining structures had been modified by erosion, appearing less prominent and in lower relief than in 2001. However, the erosion had exposed a small human trackway, consisting of three consecutive footprints, which were not recognized in 2001. Preliminary results have been published as abstracts (Bromley Figure 2. Exposure of beachrock west of Pefkos (locality 1, Fig. 1). A, general view. B, detail of A. and Uchman, 2001; Uchman and Bromley, Note low-angle bedding and variation in grain size. 2003). Description and interpretation of A few other occurrences (localities 2 and 3 Macaronichnus segregatis and associated in Fig. 1) of the trace-fossil-bearing tetrapod, including human, footprints are beachrock are scattered over a distance of the main aims of this paper. It is worth about 1 km south of the main locality emphasizing that trace fossils are rarely (locality 1 in Fig. 1). They form a rocky reported from beachrock. They document coastal zone a dozen meters wide that is several interesting aspects of surface and elevated up to 1 m above present sea level. subsurface life on a sandy beach dated to The beachrock has been extensively the Roman period. exposed and partially destroyed by recent wave action. In places, large blocks have GEOLOGIC SETTING been pulled out by the action of the sea and The beachrock deposits studied here scattered about, exposing bedding-plane are exposed over a distance of a few surfaces and vertical sections (Fig. 2). hundred meters along the shore at the

3 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

beachrock was formed in the swash zone of the foreshore. Such an environment may be observed today in the vicinity of the studied locality 1 and along the neighboring shores of Rhodes. Beachrocks occur extensively on the Rhodes coast at several localities (Alexandersson, 1972). In the present study, however, careful inspection of these localities has generally revealed no trace

fossils. The beachrocks around Pefkos are Figure 3. Lithologic log of the beachrock at Pefkos (locality 1, Fig. 1, exposure in Fig. 2) showing almost the only exception. This distribution of Macaronichnus segregatis. distribution of trace fossils may be A section of beachrock was studied; correlated with the presence of finer- about 50 cm in thickness and bearing trace grained sand at Pefkos than elsewhere. The fossils (Fig. 3). It is composed of invertebrate trace fossils are restricted to conglomeratic, very coarse to medium- these finer-grained sediments. grained calcarenite containing a considerable amount of black grains and AGE DETERMINATION pebbles. Light grains are composed mostly Determination of the age of the of different limestones but include skeletal Pefkos beachrock was approached using grains, especially of red algae. Quartz three well-known techniques (Hansen, grains are less common. The black grains 1994). are composed mostly of ophiolitic detritus, (1) The beachrock is cemented by in which pyroxene crystals are common. high-Mg calcite, which was precipitated The calcarenites display large-scale low- after the trace fossils were emplaced. Since angle (4-9°) lamination dipping toward the the cementation has contributed to the sea. A current lineation occurs on the preservation of the trace fossils, the cement surface of some laminae. This lineation is and the trace fossils might be expected to perpendicular to the present shoreline. have similar ages. Samples of the micritic The sedimentary features, including cement were extracted and submitted to low-angle cross stratification (e.g., Frey Accelerator Mass Spectrometry (AMS) 14 and Howard, 1988), indicate that the using C.

4 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

(2) Conventional 14C dating methods from the algal crust gave a much younger were used on mollusc shells embedded in age than the crust within which they were the beachrock. These gave a much older embedded, which is unconvincing. But the age than the sediment they were embedded ages given by potsherds within the algal in and the results are discarded. crust are comparable with (407 to 707 AD) The beachrock is overlain by a thin or much younger (988 to 1188 AD) than the red-algal biolithite. This encrustation must beachrock age. Thus a “Roman age” for the post-date the beachrock, which first must beachrock cement and the trace fossils is have been exposed by coastal erosion, and indicated: 45 BC - 430 AD. dates should post-date trace fossil production. Samples of the algal crust and shells embedded therein were analysed TRACE FOSSIL MACARONICHNUS using conventional 14C methods. SEGREGATIS (3) Potsherds embedded within the algal crust were analysed using Morphology and orientation thermoluminescence dating methods. In the Pefkos beachrock, Again, the time of burning of the ceramics Macaronichnus segregatis Clifton and in the oven will have post-dated the Thompson, 1978 occurs on bedding surfaces emplacement of the trace fossils as virtually straight, winding, or looping considerably. burrow-fills, 2-3 mm in diameter, preserved It is well known that many sources of in full relief and composed of a core error and uncertainties must be considered surrounded by a mantle. The core is when dating material using isotopes or composed of sediment that is lighter in color thermoluminescence. Nevertheless, the than the host rock, lacking black grains. results obtained here give a clear indication Black grains are scattered in the surrounding of a Roman age for the cementation of the rock and concentrated along the trace fossil Pefkos beachrock. margin in the mantle (Fig. 4A). The most believable age range is provided by the AMS analyses of the cement itself, giving two samples an age of 45 BC to 175 AD and 249 to 430 AD. 14C dating of the shells from the beachrock gave a much older age and is discounted. Shells

5 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

Figure 4. Macaronichnus segregatis, locality 1. A, horizontal polished surface. Contrast enhanced image, showing distribution of light- and dark-colored grains. Dark mantle and light-colored core are visible, 190P2. B, C, M. s. lineiformis isubsp. nov., locality 1, B, holotype, MGUH 27804, C, field photograph.

Three different types of form meanders loosely (Fig. 5) and the Macaronichnus segregatis can be third is irregularly spiralled (Fig. 7). These distinguished on the basis of morphology two forms show little phobotaxis. The (Figs. 4-7), each of which receives three forms rarely occur together in the ichnotaxonomic treatment in the appendix. same lamina. They are normally restricted One form is linear and shows phobotaxis to individual laminae, but commonly and strong rheotaxis (Fig. 4B-C) at right deviate into neighboring laminae. angles to the shoreline (Fig. 8). The second Macaronichnus segregatis occurs

6 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______exclusively in medium-grained calcarenitic grained deposits of the beachrock. The laminae that comprise the most fine- trace fossil occurs gregariously in patches.

Fig. 5. Macaronichnus segregatis maeandriformis isubsp. nov. indicated by arrows, and M. s. lineiformis isubsp. nov., locality 1, A, holotype of M. s. maeandriformis, MGUH 27805, B, paratype of M. s. maeandriformis, 190P1.

Figure 6. A, Macaronichnus segregatis maeandriformis isubsp. nov., MGUH 27806, locality 1. B, M. s. maeandriformis and M. s. lineiformis. Field photograph, locality 1.

7 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

Figure 7. Macaronichnus segregatis spiriformis isubsp. nov., field photograph of the holotype (arrow), locality 1.

Ethologic interpretation Koyama (1983) observed the isopod Macaronichnus segregatis has been Excirolana chiltoni japonica burrowing in shown to be produced by organisms that intertidal Japanese sands in backswash flow. feed on epigranular microbial films deeply He suggested that this isopod had produced within well-oxygenated foreshore sands; Pleistocene oriented trace fossils that may be depths of activity within the sediment included in Macaronichnus. However, Nara normally range below 20 cm, but in extreme (1994) analyzed Macaronichnus and cases reach 1.5 m (Clifton and Thompson, experimented on Excirolana chiltoni, and 1978; Saunders and Pemberton, 1990b; concluded that this isopod cannot be the MacEachern and Pemberton, 1992, p. 68; tracemaker. Radwański et al. (1975, fig. 10) Pemberton et al., 2001). Clifton and described a Miocene trace fossil from Thompson (1978) showed in experiments Denmark that can be included in that the tracemaker can be compared to the Macaronichnus, and interpreted it as a trace recent polychaete Ophelia limacina, which of “?haustoriid amphipods” in reference to is common on the western coast of USA. the neoichnologic experiments by Howard This was confirmed by Gingras et al. (1999, and Elders (1970). The interpretation, p. 355). Another producer of Macaronichnus however, is not obvious, mostly because the may be the opheliid polychaete Euzonus concentric segregation of grains by mucronata (Saunders, 1989; Pemberton et amphipods is unclear (Nicolaisen and al., 2001, p. 126). Kanneworff, 1969).

8 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

as typical of middle shoreface bars as did Pollard et al. (1993), who suggested that subtidal bars or sand waves are the environments of Macaronichnus segregatis. Hubbard et al. (1999) indicated a shoreface environment for this ichnogenus but, in strong contrast, Maples and Suttner (1990) placed it even in an offshore environment. Figure 8. Rose diagram of the orientation of Gingras et al. (1998) noted “moderate Macaronichnus segregatis lineiformis. numbers of Macaronichnus” in prodelta Paleonvironmental interpretation deposits, and higher numbers in proximal Literature on the environmental range delta-front deposits of the Upper Cretaceous of Macaronichnus segregatis is confusing. in Alberta, Canada. It seems that the According to Clifton and Thompson (1978) “oxygenation window”, essential for deep this ichnotaxon occurs in the intertidal and sand burrowers as discussed by Saunders et shallow subtidal environments (foreshore al. (1994), is not confined to the foreshore or and upper shoreface). Saunders and upper shoreface, although these zones are Pemberton (1986, 1990a, b) described preferred for the Pefkos and other cases. sediments reworked with Macaronichnus Pemberton et al. (2001, p. 128) indicated segregatis from the middle to upper that Macaronichnus can occur in oxygenated foreshore. According to Saunders in sediments of tempestites subsequent to MacEachern and Pemberton (1992, p. 66), it storm events. On the whole, therefore, “appears to be more common near the upper Macaronichnus may be regarded as most shoreface-foreshore contact”. Similarly, typical of foreshore to shoreface Nara (1994) noted its occurrence in the environments. lower foreshore. According to Pemberton in Male (1992, p. 47), Macaronichnus segregatis is common on high-energy shorelines. However, Walker and Bergman (1993) suggested that Macaronichnus is typical of transgressive sandstones. Curran (1985) regarded Macaronichnus segregatis

9 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

OTHER TRACE FOSSILS According to its morphology, it is likely that the ghost crab Ocypode sp. produced A large burrow this trace fossil and the trace fossil may be A bed at a lower level than those related to the ichnogenus Psilonichnus. containing Macaronichnus exposed part of Today, ocypodid crabs are abundant on a large burrow, probably of invertebrate eastern Mediterranean beaches. origin (Fig. 9). Artiodactyl tracks This trace fossil occurs together with other tetrapod footprints at locality 1 (Fig. 10, footprint 14). It is determined to be Pecoripeda isp. (see appendix) and is represented by a single relatively well- preserved footprint in a poorly preserved trackway (Figs. 10, 11A, footprints 12–14, 23), which comprises a symmetrical V- shaped undertrack, composed of two ovoid lobes fused in the proximal part (Fig. 12). The inner edges of the lobes are almost straight. The angle between them attains 70°. The trace fossil is 55 mm in maximum width by 42 mm long. It is preserved as a positive, flat but sculptured epirelief, elevated about 3 mm above the Figure 9. Large crustacean burrow, top bedding- surrounding laminae. Similar but longer plane view, field photograph, locality 1. tracks have been illustrated by Kim et al. The structure is a slightly oblique, (2004, fig. 8) from the Late Quaternary parallel-sided, apparently cylindrical tuffaceous shoreline sediments of Korea. burrow fill, 55 mm wide, seen for about Tracks, in which width is greater 300 mm length, filled with coarser material than length, and having a wide angle than the surrounding sediment. Margins of between the digits, are produced by even- the fill are sharp. The lower termination is toed artiodactyls like cervids when hemispherical; the upper is broken. walking in soft substrates.

10 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

Figure 10. Map of the trackbearing beachrock surface as it appeared in February 2001 (A) and March 2005 (B). The north-western two thirds of the exposure were destroyed by wave erosion in the elapsing 4 years. The three footprints, 24–26, interpreted as a human trackway, thereby became visible. Other tracks were eroded and appear in less relief than previously (dotted outline). It is preserved as an indistinct shallow Domestic pigs also produce such depression that is filled with a coarser- tracks by walking in soft, moist sand (AU, grained sediment than the surrounding personal observation). The young age of calcarenite. The filling is coarsest in the the beachrock makes it likely that cervids toe end, where this part of the print or pigs produced the tracks. consequently is poorly preserved. The heel outline is semicircular, whereas the toe end Human footprints is pointed, suggesting the wearing of a A probable, single human footprint shoe. The pointed morphology could also (Fig. 13) was found in 2001 as an isolated, be the result of digit I being dragged flat depression, 280 mm long and through the sediment during the forward maximally about 100 mm wide at the swing of the foot. The track appears to be a expected base of the toes. In the heel true track filled with coarser, well- region, it is maximally about 90 mm wide. cemented sediment.

11 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

show a division between the heel and the front part of the footprint. The latter and least well-preserved part is only represented as an elongated impression. None of the studied footprints have preserved impressions of individual toes (Figs. 13-14).

Figure 11. Tracks at locality 1, field photographs. A, Pecoripeda isp. (tracks 12-14, 23), large, rounded undertracks (footprints 11, 18, 19), and the large, elongate undertracks (footprints 15, 20). B, Detail of A showing the large, rounded undertracks (footprints 18-19) with fluid flow of sediment carrying the coarse gravel upward toward the surface as pressure-pad-like structures. This shows influence of the slight slope of the beach (downslope vector is down the figure).

By 2005, wave erosion had destroyed two thirds of the mapped area and altered the relief of the remaining trace fossils, but had also exposed a new series of tracks interpreted as a human trackway (Figs. 10B, 14, footprints 24-26). The trackway consists of three consecutive tracks, here interpreted as the imprints of Figure 12. Some details of tracks at locality 1, field photographs. A, Pecoripeda isp. (track 14) and the right, left and right feet. The tracks are large, elongate undertracks (footprints 15, 20). B, aligned in a narrow gait almost directly in Detail of Pecoripeda isp. (track 14). front of each other with a pace angulation The best preserved print, 25, has a close to 180 degrees. Two of the tracks well defined heel print. However, there is

12 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______no emphasis on the major toe as is surface is pushed backward by the pressure characteristic of human feet unrestricted by of the foot, is described by Brown (1999) footwear (cf. Tuttle, 2004; Roberts, 2004). as a disk or disk-fissure; i.e., it is produced This may indicate that this trackmaker was when the forward movement of the foot wearing shoes. pushes a disk of sediment backwards. The orientation of the trackway is almost directly westward (Fig. 10B). In contrast, the single footprint (Fig. 13) is oriented parallel to the linear Macaronichnus and perpendicular to the modern seashore, indicating that the trackmaker was walking away from the sea.

Although the tracks in the trackway Figure 13. The isolated human footprint oriented parallel with Macaronichnus segregatis lineiformis segment are eroded, an average foot length isubsp. nov. (Ma), locality 1, field photograph. from the three tracks was measured to 19.5 All footprints in the trackway are cm. The stride length between tracks 24 preserved in positive relief, as a raised and 26 measured 74 cm and the step length pedestal of finer-grained sand. Such between prints 24 and 25 measured 37 cm pedestal preservation has been observed in from heel to heel. recent sediments and in snow (Linke, 1954; Reineck and Flemming, 1997). This A multiple trackway type of preservation is also known from On a single bedding plane, a short several fossil examples i.e. the Morrison distance below the horizon of the possible Formation, where the sediments around Psilonichnus, a considerable number of large sauropod tracks are eroded, leaving undertracks was exposed (Fig. 10). The the individual tracks as raised pedestals undertracks fall into three groups. (Barnes and Lockley 1994; Reineck and Flemming 1997). The middle track in the trackway is the best preserved, and it appears that the part of the footprint originating from the distal part of the foot is dislocated slightly backward. This phenomenon, where part of the tracking

13 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

Figure 14. The human trackway (footprints 24-26) as it appeared at locality 1 in March 2005. The tracks are set in a line, forming a narrow-gait trackway. A, field photograph. B, line drawing from A. Knife-handle shows centimeters. The large, rounded undertracks (2) are (1) Small structures that include the 20 cm in length. The six examples again undertracks referable to Pecoripeda isp. are arranged as belonging to a single (Figs. 10, 11A, footprints ?12, 13-14, 23) − trackway. However, the undertracks occur the most complete example, 14, is in rather coarse-grained sediment, described above (Fig. 11); (2) large, subjacent to the bed preserving the small rounded undertracks (Fig. 10, footprints 1, undertracks. They thus show little 4-5, ?6-7, 11, 18-19); and (3) large, morphological detail and probably elongate undertracks (Fig. 10, footprints represent sediment deformation at some 15, 20). distance below the level of the true track. The small undertracks (1) appear to Considering the fauna available in Roman be arranged as a single trackway. A few times on Rhodes, the size and rounded specimens, and one in particular, are fairly form might suggest cattle as the well preserved, being emplaced in trackmakers. relatively fine-grained sediment, and the The two elongate large undertracks size and morphology suggest the (3) likewise have been emplaced in coarse- trackmaker to be a pig-sized artiodactyl. grained sediment, and seem to represent a single trackway. They are narrower and

14 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______longer than the large, rounded undertracks such an interpretation to the Quaternary (lengths ca. 25 cm). Rather considerable Narita Formation, Japan. The Quaternary disturbance of successive beds of different Macaronichnus displays strong and grain-size (Fig. 11) suggests that these consistent parallel orientation in many prints represent undertracks at only a short localities, which may be interpreted as distance beneath the true track. perpendicular to the paleoshoreline. Considering the likely available The above interpretation can be trackmakers, the size, shape, and similarity applied with a restriction to the studied to the smaller tracks 24 to 26, these tracks Macaronichnus segregatis from Rhodes. are interpreted as undertracks from The eastern Mediterranean Sea, in contrast humans. to the Japanese Pacific coast, is microtidal, and tidally influenced behavior can be DISCUSSION discounted. However, the seasonally changing wave action can substitute for the Paleoshoreline orientation tidal water movement. Under fair-weather conditions, no porewater movement would Animals that feed in the manner of be expected at a few tens of centimeters the Macaronichnus segregatis producer below seafloor. However, winter storms burrow below the zone of physical may push water masses well inland, and disturbance, yet in oxygenated sediment. the resulting significant backswash Such conditions are found foremost in the drainage would be expected to influence “toe of the beach”, where the zone of the behavior of the tracemakers. The linear “nutrient convergence” is formed by low- form of Macaronichnus segregatis, aligned tide drainage from land and wave action at right-angles to beach slope, may from the sea (Saunders et al., 1994; represent high energy conditions. This may Pemberton et al., 2001). Observations of also be the case in the winding forms, Exirolana sp. burrowing towards the land which meander about a generally linear during low tide (Koyama, 1983), whether course. In contrast, the spiraled forms do or not this species is a Macaronichnus not show such influence. It is suggested, segregatis producer, clearly show behavior therefore, that the linear forms may be adapted to tidal conditions, and help to related to high-energy winter conditions explain the linear orientation of and the spirals to quieter summer Macaronichnus. Koyama (1983) applied

15 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______conditions. The winding forms may Tetrapod footprints indicate intermediate conditions (Fig. 15). Reports of Holocene human and cattle footprints in the literature are not frequent. Aldhouse-Green et al. (1993) described Mesolithic footprints of bare- footed humans from estuarine clay from the Severn Estuary, Wales. Early Holocene human footprints in association with bird and artiodactyl footprints have been described from shore-near settlements in

Argentina, where the footprints were found Figure 15. Relationships of the ichnosubspecies of together with a large number of bones, Macaronichnus segregatis to environmental conditions. especially from sea mammals, but also

artiodactyls and birds (Politis and Bayón, The distribution of Macaronichnus 1995). Direct association of human and segregatis is additionally controlled by the animal footprints is described from grainsize of sediment. In the beachrock Formby Point at the Mersey estuary, sediments described here, Macaronichnus northwest England, where 145 trackways is restricted to the finest-grained beds or of humans, red deer, roe deer, aurochs and laminae, of medium-grained calcilutite. cranes are preserved in intertidal silts and It may be concluded, therefore, that sands dating from the Neolithic to Bronze the orientation of Macaronichnus Age (Roberts et al., 1996; Huddart et al., segregatis in the Pefkos beachrock 1999). depends on the direction and degree of When two or more consecutive movement of pore-water within the footprints are preserved it is possible to sediment. The direction of this movement calculate the speed of progression using is perpendicular to the paleoshoreline. the formula conceived by Alexander Therefore, the oriented Macaronichnus can (1976), which applies to all tetrapods be considered an indicator of having an erect leg posture: velocity (m/s) paleoshoreline location and orientation. 0.5 1.67 -1.17 = 0.25g * SL *h , where SL is stride

length between two footprints from the same foot, and h is hip height, being the distance from the acetabulum to the

16 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______ground. When hip height is unknown, as in trackmaker. By using this value, the height most cases with fossil footprints, it can be of the individual responsible for the small estimated as a multiple of the foot length. trackway is 130 cm; this suggests a rather To obtain an average relationship small, perhaps juvenile human as between footprint length and hip height in trackmaker. Speed of progression of 2.6 humans, measurements were taken on 30 km/h is consistent with comfortable random students and members of staff at walking along a sandy beach. the Geological Institute, University of Copenhagen. This gave a relative hip Trace fossils in beachrock height of 3.6 * foot length and, when used Trace fossils are not commonly on the fossil trackway, a hip height (h) of preserved in beachrock. Frey and the trackmaker of 70.2 cm was calculated. Pemberton (1986) studied vertebrate traces With this value in the formula, the in recent beaches of the coastal islands of progression speed of the Pefkos Georgia, USA. They concluded that trackmaker is calculated as 0.72 m/s or subsequent marine and aeolian processes 2.59 km/h. The foot length of 19.5 cm and and invertebrate burrowing obliterate most estimated hip height of 70.2 cm are smaller tracks. Reineck and Flemming (1997) and than for adult humans. Roberts et al. Demathieu & Demathieu-Mallet (2000) (1996) calculated body height from the described traces on recent beaches and Neolithic footprints at Formby Point by emphasized the poor preservation potential using an average relationship that foot of such structures in that environment. length is 15% of the body height, and the Mountain (1966) described some traces of same relationship was used by Mietto et al. Holocene humans, probable hyena, and (2003) to estimate the height of hominid birds from sediments that were most trackmakers in Pleistocene ash deposits of probably classifiable as beachrock in South Italy. More accurate values exist for the Africa. Badve and Ghare (1984) described relationship between foot length and body Arenicolites isp., the large tubular trace height, differentiating between races and fossil Diplotuba holocenia, and bivalve gender (e.g. White, 1980; Charteris et al., borings from a Holocene beachrock in 1982; Tuttle et al., 1990) but, as these are India. Indeed, bioerosion traces, in having unknown factors in the herein described a much higher preservation potential, are material, the average value of 15% is relatively well known in beachrock (e.g., sufficient to estimate the height of the

17 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

Ginsburg, 1953; Rice, 1971; McLean, cattle), and to humans, interrelationships 1974; Ekdale et al, 1984, p. 205). between the different undertracks, their Preservation preservational similarities, and the near- Macaronichnus was produced within identical direction of travel suggest that the the sediment, probably at least a few tens tracemakers may have headed eastward, of centimeters subsurface. It is exposed on possibly as a group. The disturbance the lamination surfaces because of structures associated with the large round selective erosion, its fill being better and the elongated types of undertracks cemented than the surrounding rock and clearly indicate pressure that would therefore more resistant to erosion. indicate travel in an eastward direction. The mammal footprints, preserved in This is particularly well shown by the large pedestal relief, are undertracks sensu round undertracks (Fig. 11, tracks 18 and Goldring & Seilacher (1971), Lockley 19), the structures resembling the “pressure (1991) and Milàn & Bromley (2006). Only pads” described by Fornós et al. (2002). the exceptionally deep undertracks seem to The orientation of the pressure pads also have been preserved. This may be shows influence of the slight slope of the explained by the softness of the sediment beach, in that a gravitational downslope and the degree of impact by the foot of the vector is also displayed (Fig. 11B). tracemaker. Most of the tracks are isolated. The orientation of the three tracks However, the three tracks 24-26 probably ascribed to a small human, however, points constitute a human trackway segment, and westward. But the orientation of the possibly 1, 4, 5, 11, 18 and 19 represent trackway parallels that of the eastward- fragments of the trackway of a large heading tracks. Thus, multiple use of a artiodactyl, probably an ox. The beach pathway may be indicated. The Pecoripeda track is preserved in positive isolated human track (Fig. 13) is the sole relief. This track may be part of a trackway exception to this east-west trend. in connection with the linear distribution of tracks 6, 7, 13, 14 and 23. CONCLUSIONS Although beaches today are Direction of travel and effect of beach frequently visited by continental slope tracemaking animals, and inhabited by Considering the group of undertracks marine endobenthic animals, sharing the ascribed to artiodactyls (probably pigs and same substrate, the periodically high

18 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______energy of eolian, tidal and wave processes sand substrate, as a response to internal of the intertidal regime renders this current regimes within the porosity of the environment generally unsuitable for the sediment. During higher energy periods preservation of, at least shallow-tier, trace (predominantly in winter) water movement fossils. Nevertheless, the co-presence of perpendicular to the coastline caused M. marine invertebrate trace fossils and segregatis to be oriented likewise. During mammalian tracks in early-cemented, periods of stationary pore-water marginal marine sediment is indicative of (predominantly in summer), the M. beachrock. segregatis assumed a spiral form. An The relatively fine-grained sand beds intermediate, meandering form and lamellae in beachrock at Pefkos, characterized by sinuous meanders Rhodes, together with the microtidal indicates intermediate porewater hydraulic environment, have allowed the movement. preservation of shallow-tier mammalian Three new ichnosubspecies of trace fossils in beachrock at this locality. Macaronichnus segregatis are herein These are identified as the footprints of designated as M. segregatis lineiformis, M. artiodactyls (probably pigs and cattle), and segregatis spiriformis and M. segregatis humans. A multiple trackway of these maeadriformis, on the basis of the three mammals indicates an eastward direction morphotypes. These forms of M. of progress. Three tracks attributed to a segregatis probably represent foraging small human, however, show a westward optimization under different regimes of direction, suggesting that the pore-water movement. trampleground represents multiple use of a The orientation of M. segregatis coast-parallel beach pathway. lineiformis, perpendicular to the beach The polychaete trace fossil slope, provides an accurate indicator of Macaronichnus segregatis is preserved in paleoshore orientation. This is supported medium-grained laminae and beds within by the longshore orientation of the the otherwise coarser-grained beachrock. mammalian trackways. The microtidal regime allows sea- Samples of the micritic high-Mg level to remain stationary for much longer calcite cement of the beachrock were periods than on tidal beaches. This may be analysed for 14C dating using Accelerator the reason for the special orientation of Mass Spectrometry. This gave calibrated middle-tier M. segregatis within the fine- ages of 45 BC to 175 AD and 249 to 430

19 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

AD, indicating cementation in the Roman Type ichnospecies. Macaronichnus period. segregatis Clifton & Thompson, 1978 ACKNOWLEDGEMENTS New diagnosis. Generally winding, Field research was supported by the unbranched burrow fills having a zoned Danish Research Council (RGB, KSH). backfill structure comprising an outer mantle Additional support was given by the composed of dark-colored mineral grains Jagiellonian University (DS funds) (AU). and the central core of pale-colored grains. Dating by thermoluminescence was carried out at Risø National Laboratory and Macaronichnus segregatis Clifton & assistance from Vagn Mejdahl is much Thompson, 1978 appreciated. 14C dating was accomplished (Figs. 4-7, 13, 15) at the Danish National Museum and the Geological Survey of Denmark and Macaronichnus segregates segregates Greenland. 30 students and members of the isubsp. nov. staff at the Geological Institute, University of Copenhagen, kindly participated with Diagnosis. As for ichnogenus. measurements of foot lengths and hip Description. See section above, “Trace heights. Marco Avanzini and Jeong Jul fossil Macaronichnus segregatis. Kim provided critical reviews and spiked Morphology and orientation”. several weak points in the manuscript. This Discussion. In their original paper is published with the approval of the designation of M. segregatis, the authors Director of the Institute of Geology and did not include a diagnosis (Clifton and Mineral Exploration, Athens, under project Thompson, 1978). A new diagnosis is no. 846/10.2.1973. supplied above. The material on which M. segregatis was based occurred in unlaminated sediment. The trace fossil APPENDIX: ICHNOTAXONOMY therefore is not aligned along bedding planes as is the beachrock material. However, it Richard G. Bromley and Alfred Uchman was stated that the trace fossils “tend to be preferentially oriented about a horizontal Macaronichnus Clifton & Thompson, 1978 plane” (Clifton and Thompson, 1978, p. 1293). With the designation of three new ichnosubspecies this originally described

20 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______form of M. segregatis automatically receives the end-member forms are distinctive, we ichnosubspecies status as M. segregatis define the new ichnotaxa at segregatis. ichnosubspecies rank on the basis of these Saunders and Pemberton (1986) characters. distinguished an additional ichnospecies of Macaronichnus, a spiral form, as Macaronichnus segregatis lineiformis Macaronichnus spiralis (see Saunders and isubsp. nov. Pemberton, 1990b, fig. 1; MacEachern and (Figs. 4B. C, 5, 6B, 13, 15) Pemberton, 1992, fig. 6D) that was described in field guidebooks that cannot Diagnosis. M. segregatis in which the be recognized as publications in the sense trace fossil shows a strong, parallel of ICZN (1999). M. spiralis must therefore orientation along bedding planes. be considered a nomen nudum. Pemberton Deriviation of name. From Latin, linea, et al. (2001, fig. 103) repeated this name line; and forma, shape. but made no formal designation of type Description. Rigor of straightness varies material. The name therefore remains from high, where the tubes are virtually informal. straight, to weak, where undulations deviate We consider that the three the course of the burrow. morphological forms described in the Holotype. MGUH 27804 (Fig. 4B), section above, “Trace fossil housed in the Geological Museum of the Macaronichnus segregatis. Morphology University of Copenhagen, Denmark. and orientation”, represent different Type horizon and locality. Holocene ichnosubspecies, because they are united in beachrock. Rocky shore at the west end of their possession of the core-and-mantle the village of Pefkos, Rhodes, Greece internal structure that characterizes M. (locality 1, Fig. 1). segregatis Clifton and Thompson. Although the end members are highly Macaronichnus segregatis maeandriformis distinctive, the trace fossils grade into each isubsp. nov. other, the different specimens displaying (Figs. 5-6, 15) different degrees of spirality and meandering. In uncommon cases, Diagnosis. M. segregatis that sinuously individual trace fossils show continuity meanders along bedding planes. from one form to another. Thus, although

21 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

Deriviation of name. From Latin, nomen nudum 1990 Macaronichnus spiralis maeander, meander; and forma, shape. – Saunders and Pemberton, 1990b, p. Description. There is a degree of 47, pl. 10, figs. 1, 7, 9, 10, textfig. variation from strict sinus-wave morphology 13. to regular to irregular sinusoidal waves and nomen nudum 1992 Macaronichnus spiralis loops. – MacEachern and Pemberton, p. 68, Holotype. MGUH 27805(Fig. 5A), fig. 6D. housed in the Geological Museum, nomen nudum 1993 Macaronichnus spiralis University of Copenhagen; paratype – Walker & Bergman, p. 845. (190P1) (Fig. 5B) housed in the Institute of nomen nudum 2001 Macaronichnus spiralis Geological Sciences, Jagiellonian – Pemberton et al., p. 128, fig. 103. University, Kraków. 2001 Macaronichnus – Pemberton et al., p. Type horizon and locality. Holocene 128, fig. 104D, F, G. beachrock. Rocky shore at the west end of Pefkos, Rhodes, Greece (locality 1 in Fig. 1). Diagnosis. M. segregatis in which the Discussion. Judging from the external morphology is dominated by spiral figures morphology alone, the winding backfills degenerating into loops, closely following display features of Helminthopsis Heer. bedding planes. However, the active segregation of grains Deriviation of name. From Latin, spira, and flakes to produce a concentrically zoned spiral; and forma, shape. backfill is typical of Macaronichnus Description. The spiral development of segregatis Clifton & Thompson (1978). the trace fossil dominates over winding and looping portions of the structure. Macaronichnus segregatis spiriformis Holotype. MGUH 27807 (Fig. 7), isubsp. nov. housed in the Geological Museum, (Figs. 7, 15) University of Copenhagen. Type horizon and locality. Holocene nomen nudum 1986 Macaronichnus spiralis beachrock. Rocky shore at the west end of – Saunders and Pemberton, p. 47-50, Pefkos, Rhodes, Greece (locality 1 in Fig. 1). pl. 10, figs. 7, 9-10, textfig. 13. Discussion. Judging from the external nomen nudum 1989 Macaronichnus morphology alone, the meandering backfills segregatis spiralis – Saunders, p. 89, display features of Gordia Emmons. figs. 26f, g, 27. However, the active segregation of grains

22 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______and flakes to produce a concentrically zoned Wales. Archaeologia Cambrensis, 141: backfill is typical of Macaronichnus 14–55. segregatis Clifton & Thompson (1978). Alexander, R. M. 1976. Estimates of speeds of dinosaurs. Nature, 261: 129– Pecoripeda Vialov, 1966, emend. Sarjeant 130. and Langston, 1994 Alexandersson, T. 1972. Mediterranean Pecoripeda isp. beachrock cementation: marine (Figs. 10, 11A, 12) precipitation of Mg-calcite. In Stanley, Description. As in the section D. J. (ed.), The Mediterranean Sea: a “Artiodactyl track” above. natural sedimentation laboratory. Discussion. Ichnogenus Pecoripeda Dowden, Hutchinson & Ross: and its ichnosubgenus Ovipeda were Stroudsburg, Pennsylvania, p. 203–223. distinguished by Vialov (1965), but lacked Badve, R. M. and Ghare, M. A. 1984. a definition or a holotype for the Holocene trace fossils from beach rock ichnospecies. These were supplied in of Vales coast, Raigad District, Vialov (1966). Therefore, the latter is Maharashtra. Biovigyanam, 10: 165– considered as the proper year of 172. designation of Pecoripeda. However, Barnes, F. A. and Lockley, M. G. 1994. Sarjeant and Langston (1994) rejected Trackway evidence for social sauropods Vialov’s (1966) subgenera of this from the Morrison Formation, Eastern ichnogenus, and emended the ichnogenus Utah (USA). Gaia, 10: 37–41. Pecoripeda. Bromley, R. G. and Uchman, A. 2001. Young Holocene beachrock ichnofabrics from Rhodes, Greece. VI REFERENCES International Ichnofabric Workshop, Isla Margarita & Puerto La Cruz, Aldhouse-Green, S. H. R., Whittle, A. W. Venezuela, July 14-20, 2001. Sociedad R., Allen, J. R. L., Caseldine, A. E., Venezolana de Geólogos, PDVSA, [1 Culver, S. J., Day, M. H., Lundquist, J. p., no pagination]. and Upton, D. 1993. Prehistoric Brown, T. 1999. The Science and Art of footprints from the Severn Estuary at Tracking. Berkley Books, New York. 219 Uskmouth and Magor Pill, Gwent, p.

23 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

Charteris, J., Wall, J. C. and Nottrodt, J. Holocene barrier islands of Georgia. W. 1982. Pliocene hominid gait: new Geological Magazine, 125: 621–640. interpretations based on available Frey, R. W. and Pemberton, G. S. 1986. footprint data from Laetoli. American Vertebrate Lebensspuren in intertidal and Journal of Physical Anthropology, 58: supratidal environments, Holocene 133–144. Barrier Islands, Georgia. Clifton, H. E. and Thompson, J. K. 1978. Senckenbergiana Maritima, 18: 45–95. Macaronichnus segregatis: A feeding Fornós, J. J., Bromley, R. G., Clemmensen, structure of shallow marine polychaetes. L. B. and Rodríges-Perea, A. 2002. Journal of Sedimentary Petrology, 48: Tracks and trackways of Myotragus 1293–1302. balearicus Bate (Artiodactyla, Caprinae) Curran, A. H. 1985. The trace fossil in Pleistocene aeolianites from Mallorca assemblage of a Cretaceous nearshore (Western Mediterranean). environment: Englishtown Formation of Palaeogeography, Palaeoclimatology, Delaware, U.S.A. In Curran, A. H. (ed.), Palaeoecology, 180: 277–313. Biogenic structures: Their use in Gingras, M. K., MacEachern, J. A. and interpreting depositional environments. Pemberton, S. G. 1998. A comparative Society of Economic Paleontologists and analysis of the ichnology of wave- and Mineralogists, Special Publication, 35: river-dominated allomembers of the 261–276 Upper Cretaceous Dunvegan Formation. Demathieu, G. R. and Demathieu-Mallet, J. Bulletin of Canadian Petroleum 2000. Curiosités ichnologiques sur les Geology, 46: 51–73. plages de Languedoc. Importance du Gingras, M. K., Pemberton, S. G., vent. Société d’Histoire Naturelle Saunders, T. D. A. and Clifton, H. E. d’Audin, Bulletin, 170: 7–42. 1999. The ichnology of modern and Ekdale, A. A., Bromley, R. G. and Pleistocene brackish-water deposits at Pemberton, S. G. 1984. Ichnology - trace Willapa Bay, Washington: variability in fossils in sedimentology and stratigraphy. estuarine settings. Palaios, 14: 352– Society of Economic Paleontologists and 374. Mineralogists, Short Course, 15: 317 p. Ginsburg, R. N. 1953. Intertidal erosion on Frey, R. W. and Howard, J. D. 1988. the Florida Keys. Bulletin of Marine Beaches and beach-related facies, Science of the Gulf and Caribbean, 3: 55–69.

24 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

Goldring, R. and Seilacher, A. 1971. Conference, C.D. Volume, Paper 1229, Limulid undertracks and their p. 1–4. sedimentological implications. Neues Huddart, D., Roberts, G. and Gonzalez, S. Jahrbuch für Geologie und 1999. Holocene human and animal Paläontologie, Abhandlungen, 137, 422– footprints and their relationships with 442. coastal environmental change, Formby Hansen, K. S. 1994. En sedimentologisk, Point, NW England. Quaternary cementpetrografisk og -geokemisk International, 55: 29–41 beskrivelse og tolkning af tre holocaene ICZN (International Commission for og en pleistocaen beachrock fra den Zoological Nomenclature) 1999. graeske ø Rhodos. Unpublished M.Sc. International Code of Zoological Thesis, Geological Institute, University Nomenclature, adopted by the of Copenhagen. International Union of Biological Harbury, N. A. and Hall, R. 1988. Sciences, 4th ed. 232 p., London Mesozoic extensional history of the (International Trust for Zoological southern Tethyan continental margin in Nomenclature). the SE Aegean. Journal of the Kim, J. Y., Kim, K. S., Lee, C. Z. and Lim, Geological Society, London, 145: 283– J. D. 2004. Occurrence of hominid and 301. other vertebrate footprints of Jeju Howard, J. D. and Elders, C. A. 1970. Island, Korea. In Kim, J. Y., Kim, K. - Burrowing patterns of haustroiid S., Park, S. I. and Shin, M. -K. (eds.), amphipods from Sapello Island, Georgia. Proceedings of International In Crimes, T. P. and Harper, J. C. (eds.), Symposium on the Quaternary Trace fossils. Geological Journal, Special Footprints of Hominids and Other Issues, 3: 243–262. Vertebrates. Namjejugun, p. 1–26. Hubbard, S. M., Gingras, M. K. and Koyama, S. 1983. A new method for Pemberton, G. S. 1999. The Bluesky reconstructing of palaeogeography Formation in the Cadotte Region of using burrow orientation created by Alberta: an example of an economically Excirolana chitoni japonica significant estuarine deposit. (Thielmann). Journal of the Geological Geoscience Canada, G.A.C/C.S.P.G. Society of Japan, 89: 117–123. Annual Meeting, Calgary, Core Lewis, D. W. and Titheridge, D. G. 1978. Small scale sedimentary structures

25 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

resulting from foot impressions in dune Colorado Journal of Paleontology, 64: sands. Journal Sedimentary Petrology, 859–880. 48: 835–838. McLean, R. F. 1974. Geological significance Linke, O. 1954. Erhabene Fährten im of bioerosion of beachrock. Proceedings Sand. Natur und Museum, 84: 18–21. of the 2nd International Coral Reef Lockley, M. 1991. Tracking Dinosaurs - a Symposium, 2: 401-409. New Look at an Ancient World. Mietto, P., Avanzini, M. and Rolandi, G. Cambridge University Press, 2003. Human footprints in Pleistocene Cambridge, 238 p. volcanic ash. Nature, 422: 133. MacEachern, J. A. and Pemberton, S. G., Milàn, J. and Bromley, R. G. 2006. True 1992. Ichnological aspects of Cretaceous tracks, undertracks and eroded tracks, shoreface succession and shoreface experimental work with tetrapod tracks variability in the Western Interior seaway in laboratory and field. of North America. In Pemberton, S. G. Palaeogeography, Palaeoclimatology, (ed.), Application of Ichnology to Palaeoecology, 231: 253-264. Petroleum Exploration. A Core Mountain, E. D. 1966. Footprints in Workshop. Society of Economic calcareous sandstone at Nahoon Point. Paleontologists and Mineralogists, 17: South African Journal of Science, 62: 57–84. 103–111. Male, W. H. 1992. The sedimentology and Mutti, E., Orombelli, G. and Pozzi, R. ichnology of the Lower Cretaceous 1970. Geological studies on the (Albian) Bluesky Formation in the Karr Dodecanese Islands (Aegean Sea). IX. Area of west-central Alberta. In Geological map of the Island of Rhodes Pemberton, S. G. (ed.), Application of (Greece). Explanatory notes. Annales Ichnology to Petroleum Exploration. A Géologiques des Pays Helléniques, Core Workshop. Society of Economic Series 1, 22: 77–226. Paleontologists and Mineralogists, 17: Nara, M., 1994. What is the producer of 33–55. “trace fossil of Excirolana chiltoni”? - Maples, C. G. and Suttner, L. J. 1990. Trace Tracemaking mechanism of fossils and marine-nonmarine cyclicity in Macaronichnus segregatis. Fossils, 56: the Fountain Formation (Pennsylvanian: 9–20. [In Japanese, with English Morrowan/Atokan) near Manitou Spring, abstract].

26 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

Nicolaisen, W. and Kanneworff, E. 1969. Geological Journal, Special Issues, 9: On the burrowing and feeding habits of 227–264. the amphipods Bathyporeia pilosa Radwański, A., Friis, H. and Larsen, G. Lindström and Bathyporeia sarski 1975. The Miocene Hagenør-Børup Watkin. Ophelia, 6: 231–250. sequence at Lillebælt (Denmark): its Pemberton, G. S., Spila, M., Pulham, A. J., biogenic structures and depositional Saunders, T., MacEachern, J. A., environment. Bulletin of the Geological Robbins, D. and Sinclair, I. K. 2001. Society of Denmark, 24: 229–260. Ichnology and Sedimentology of Reineck, H.-E. and Flemming, B. W. 1997. shallow to marginal marine systems: Unusual tracks, traces, and other oddities. Ben Nevis & Avalon Reservoirs, Jeanne Courier Forschungs-Institut D’Arc Basin. Geological Association of Senckenberg, 201:349–360. Canada, Short Course Notes, 15: 343 p. Rice, M. E. 1971. Survey of the Sipuncula of Politis, G. G. and Bayón, C. 1995. Early the coral and beachrock communities of Holocene human footprints and sea the Caribbean Sea. Proceedings of the mammals in the tidal zone of the International Symposium on the Biology Argentinian foreshore. Past, 20: 5–6. of the Sipuncula and Echiura 1, Kotor, Pollard, J. E., Goldring R. and Buck, S. G. June 1970, 35–49. 1993. Ichnofabrics containing Roberts, G. 2004. Ephemeral, subfossil Ophiomorpha: significance in shallow- mammalian and avian tracks within water facies interpretation. Journal of the intertidal-zone, Flandrian sediment Geological Society of London, 150: 149– exposures at Formby Point, Sefton 164. Coast, north west England . In Kim, Radwański, A. 1969. Lower Tortonian J.Y., Kim, K.-S., Park, S.I. and Shin, transgression into the southern slopes of M.-K. (eds.), Proceedings of the the Holy Cross Mts. Acta Geologica International Symposium on the Polonica, 19: 1–164. [In Polish with Quaternary Footprints of Hominids and English summary]. Other Vertebrates. Namjejugun, p. 54– Radwański, A. 1977. Present-day types of 83. traces in the Neogene sequence; their Roberts, G., Gonzalez, S. and Huddart, D. problems of nomenclature and 1996. Intertidal Holocene footprints and preservation. In Crimes, T. P. and their archaeological significance. Harper, J. C. (eds), Trace fossils 2. Antiquity, 70: 647–651.

27 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

Sarjeant, W. A. S. and Langston, W. 1994. Saunders, T. D., MacEachern, J. A. and Vertebrate footprints and invertebrate Pemberton, S. G. 1994. Cadotte traces from the Chadronian [?] (Late Member Sandstone: progradation in a Eocene) of Trans-Pecos, Texas. Texas boreal basin prone to winter storms. In Memorial Museum Bulletin, 36: 86 p. Pemberton, G. S., James, D. P. and Saunders, T.D.A. 1989. Trace fossils and Wightman, D. M. (eds.), Canadian sedimentology of a Late Cretaceous Society of Petroleum Geologists progradational barrier island sequence: Manville Core Conference, p. 331–349. Beapraw-Horseshoe Canyon Formation Tuttle, R. H. 2004. Footprint clues in transition, Dorothy, Alberta. University forensics and hominid evolution: of Alberta, M.Sc. thesis, 187 pp. limitations and lesson. In: Kim, J. Y., Saunders, T. D. A. and Pemberton. S. G. Kim, K. -S., Park, S. I. and Shin, M. -K. 1986. Trace fossils and sedimentology of (eds.), Proceedings of International the Appaloosa Sandstone: Bearpaw- Symposium on the Quaternary Horseshoe Canyon Formation transition, Footprints of Hominids and Other Dorothy, Alberta. Canadian Society of Vertebrates. Namjejugun, p. 44–53. Petroleum Geologists, Field Trip Guide Tuttle, R., Webb, D., Weidl, E. and Baksh, Book, 117 p. M. 1990. Further progress on the Saunders, T. D. A. and Pemberton. S. G., Laetoli trails. Journal of Archaeological 1990a. On the palaeoecological Science, 17:347-362. significance of the trace fossil Uchman, A. and Bromley, R. 2003. Macaronichnus. 13th International Beachrock ichnofabrics of Roman age Sedimentological Congress, Nottingham, from Rhodes, Greece. In: VII England, 26-31st August, 1990. Abstract International Ichnofabric Workshop, of Papers, p. 416. Program, Abstracts, Contributors, Saunders, T. D. and Pemberton, S. G. Basel, Switzerland, 14th to 16th of July 1990b. Trace Fossils and 2003, p. 55. Basel. Sedimentology of the Appaloosa Vialov, O. S. 1965. Stratigraphia Sandstone: Bearpaw-Horseshoe Canyon Neogenovih Molass Predkarpatskoho Formations Transition, Dorothy, Progiba. Naukova Dumka, Kiev, p. 192. Alberta. Edmonton Geological Society [In Russian.] Field Trip Guidebook. Vialov, O. S. 1966. Sledy zhiznedelatelnosti organismov i ich

28 Bromley et al. – Macaronichnus and human and cattle trackways in Holocene beachrocks ______

paleontologicheskoye znachene. Kiev, 220 p. [In Russian.] Walker, R. G. and Bergman, K. M. 1993. Shannon Sandstone in Wyoming: a shelf-ridge complex reinterpreted as lowstand shoreface deposits. Journal of Sedimentary Petrology, 63: 839–851. White, T. D. 1980. Evolutionary implications of Pliocene hominid footprints. Science, 208: 175–176.

29 Jesper Milàn – Vertebrate Ichnology Appendix ______

Paper 12 Do shod humans leave true tracks? Jesper Milàn & Richard G. Bromley Accepted for publication in Ichnos

Milàn & Bromley – do shod humans leave true tracks? ______

SHORT COMMUNICATION

Do shod humans leave true tracks?

Jesper Milàn & Richard G. Bromley Geological Institute, University of Copenhagen, Oester Voldgade 10, DK-1350 Copenhagen K, Denmark. E-mail: [email protected] Phone: 0045 353224236

ABSTRACT Incorporating footprints from shod humans into ichnotaxonomical nomenclature presents several problems in that the track maker does not actually touch the sediment, and further the wear of shoes represents a behavioral choice as well as evidence of technological developments in shoemaking. If footprints of shod humans were to be treated ichnotaxonomically, they should be regarded as traces of compound behavior comprising bipedal walking, wearing and production of shoes. Footprints of naked feet offer no restrictions in being classified into the ichnotaxonomical system. ______

INTRODUCTION broadest sense, not only footprints in The current interest in human ichnology sediments, but also traces of their various initiated by the international symposium on activities like hunting and preparing of food the Quaternary footprints of hominids and (Noe-Nygaard, 1989), settlements, cave other vertebrates (Kim et al., 2004), will paintings, sculptures and various tool hopefully result in increased utilization, communities (Kim et al., this volume). documentation, and awareness of human Footprints from more recent track makers, in footprints in both paleontological and addition to providing information about the archeological contexts. One important factor stance, gait and foot morphology of the track that should be kept in mind when dealing with maker, also reflect behavior in the form of human footprints is that they do not entirely wearing footwear as well as technical follow the criteria of other trace fossils. developments in shoemaking. This paper will Traces of modern humans include, in the

1 Milàn & Bromley – do shod humans leave true tracks? ______only deal with the question of footprints from preservation of fine anatomical details. To unshod versus shod humans. consider footprints of shod humans as undertracks would make sense, as no DISCUSSION anatomical details of the track maker’s foot, The occurrence of footprints of humans except for the general outline, are preserved, wearing shoes in lithified Roman-age as experimental work with track and beachrock from the Greek island of Rhodes undertracks predicts (Milán and Bromley, in (Bromley et al., this volume) raises the press). For example, the tracks of a giant interesting ichnophilosophical question: ground sloth from the Pleistocene Carson City should a footprint of a shod human be tracksite, Nevada, in the late 1880s were considered a true track or an undertrack, and misidentified as tracks from a giant human should footprints from humans wearing shoes wearing sandals (Lockley and Hunt, 1995 and be the subject of ichnotaxonomical refs. therein). nomenclature? If, on the other hand, shoes are considered Undertracks are defined as the as a part of the track maker, then another deformation in the sedimentary layers problem arises. The morphology of the tracks subjacent to the track maker’s foot (Lockley, of shod humans will vary fundamentally 1997). Gatesy (2003) took the discussion according to which footwear is in use, as further, stating that, strictly speaking, only the human footwear comes in a wide range of actual sediment grains touched by the track morphologies, ranging from plain shoes to maker’s foot should be regarded as the true high heels. Further, the choice of footwear in track. It is here that the problem with shoes humans reflects both a behavioral choice and arises, as a shod human actually does not a behavioral adaptation to local conditions. As touch the sediment, but is separated from the a consequence of this, the same track maker tracking surface by the sole of the shoe, which can produce very different tracks of very in this case acts as an impermeable different appearance, from different shapes of “sediment” layer on top of the actual tracking heels and toes to the tracks of snorkeling fins surface. or the long sinusoidal traces of skis (Fig. 1). Nadon (2001) suggested that the apparent To apply ichnotaxonomic nomenclature to the lack of fine anatomical detail in dinosaur huge variety of structures that human tracks is due to a layer of mud that adheres to footwear is capable of producing, would be a the foot during walking and thereby hinders meaningless occupation.

2 Milàn & Bromley – do shod humans leave true tracks? ______

Figure 1. The many faces of footprints of shod humans. (A) The footprint from a stiletto heeled shoe is able to leave puncture marks from the heel in even hard substrates. (B) Parallel meandering traces in snow produced by humans wearing skis. (C) Apparent didactyl human footprint produced by (D) RGB in Japanese free-toed boots. (E) Footprint in beach sand from a human wearing fins. If found fossilized, this could be described as the footprints from an aquatically adapted species of human. All photos by RGB

3 Milàn & Bromley – do shod humans leave true tracks? ______

For such a reason, it has been suggested of Natural Science, University of that structures created by human civilization Copenhagen. should be excluded from ichnonomenclatural treatment (Bertling et al., in press). REFERENCES Nevertheless, fossil or subfossil tracks produced by shod humans, even from Bertling, M., Braddy, S., Bromley, R. G., historical times, should by all means be Demathieu, G. D., Mikuláš, R., Nielsen, documented, described and published. As an J. K., Nielsen, K. S. S., Rindsberg, A. K., example an imprint of a modern hobnail boot Schlirf, M. and Uchman, A., in press. was identified in association with tracks of Names for trace fossils: a uniform unshod Paleolithic track makers in the floor of approach. Lethaia. the French cave of Niaux (Pales, 1976). Bromley, R. G., Uchman, A. Milàn, J. and However, tracks made by shod humans Nielsen, K. S. (this volume). Rheotactic should not be considered a basis for Macaronichnus, and human and cattle ichnotaxonomic nomenclature. If tracks from trackways in Holocene beachrock, humans wearing shoes were to be treated Greece: reconstruction of palaeoshoreline ichnotaxonomically, they should then be orientation. Ichnos. regarded as traces of a compound behavior Gatesy, S. M. 2003. Direct and indirect track comprising bipedal walking, wearing and features: what sediment did a dinosaur production of shoes. Given the above touch? Ichnos 10: 91–98. arguments however, imprints of naked feet Kim, J. Y., Kim, K.-S., Park, S. I. and Shin, offer no such restrictions to being classified M.-K. 2004. Proceedings of the into the ichnotaxonomical system. international symposium on the Quaternary footprints of hominids and ACKNOWLEDGEMENTS other vertebrates, Jeju Island, Korea. 175 We are grateful to Martin G. Lockley and pp. Joeng Yul Kim for promoting the study of Kim, J. Y., Lockley, M. G. and Kim, K. S. human ichnology by editing this volume, and (this volume). Hominid Ichnology: an for their constructive reviews and comments introduction. Ichnos. on the manuscript. The research of JM is Lockley, M. G. 1997. The paleoecological supported by a Ph.D. grant from the Faculty and paleoenvironmental utility of dinosaur tracks. In Farlow, J. O. and

4 Milàn & Bromley – do shod humans leave true tracks? ______

Brett-Surman, M. K. (eds.), The Nadon, G. C. 2001. The impact of Complete Dinosaur. Indiana University sedimentology on vertebrate track Press, Bloomington, pp. 554–578. studies. In Tanke, D. H. and Carpenter, Lockley, M. G. and Hunt, A. P. 1995. K. (eds.) Mesozoic Vertebrate Life. Dinosaur tracks and other fossil Indiana University Press, Bloomington, footprints of the western United States. 395–407. Columbia University Press, New York, Noe-Nygaard, N. 1989. Man-made trace 338 p. fossils on bones. Human Evolution 4: Milàn, J. and Bromley, R. G. in press. True 461–491. tracks, undertracks and eroded tracks, Pales, L. 1976. Les impreintes de pieds experimental work with tetrapod tracks in humains dans les caverns. Archéologie laboratory and field. Palaeogeography, Institut Paléontologie Humaine, 36: 1– Palaeoclimatology, Palaeoecology. 166.