INTRODUCTION

Footprints record the dynamic interaction between an animal’s limb and a malleable substrate (Baird 1957). As the foot penetrates and is extracted, nearby sediment is pushed, sheared, and dragged into a new configuration that variably records aspects of pedal anatomy and locomotor movement (Padian and Olsen 1984; Thulborn and Wade 1984, 1989; Gatesy et al. 1999; Gatesy 2003). This imperfect mold is further altered, if not completely destroyed, by pre-burial erosion, diagenesis, post-exposure weathering, collection, and preparation. A fossil track that survives these harsh filters can offer unique and valuable evidence of behavior in extinct taxa (e.g., Seilacher 1967). For ichnological analyses to be well founded, however, footprints must be documented by methods that minimize loss of information. Inaccurate representations of track morphology can distort or obscure potentially important data.

Compared to many other fossils, description of footprints can be particularly challenging. The most important distinction is that the internal morphology of bones, eggshells, and soft tissues are preserved when actual biological material serves as a template for mineral replacement. By contrast, tracks are purely sedimentary structures that only reflect a foot’s external morphology and frequently lack discrete borders. Second, images of track surfaces are informative from a limited range of perspectives, whereas skeletal elements can often be figured from many viewpoints. Finally, many footprints are studied only in the field, rather than collected or cast. Even under controlled laboratory conditions, a track’s three-dimensional contours and textures are notoriously difficult to quantify and illustrate, making these phases of analysis especially prone to inaccuracy or bias.

The human visual system is adept at determining the distance and orientation of surfaces. However, just as a camera compresses a 3-D world onto a planar CCD or film, depth information is lost when the environment is projected onto the 2-D receptor array in our retina (e.g., Palmer 1999). Our brain must therefore use a number of different signals to perceive spatial arrangement and resolve ambiguity. Some sources of information are intrinsically dynamic, such as the differential motion of objects at unequal distances (motion parallax). However, most other depth cues are static and potentially useful for extracting 3-D information from 2-D images on a monitor or in print (Palmer 1999; Ware 2004). These cues include occlusion (near overlaps far), perspective (convergence of parallel lines, position relative to the horizon, relative size, atmosphere), focus (depth of field), shading, cast shadows, and stereopsis. Unfortunately, tracks rarely show features with enough topography to benefit from occlusion, perspective, or depth of field. This leaves shading, cast shadows, and stereopsis as signals available for elucidating and communicating track geometry.

Stereopsis is the extraction of depth information from differences between images recorded by our two retinas (binocular disparity; e.g., Palmer 1999). Paleontologists have used stereophotographic techniques for more than 90 years (e.g., Hudson 1913, 1925) and the methodology is well described (Gott 1945; Evitt 1949; Feldman 1989; Knappertsbusch 2002). Yet, although several stereo pairs of tetrapod track photographs have been published (Sarjeant and Thulborn 1986; Ishigaki and Fujisaki 1989; McAllister 1989), this approach has been underutilized relative to the widespread use of stereophotography to illustrate skeletal fossils. There are drawbacks that make traditional stereo pairs less than ideal for publication and presentation (e.g., Evitt 1949). Printed figures are constrained to relatively small widths (less than ca. 8 cm), which preclude highly detailed images spanning a full page. At the same time, separate left and right images require at least twice the area of an individual plate. Traditional stereo pairs can only be projected before an audience using specialized polarizing or LCD equipment.

Anaglyph stereo imaging offers an alternative method of presentation that is scale-independent. An anaglyph is a color image formed by superimposing left and right members of a stereo pair. The two original images are color-converted so that each is invisible when viewed through a correspondingly colored gel. Inexpensive and widely available “3-D glasses” with different lenses (red-blue, red-green, or red-cyan) are worn to provide each eye with its appropriate image. Instructions for creating anaglyphs using imaging software such as Adobe Photoshop are available in Purnell (2003) and on many websites. Anaglyphs can be printed and projected at any size, making them ideal for journals, websites, museum displays, poster sessions, small seminars, and large conference halls. Sequential anaglyphs can be easily combined to create compelling stereo animations. Paleontologists have recently taken advantage of this technique by publishing static and animated anaglyphs of conodont and invertebrate microfossil material (Knappertsbusch 2002; Purnell 2003).

Herein, we address the utility of anaglyph stereo imaging to the exploration and exposition of dinosaur tracks. Our examples are tracks attributable to small theropods that are preserved in the Late Triassic Fleming Fjord Formation of Jameson Land, East Greenland (Jenkins et al. 1994; Gatesy et al. 1999; Gatesy 2001, 2003). We present three case studies ranging from field to laboratory, from whole footprints to minute skin impressions, and across a range of imaging techniques. We include specific methods as part of each case study. Our goal is to focus on the benefits of anaglyphs for footprint studies in general, rather than on specific descriptions or interpretations of these specimens.