INTRODUCTION

There are two primary categories of fossil deformation, brittle and plastic. Brittle deformation may be described as structural cracking without shape change of the individual broken pieces, whereas plastic deformation is described as shape change without breakage. Plastic deformation alters the true shape of a fossil, the shape of the body part during life.

Fossil deformation is the result of many different forces and processes. Overburden stress and resulting porosity loss are the primary processes that cause fossil deformation. As fossils are initially buried the weight of the overlying sediments linearly compacts the fossil from above. Depending on numerous associated factors, including confining pressure and strain rate, this can cause the fossil to break and/or warp. All fossils undergo these forms of diagenetic deformation, implying that the true fossil shape is rarely, if ever, preserved.

Secondary causes of fossil deformation include tectonic stresses and sediment cracking. Tectonic stresses from large- and small-scale plate motion can be a major source of both brittle and ductile deformation. Tectonic deformation is usually nonlinear and very complicated (Davis and Reynolds 1996). Because of this it is very difficult to predict and correct for tectonic deformation. Likewise sediment cracking/expansion is a poorly understood process that involves the invasion and expansion of sedimentary matrix material into weak areas of a fossil, such as sutures (White 2003). This process can greatly alter the shape of a fossil.

It is important to consider the effects of deformation in almost all paleontological studies. Many false conclusions may be drawn if a plastically deformed fossil is thought to retain its true shape. Functional morphology relies on fossil shape to hypothesize motion capabilities and lifestyle of extinct organisms. Systematic and phylogenetic studies use fossil shape when recognizing a taxon and coding its characters in the process of tree building. Using plastically deformed fossils for either type of study carries great inherent risk and adds enormous amounts of error to all results. Because of this, numerous methods have been created for removing both brittle and plastic deformation. Such methods have been called retrodeformation (Hughes and Jell 1992).

Plastic deformation is difficult to remove and retrodeformation can only be attempted digitally (Zollikofer and de León 2005, Zollikofer et al. 2005) or optically (Lake 1943, Hills and Thomas 1944). Almost all plastic retrodeformation methods require that the original fossil shape was symmetrical. These methods use the re-acquisition of symmetry as the basis for the transformations applied to the fossil. The geometric average method involves identifying symmetrical landmark pairs and equalizing their distance to an established sagittal plane (Ogihahra et al. 2006). The minimum stretch method calculates the minimum amount of stretch necessary that a fossil would have to undergo to re-attain its symmetry and deforms the fossil according to this calculation (Zollikofer and de León 2005). It is not our purpose to examine the methods for plastic retrodeformation.

Brittle retrodeformation is generally accomplished using a method we refer to as "puzzle piecing." The method consists of taking the broken pieces of a fossil and piecing them back together in their presumed original shape, much like constructing a model airplane. This can be done both physically, with the actual fossil fragments or fragments of a cast, or digitally, using digitized representations (Zollikofer et al. 1995, 2005). We believed that this method would yield unreliable and unrepeatable results, which could lead to inaccurate studies of the fossils in question. Therefore, we tested this method of reconstruction in a completely controlled environment by piecing together digitally deformed pieces of an extant specimen using both the original specimen and an alteration of the original specimen as reconstruction templates. We wanted to know if 1) puzzle piecing only results in the true skull shape; and 2) the fact that deformed skull pieces could be arranged into a symmetrical 'skull' justifies the assumption that plastic deformation from overburden compaction does not cause additional significant losses of shape data.