Project History
It may be some time before this website attains the level of balance and diversity to which we aspire. However, whatever semblance of balance it has now is a result of drawing together multiple strands from the research and educational efforts of a group of colleagues. Each has contributed to the whole, and like any integrated, organic form, it is difficult to trace all the contributions. We anticipate, moreover, that it will continue to evolve, so its full history cannot yet be written. For now, we offer an account of just one thread in the tapestry.
Mastodons and Mammoths
Non-digital methods
The proboscidean component of this enterprise, featuring mainly mastodons and mammoths, can be traced back at least to 1989 and the discovery of the Burning Tree mastodon, in Licking County, Ohio (Lepper et al., 1991; Fisher et al., 1991, 1994). This specimen was remarkably complete and well preserved, offering an opportunity to deduce bone identifications not only from shape, but also through direct inspection and matching of features on articular surfaces throughout the body. For the better part of a year, Professor Dan Fisher made frequent weekend trips to central Ohio, working with a local photographer to document this specimen with six orthogonal views of each bone. Both conventional B&W photography and stereophotography were used to capture as much 3D information as possible with the available technology. We were thankful for every hour we spent with this specimen, but unfortunately, it was sold to a commercial dealer and shipped overseas before our work was done.
After losing access to the Burning Tree mastodon, Prof. Fisher was determined to do everything possible to secure more information in less time, should any similar opportunity arise again. Years passed, but not until 1998 and the discovery of the Buesching mastodon was anything comparable found. This was not a complete specimen, but it was promising from a scientific perspective, and in early 1999, through an agreement with the Buesching family and Indiana University/Purdue University at Fort Wayne (IPFW), it was loaned to the University of Michigan so that work documenting its morphology could begin. The ultimate disposition of this specimen was still undecided, so our first priority was again photography, because it could be completed most quickly, followed by molding of each bone recovered from the site. These molds would allow accurate physical replicas (casts) of the bones to be produced later.
This process had barely begun when, later in 1999, the Hyde Park mastodon, was discovered. Excavated in 2000, this skeleton was almost as complete as the Burning Tree mastodon, and so another cooperative agreement was initiated, this time with the Paleontological Research Institution (PRI), in Ithaca, New York. PRI was on a fast-track to mount the Hyde Park mastodon as a focal specimen in their new Museum of the Earth. We therefore started Hyde Park photography immediately after Buesching photography and shifted to Hyde Park molding even before Buesching molding was done. Additional information on the Hyde Park mastodon is available in Fisher (2008) and other chapters in Allmon and Nester (2008).
Following recovery of the Buesching and Hyde Park mastodons, we committed to comprehensive molding because we knew that bones of the Buesching mastodon would most likely end up in an Indiana institution, where a state tax credit would favor their donation, and those of the Hyde Park mastodon would become effectively inaccessible once they were mounted. Molds, and the casts they allow us to produce, were, and in some cases still are, the “gold standard” for capturing information on form, but they are time-consuming to create, and they are not as accessible remotely as photographs or digital models have the potential to be.
References:
Allmon, W. D., and P. L. Nester (Eds.), 2008. Mastodon paleobiology, taphonomy, and paleoenvironment in the late Pleistocene of New York State: studies on the Hyde Park, Chemung, and North Java sites. Palaeontographica Americana 61.
Fisher, D.C. 2008. Taphonomy and paleobiology of the Hyde Park mastodon. In Mastodon paleobiology, taphonomy, and paleoenvironment in the late Pleistocene of New York State: studies on the Hyde Park, Chemung, and North Java sites. W. D. Allmon and P. L. Nester, eds., Palaeontographica Americana 61: 197-290.
Fisher, D.C., B.T. Lepper, and P.E. Hooge. 1991. Taphonomic analysis of the Burning Tree mastodont. Current Research in the Pleistocene 8: 88-92.
Fisher, D.C., B.T. Lepper, and P.E. Hooge. 1994. Evidence for butchery of the Burning Tree mastodon. In: The First Discovery of America: Archaeological Evidence of the Early Inhabitants of the Ohio Area (W.S. Dancy, editor), Ohio Archaeological Council, Columbus, Ohio, pp. 43-57.
Lepper, B.T., T.A. Frolking, D.C. Fisher, G. Goldstein, D.A. Wymer, J.E. Sanger, J.G. Ogden, III, and P.E. Hooge. 1991. Intestinal contents of a late Pleistocene mastodont from midcontinental North America. Quaternary Research 36: 120-125.
Digitization
After access to the Burning Tree mastodon was lost, it became clear that some more effective way to archive and share information on bone shapes was needed. Photographs were a good starting point, with a favorable balance of information conveyed relative to time expended in taking them, but two dimensions are ultimately no match for three. What we needed was some way to express the shapes of bones in digital terms. In principle, recording 3D coordinates for points on a bone surface is not complicated, and yet in the early days of personal computing, commercially available implementations of this capability were few. Convinced nonetheless that this was the way forward, early in 1991, Prof. Fisher purchased an Isotrack electromagnetic digitizer and Hyperspace software with which point coordinates could be acquired. This early software required explicit construction of each polygon added to a mesh to define a surface, but surfaces did ultimately emerge, leading to our first digital bone models. However, we learned that even after removing all metallic objects from the vicinity of the digitizer, nonlinear aspects of sensor performance resulted in distortions of the shapes that were recorded, and because of these inaccuracies, no Isotrack models remain in our current repository.
By 1998, our ailing Isotrack digitizer was replaced by a Microscribe-3DLX, an articulated-arm contact-digitizer with significant improvements in accuracy (without the systematic distortion of the electromagnetic system) and precision. For software, we graduated to Geomagic Wrap, which allowed us to acquire points in a free-form sequence, followed by computation of a surface “wrapped” over those points.
Extended access to the Buesching material allowed us to digitize the original bones before molding, while fast-tracking of the Hyde Park mastodon meant that digitizing had to be done on casts produced following the completion of molds. The majority of bones from the Buesching and Hyde Park mastodons were digitized using the Microscribe-3DLX. This digitizer allows construction of “surface models”, but often there is important information in the color or fine-scale texture of the bone surface that is missing in such models. So, in the early 2000s, we began to work with DeepPaint 3D software, which allowed us to do a primitive form of texture-mapping, “painting” the surfaces of bone models with custom textures that schematically represented synovial surfaces, rugose muscle attachment areas, and other conditions typical of bone exteriors. These schematic representations were abandoned in 2010, when we began to use 3DSOM software, which permits production of photorealistic models by projecting actual photos of the specimens onto surface models. All texture-mapped mammoth and mastodon specimens currently on the site use photograph-derived textures.
In 2006, new standards in digitizing were being set, and we purchased a Creaform HandyScan 3D laser scanning digitizer. This offered some improvement in precision but mainly a great leap forward in the density of points that could be acquired and the speed of surface model production. With the associated VXelements software, model creation is essentially automated, although we often need to stitch together a series of scans, which we do with Magics software from Materialise. Some proboscidean specimens on this site (e.g., one tusk of the Buesching mastodon) are reduced-resolution laser-scanned models. The size of the full-resolution scans generally precludes online use.
Continuing work
The models served on this site include both surface models and texture-mapped models. Where possible, especially if it would assist with some analysis, we plan to add texture-mapping later. Many of our newer models are generated using the HandyScan, but increasingly, we also incorporate models generated using photogrammetry and X-ray computed tomography – industrial CT, medical CT, and microCT. Standards will no doubt continue to change, and we will strive to present the best available models to our users.