Research using engineering tools is making new options possible in the study and preservation of ancient finds. According to the manager of Dinosaur Hall at the Smithsonian Institution, the plan is to keep the original fossil bones off the floor, in a collection available for study by scholars, and to replace the skeleton with a replica created by methods that will include 3D modeling and rapid prototyping. Computer animations will help recreate the way the animal stood. The optical scanner analyzes an object in a series of small areas, or patches, which must be assembled to create the overall picture. In order to set up the later alignment of the scans, an engineer mapped a digital outline of the Triceratops. Restorers are expected to use the conserved Triceratops bones as patterns for molds, but when parts are missing or mismatched, mirror images will fill in. Paleobiologists will also use the scale model, along with Virtual Surfaces' animations, to study possible reconstructions of the dinosaur's stance.
Late 20th century engineering technology is adding new terms to the vocabulary of the Dinosaur Hall at the Smithsonian Institution's National Museum of Natural History. "Digital scanning" and "stereolithography" are gaining currency alongside all those words that begin with "paleo," which means "very old." At the same time, research using engineering tools is making new options possible in the study and preservation of ancient finds.
The Sl11ithsonian is one of the country's most venerable establishments, where change is not frivolous. The public perception of the Dinosaur Hall is more likely associated with the patient adjustment of calipers than with automated calculations at the speed of light. But behind the scenes modern techniques have been catching on, and now the paleobiology department has enlisted a group of high-tech companies in an effort to save a deteriorating piece of the national heritage.
The presence of electronic engineering tools isn't new to the study of fossils at the National Museum of Natural History. Ralph Chapman heads the museum's morphometrics laboratory, which studies the shapes of fossils and also has applied its measurement skills to other tasks, such as criminal investigations. According to Chapman, fossils and the remains of crime victims have been studied and identified on his watch. Tools at the lab include a point digitizer from Polhemus Inc. of Colchester, Vt. Chapman also uses AutoCad and 3-D Studio from Autodesk Inc. of San Rafael, Calif., and writes some of his own code.
What makes the latest job a little different is that this time the goal is to build a whole animal in a process that will include polygonized point clouds and .stl files.
"The electronic thing is my bandwagon," Chapman said. "I've been wanting to do this for a dozen years."
It all began last year when Kathy Hawks, a conservator retained by the Smithsonian to look over its fossils, discovered that the museum's Triceratops skeleton was weakening and in danger of collapse. The museum had had some doubts about the creature's physical condition because it is at least 65 million years old and has been on its feet at the Smithsonian for close to a century. The museum believes it was the first Triceratops ever mounted. Completed in 1905 by Norman Boss, a Smithsonian preparator, and Charles Gilmore, a dinosaur curator, it originally stood in the Smithsonian's Arts & Industries Building.
One plan of preservation was to dismantle the skeleton, glue and harden the bones, and remount them in a padded armature. But, eventually, Richard Benson, director of the museum's Department of Paleo biology, became interested in looking at alternatives.
Now the plan is to keep the original fossil bones off the floor, in a collection available for study by scholars, and to replace the skeleton with a replica created by methods that will include 3-D modeling and rapid prototyping. Computer animations will help recreate the way the animal stood.
This latest-meets-oldest motif came together, appropriately enough, at a paleobiology meeting last fall in Utah, a place with a considerable capacity for yielding dinosaur remains. Among its attractions for paleozoologists, the state contains part of the fossil trove designated Dinosaur National Monument.
Benson, Chapman, and Steve Jabo, a preparator in the Department of Paleo biology at the Smithsonian, went to Utah to meet two technology advocates who would be key players in the preservation effort.
One of them was Arthur Andersen, president of Virtual Surfaces Inc. of Mount Prospect, Ill. Andersen describes himself as a physicist and mathematician turned engineer, and his company's chief line of business is providing engineering services to commercial clients. But Andersen also calls himself a "paleotechnologist;' that is, not quite a paleontologist, but near enough to belong to the club. What's more, he said his firm has built up a sideline in the field.
He "had done dinosaur work in the past at the University of Chicago," he said. Working with Hans Larsson, a graduate student in paleontology at the university, Andersen had taken CAT scans of dinosaur braincases. The scans were converted to IGES format using Mimics software by Materialise U.S.A. of Ann Arbor, Mich., and edited further with Surfacer, software from the Image-ware division of Structural Dynamics Research Corp., also of Ann Arbor. Objects representing the interiors of the braincases were reproduced by Baxter Healthcare Corp.'s Advanced Engineering Design Center in Round Lake, Ill. Baxter used LOM (for laminated object manufacturing) rapid prototyping equipment from Helisys Inc. of Torrance, Calif., to create the objects. Andersen had come to the conference to deliver a talk about the work he had done with Larsson.
Also in Utah for the meeting was Lisa Federici, president of Scansite 3-D Services of San Francisco. Her company also had experience in applying the digital. scanning technology to fossils.
That meeting led to the recruitment of Virtual Surfaces and Scansite into the museum's preservation project.
They, in turn , brought in Dimension Data, a Farmington Hills, Mich., company that markets digital scal1lung equipment and offers contract scanning services. The company put Henry We de, an engineer, in charge of its part of the scanning operations. As Wede recalled it, "We had a person who does computer work in Chicago ask us, 'Can you do dinosaur work?'"
From Doll Heads to Windshields
According to Wede, "We've scanned everything from doll heads to the windshields of buses." Since the company already had been hired for projects as diverse as mold-making for figurines and prototyping automotive components, Triceratops bones didn't seem that far out of its line. So Wede set up shop in the Dinosaur Hall and made about 40 scans of the dinosaur, which included more than 100 bones. Some scans were of individual bones and others were of parts with several bones. For example, Wede scanned the spine for a record of the structure.
The spine was an area where Boss and Gilmore had to fill in with probable structures, derived from only three authentic bones. We de scanned these three vertebrae separately, and they will serve as models from which the rest of the spinal bones will be generated.
Wede used a Comet white-light scanner made by Steinbichler GmbH of Neubeuern, Germany. The company operates the scanner with proprietary software that runs on a generic PC, which needs about 256 ME of random-access memory to handle the resulting data files.
The optical scanner analyzes an object in a series of small areas, or patches, which have to be assembled to create the overall picture. To set up the later alignment of the scans, Wede mapped a digital outline of the Triceratops.
He took round stickers, like the circular labels that are sold at any stationery store, and affixed them to several places on the skeleton. When he touched a hand probe to the center of each sticker, We de got a pattern of reference points that would help in reassembling the creature from its many parts.
Wede spent three weeks scanning. Besides the large bones of the skeleton, he also scanned an extra Triceratops skull and the skull of a related type of dinosaur called Centrosaurus.
An animal as old as a Triceratops doesn't come in one piece. Plenty of parts have been lost since the Cretaceous period, when these herbivorous dinosaurs ran in herds. The skeleton at the National Museum of Natural History is the amalgam of bones from as many as 15 individuals. Some parts are the result of conjecture, based on knowledge available in 1905. Certain parts of the mounted structure contain more original bones than others.
Wede scanned the left side of the skeleton because it contained more original ribs than the right. A mirror image of that scan will generate the right ribs. The process was reversed for the legs, where more authentic parts are found on the right.
Restorers will use the conserved Triceratops bones as patterns for molds, but when parts are missing or mismatched, mirror images will fill in.
Jabo explained that the left upper arm bone, for example, is smaller than the right because the bones came from different Triceratops. The mold for the right upper will be made from the conserved fossil bone itself. The cast for the left upper will begin as a digital mirror image of the bone on the right; then a replica bone will be milled to make a pattern for a mold.
A scanner describes a surface as a dense cloud of points, each fixed by its x, y, and z coordinates. The result is a file too big for most computers to work with. Someone along the way has to simplify the image by reducing the number of points so a computer can handle the information efficiently. The density of the Triceratops scan described a point every 0.2 mm, Wede said.
Andersen's company, Virtual Surfaces, has the job of simplifying the images. According to Andersen, the 0.2 mm density of Dimension Data's scan comes out to about 14 million points for the Triceratops skull alone. Virtual Surfaces scales back the density to a point every 1 or 2 mm, depending on the size of the specimen.
According to Andersen, the reduced point cloud will retain the accuracy of the original, and will be more manageable for graphic and rapid prototyping uses.
Many of the bones contain steel support structures called armatures. Surfacer point cloud editing software will digitally remove the armatures and close the resulting gaps.
Andersen said his "animation guru," Kyle Kenny, has worked out a simulation that shows the skeleton in motion. One aim of animating the 3-D computer model is to help in "reconstructing the posture of the beast," Andersen said.
Jabo, one of the Smithsonian's three preparators (so called because they prepare bones and assemble skeletons for a living), said the original Triceratops skeleton will be preserved and put in the museum's collection for study.
The public will see a cast based on the original, perhaps with real skulls. The replica will have a modular design that will be practical for dismantling and reassembling, so it's conceivable that the new Triceratops may travel from time to time. According to a brochure published by the museum, the last time the Smithsonian shipped out a fossil vertebrate skeleton for a temporary exhibit was in the 1930s.
THE DINOSAUR HALL at the Smithsonian Institution is looking for a few interested mechanical engineers. There's no cash involved, but volunteers will get a closeup introduction to some of the rarest and oldest creatures on Earth.
Richard Benson, head of the Department of Paleobiology at the National Museum of Natural History, expects that an engineering perspective can lead to new insights into the structures of fossil animals, possibly adding to the understanding of how extinct animals once stood and moved. Benson's idea of connecting mechanics and fossil anatomy comes as little surprise to his colleagues. He started his college education in mechanical engineering and shifted toward geology, with an eye toward becoming a mining engineer, before the fossil bug bit him and he turned to paleontology.
A paleontologist looks at fossils and sees a wealth of anatomical information about structures, similarities and differences of shapes, and common origins of diverse forms. The engineer, Benson believes, will see something else: the skeleton as a design of a dynamic system.
Experience with fossils is not particularly wanted. Ralph Chapman, who oversees the museum's morpho-metries laboratory, said the museum is interested in volunteers "with no biological baggage." Prospective paleo volunteers can reach Chapman at email@example.com.
Plenty has been discovered about the Triceratops since Boss and Gilmore finished their work more than nine decades ago, and the new exhibit will reflect what paleozoology has learned. For instance, paleontologists have come to believe that the animal stood more erect than the 1905 mounting does. As a result, in the new exhibit the Triceratops will stand straighter on its front legs.
Another important change will be back feet. In 1905, no one had yet found any for a Triceratops, so Boss and Gilmore substituted those from a duck-billed dinosaur, which was the right size but of a different sort altogether.
Rear Triceratops feet have been found since then, and new parts will replace the stand-ins of the original mounting.
The museum is having casts made of bones from real Triceratops back feet and plans to send them out for scanning. Dimension Data will scan a cast of the assembled feet.
According to Chapman, it is likely that Scansite 3-D will scan castings of individual foot bones later. Scansite 3-D, which markets scanners, scanning services, and engineering consulting services, plans to use an MS-3030 laser scanner from Cyberware of Monterey, Calif. These scanners are controlled by Silicon Graphics workstations. David Bassett, Scansite's director of technology, said the Cyberware scanner has a size limit of a half-meter. Some of the bones in the Triceratops are as long as 54 inches.
The replica in the Dinosaur Hall will reflect all the revisions and updates, and will be in the same scale as the original. The 1905 interpretation has its own historical value, and the museum intends to preserve it. A one sixth scale model is being built to serve as a record of the original mounting.
Paleobiologists will also use the scale model, along with Virtual Surfaces' animations, to study possible reconstructions of the dinosaur's stance. As Jabo pointed out, one-sixth scale is easier to manipulate than a full-size Triceratops. The actual skeleton is just under 25 feet long and more than 8 feet high.
Construction of the replicas has begun under Federici's company, Scansite 3-D Services, which is coordinating the work of the various companies involved and is overseeing the creation of prototypes for the models.
The one-sixth scale replica is being turned out through stereolithography. Virtual Surfaces is preparing the .stl files. According to Andersen, the files for stereolithography are sampled down until they contain between 2 and 3 percent of the points of the original scans.
Federici said she has enlisted the aid of a large consumer products manufacturer to donate stereolithography services and materials to the project. The company, she said, prefers to remain unidentified.
Milling of the larger replica bones is necessary because of the practical limit to the scale of stereolithographic parts. Satellite Models of Mountain View, Calif., has been lined up to do the milling. The company will use a five-axis computer numerically controlled router from Quintax Inc. of East Canton, Ohio. The material, according to Frank Stoeckel, a project manager for the company, has not been decided. Satellite Models tends to use either a urethane foam that's supplied by General Plastics of Tacoma, Wash., or a foam called Ren Shape, made by Ciba Specialty Chemicals of East Lansing, Mich.
The current project began late last year, and the Triceratops has largely been dismantled for conservation. Jabo said the museum intends to have the full-scale display replica ready by next spring for a temporary exhibit called Dinofest 2000. If the new Triceratops makes that date, it will come home to the Dinosaur Hall later in the year.