5. Digging In

Television documentaries about dinosaurs notwithstanding, the life of a vertebrate paleontologist is not an endless series of exciting discoveries. Like all other scientific research, advances in paleontology come at a price: financial, physical, and emotional. An excavation like that of Sam is never easy, and it isn't accomplished overnight.

In dinosaur research three-fourths of our time goes to organizational details, planning, coordination of personnel, preparation of grant proposals, budget management, report writing, and communicating with landowners and land managers. We have to be good at persuading supervisors and museum directors of the worthiness of our plans to excavate and study yet another specimen. We have to train volunteers, lead site visits or tours of the laboratory, all the while responding to memos and deadlines. The other fourth of our research time goes to reading, corresponding with colleagues, attending meetings, and actually working with the fossils. The popular image of paleontologists as educated ruffians or displaced cowboys is as misdirected as the notion that our science proceeds in an orderly and predictable fashion.

Just as we carve out of our personal lives large blocks of time for the research, so do we scrape and scrounge for support. Even well endowed museums (and these are few indeed) dedicate little funding for dinosaur research. It's expensive, time-consuming, slow to generate tangible returns, and often requires extensive travel to study other collections.

For every hour of field time in an excavation, moreover, several hours of laboratory work are required before the bones are ready for study; for my work on dinosaurs in the Morrison Formation I estimate a ratio of 10:1. That is, for every hour of labor on-site, ten hours of laboratory work follow. Alternatively, blocks of bones encased in plaster-and-burlap will accumulate in museum storage areas for decades. The excitement of discovery and excavation fades, and the harvest of unprocessed fossils becomes a museum director's nightmare and a curator's embarrassment.

A large excavation is risky from the standpoint of professional time and institutional money. To undertake a large-scale field operation requires a personal commitment to years of work, with the expectation that tangible results, such as publication and mounted specimens, may take a decade or longer to achieve. The larger the dinosaur, the greater the commitment. Likewise, the more complete the skeleton or the more extensive the site, the greater the commitment. All told, for a skeleton of a large dinosaur that may be complete or half complete, the expectations are daunting at best.

Location of bone fragment. The isolated bone fragment that we left on site in 1985 was buried in the sandstone ledge in the foreground. It so perfectly matched the color and texture of the rock in which it was embedded that it posed little danger of vandalism or unscrupulous collection. This fragment was evidence that we might find additional bones at the site, and it was thus vital for securing our first excavation grant from the National Geographic Society.

A full excavation of whatever bones might remain of Sam thus seemed an awesome prospect. We could have abandoned the site, bidding farewell to the embedded fragment of bone. In that case, all of our work would have been completed upon final preparation of the eight tail vertebrae and publication of the results. But that one bone called to me as the sirens did to Ulysses. Perhaps it would lead to more bones, and just possibly, given the right orientation of the skeleton as it lay on the edge of the mesa, much more would still be preserved beneath the ten-foot layer of sandstone and cap rock.

I had to find out. Besides, from a curatorial standpoint, if more of the skeleton were to be found, my colleagues and I could more thoroughly describe the bones and more confidently identify this dinosaur new to New Mexico.

Following my visit to the Smithsonian Institution and other museums of the eastern United States and my seminar at Los Alamos National Laboratory that had generated so many novel ideas for remote sensing, I applied to the National Geographic Society for excavation support. The single bone fragment we left on site in 1985, a bone measuring no more than ten square inches as exposed, proved suggestive enough to bring us this vital financial assistance. The initial grant was for a one-year excavation. Later, the Society would award a two-year extension, the Martin Marietta Corporation Foundation would supply a one-year grant for more excavation, and we would win a one-year grant from the National Science Foundation for remote-sensing experiments.

In early summer of 1987, supported by the National Geographic Society, we reopened Sam's quarry with great optimism, but with a tremendous sense of foreboding, too. By all our calculations, the skeleton would lead into the mesa, below a ten-foot layer of cap rock that would be as hard as concrete. On the other hand, that one bone fragment might lead to nothing: it was possible (but not likely by my reckoning) that we had found all the bones we would ever find, and that the grant would be wasted on digging in barren rock.

The first attempts by Los Alamos and Sandia scientists to locate buried bone had been provocative. We had potential targets in the subsurface, but we could verify them only by excavating. There might be nothing to excavate, or the bones might go so far and so deep into the hill that the award would be insufficient. Nevertheless, in this particular excavation, as in many dinosaur excavations, we were blessed with a crew of hearty volunteers that made it possible to stretch grant dollars much further than in other forms of scientific research.

We set to work once again with jackhammers, picks, chisels, and shovels. I predicted the skeleton would turn gradually to the north and remain at the same level. With that seat-of-the-pants projection, we laid out the area where we should excavate. Like warriors armed with toothpicks, we began our assault on the mountain of rock.

Meanwhile, the new Museum of Natural History in Albuquerque had opened in early 1986, and we were overrun with visitors and administrators. The public loved the wizardry of the state-of-the-art exhibits, and we were ill-prepared for the masses that crowded the museum halls. Volunteers were essential.

Peggy Bechtel became the museum's coordinator of volunteers, one of the most harrowing jobs in a museum. She trained and managed scores of volunteers, leading them through museum orientation, formal instruction on museum exhibit themes, details of the museum exhibits, and museum ethics. She trained volunteers to serve as docents and for behind-the-scenes labor such as carpentry and exhibit production. Her volunteers immediately became indispensable. They doubled and tripled the museum's work output. Peggy's organizational talents made the difference between opening with a mediocre museum and opening with a spectacular start. More importantly, during the next two years her volunteers contributed thousands of hours of unpaid help to the museum's operations. They became the life force of the museum.

Many of the volunteers wanted more than indoor museum work. Some hoped to excavate fossils, especially dinosaurs. After all, a "save the dinosaurs" sentiment generated the funds that built the museum. We thus had a ready and willing work force for excavating Sam.

Between the initial excavation of Sam's eight tail vertebrae in June of 1985 and the museum's opening in January of 1986, a special group of New Mexico volunteers rallied around paleontology, hungry to participate in fieldwork and research on fossils. These dedicated people organized fund-raising efforts to support an international conference on dinosaur tracks. They assisted with the excavation of an eight-ton block containing dozens of skeletons of Coelophysis, a small bipedal predatory dinosaur from the Triassic, which is now on display in the Ruth Hall Museum of Paleontology at the Ghost Ranch Conference Center in northern New Mexico. And they provided the labor for scientific study of dinosaur tracks at Clayton Lake State Park in the northeastern part of the state. Eventually they organized as a separate support group, the New Mexico Friends of Paleontology. This group is now a nonprofit corporation whose activities are dedicated to the support of paleontology at the New Mexico Museum of Natural History.

The members lent a camper shell for shelter and storage at the site of Sam's excavation. And, most important, during the first two years of excavation they contributed their time. The early stages of excavation would have been impossible without them.

Among the most avid volunteers was Wilson Bechtel, Peggy's husband, who was near retirement from his job in a local movie theater. He and Peggy together assumed full excavation responsibility in the third year of the project, under the sponsorship of the Southwest Paleontology Foundation, after I moved to Utah. Their dedication and experience during this time introduced them to the world of professional paleontologists. Their contributions as professional-level researchers in this project have culminated in several published papers as coauthors with me.

So, after a frustrating delay in finalizing the National Geographic Society grant, and after training our first group of volunteers at a nearby dinosaur site, we set to work at last. Fully two years after Frank Walker had shown me the string of bones on the edge of this remote mesa west of Albuquerque, we were finally digging in.

The bone we left in the quarry as a marker proved important for orientation, marking the level of the skeleton and its position on the side of the mesa. But it was not a vertebra. Instead it was a downward projecting bone, called a chevron, positioned at a joint in the tail. Most vertebrates with long tails, including many mammals and all dinosaurs, have chevrons. These bones protect the venous blood vessels that return blood to the body from the tail and they separate the muscles of the lower half of the tail.

But after that chevron was removed the next move was not obvious. Where we expected to find the continuation of the tail vertebrae lay only barren sandstone. The realization that the tail might not continue into the side of the mesa became a recurrent nightmare, an unstated fear that we might not find any more bones and the National Geographic grant would prove fruitless.

But within a month we did find more bone--a few yards away from the bone fragment and in the same line and orientation as the original eight vertebrae. This new bone proved to be the seventh vertebra of the tail, and we soon uncovered the edges of the sixth, and then the fifth, still tightly connected. We seemed to be approaching the pelvis.

Discovery of more vertebrae. Subsequent excavation revealed a continuation of the vertebral column from the bone fragment left embedded in 1985. The four vertebrae beneath the square-meter scale are encased in plaster and burlap for protection during removal and transfer. These are caudals 4 through 7. In the foreground caudals 1 through 3 are partially exposed, showing the rigor mortis curvature of the vertebral column that made the skeleton turn "into the mesa."

Close-up view of the four vertebrae (caudals 4, 5, 6, and 7). The bones are lying on their right sides.

Our hopes revived, we bore down with intensity. The vertebral column turned slightly toward the mesa, as I had predicted from my expectation that the skeleton had arched in a rigor-mortis curve before the carcass was buried. Our spirits soared, despite the hard labor. In some places, especially near the bone, the sandstone was hard and unyielding. Our quarter-size jackhammers and shovels gave way to hammers and chisels as we carved the sandstone away from the surface of the bones, following them into the edge of the mesa. One vertebra led to another, and we were immediately encouraged. Maybe the skeleton would be articulated, and (hoping against all reasonable expectation) just maybe it would be complete all the way to the neck and head.

By late summer we found ourselves facing the full ten-foot wall of sandstone that buried Sam's skeleton. I was pleased, however, knowing the National Geographic grant money would be fruitful. But looking at that formidable rock, I felt a special bond with my namesake, David, facing Goliath.

The tail vertebrae were so tightly connected that I could now entertain my wildest dream: that the skeleton might continue in this state of articulation to the pelvis and beyond, to the ribbearing vertebrae and the neck. Except for the missing end of the tail (which would have eroded off the edge of the ledge thousands of years ago), we might have a complete skeleton. Now that's optimism, but without X-ray eyes to look into the rocks, and with the remote sensing tests only in the experimental phase, we could do little else but hope and dig.

Sam's bones presented an unusual problem. Two features of their preservation made our work exceedingly difficult. First, the bones and the surrounding sandstone were the same buff color, an unusual condition for dinosaur bone. Generally, fossil bones are dark, even in light-colored rock, and almost always they are easy to distinguish from the host rock. Not so with Sam's bones. The colors are so perfectly continuous that we had to dedicate considerable time to training volunteers just to differentiate bone from sandstone. And even professionals could be fooled. Working at the site alone one day, I spent several hours chopping gently with hammer and chisel through what I thought was sandstone when I suddenly realized I was well into the neural spine of a vertebra. We all had trouble with the lack of color distinctions.

Second, the contact, or boundary, between bone and rock was gradational rather than abrupt as is usual for fossil bones. Before Sam's skeleton was buried, the bones had checked and cracked as they desiccated in the Jurassic sun. As sand settled around the skeleton on the sand bar, individual sand grains worked their way into the internal fabric of the bones, beneath and between the checks and cracks on their surface. One hundred and fifty million years later, distinguishing bone from rock was sometimes impossible, and often we worked in fear that we were chopping right through fossil bone. Some volunteers insisted on working well away from the bone, preferring heavy labor such as shovel-and-wheelbarrow to the close-in work with hammer and tiny chisel.

Progress became at times almost imperceptible, even after hundreds of hours of work on-site. Satisfaction came only with persistence and long-term perspective, our only antidotes for the frustrating tedium of separating bone from rock. But the reward was there bit by bit, as we followed the skeleton, uninterrupted, into the mesa.

By October, 1987, the end of the field season, we had followed the tail forward to the sacrum, which consists of five enormous vertebrae all fused into a huge boxlike structure for transferring the weight to the hind legs. Pelvis bones, too, were attached. The left ilium, oriented upward as the skeleton lay on its right side, had been eroded slightly before this region of skeleton was buried. But the right ilium on the underside of the skeleton was intact, still fused with the lateral projections from the sacral vertebrae. Three of the four lower pelvis bones (the left ischium, right ischium, and right pubis) were in place, too, but separated and collapsed from the position in life. Quite unlike the upper surfaces of the caudal vertebrae and sacrum, the ischium and pubis lay in unconsolidated sand, almost soft enough to sweep away with a brush.

Discovery of the pelvis bones was a tremendous find, and it seemed to confirm the value of one of the remote-sensing experiments. The sacrum lay beneath one of the targets indicated by the first ground penetrating radar experiments.

Removal techniques have changed little in the past hundred years. To extricate a large fossil bone from the ground requires nothing more than labor, some cheap materials, and considerable ingenuity. No one has produced a "cook book" manual on how to excavate a skeleton--probably because no two excavations are alike. Each skeleton, and each bone, presents special problems to the excavators. Keeping the bones intact is the primary concern: once freed from above and on their sides from the rock in which they have been encased for millions of years, fossil bones naturally expand. They crack and break along fractures that were microscopic during burial but which open widely when the bone is no longer confined by rock. Bones exposed on the surface of the ground for a long time almost always fracture and split into small chips and splinters; fossil bones only infrequently erode from their surrounding matrix and lie on the surface, unbroken and whole, like the string of Sam's tail vertebrae exposed by natural erosion.

Today's paleontologists use a technique invented by our nineteenth-century counterparts to stabilize bones for removal and transport. To keep them from expanding and falling apart, we apply wet tissue paper, then wet newspaper in layers, onto the surface of the exposed bone to function as a separator and cushion, then strips of burlap soaked in wet plaster laid over the contours of the bone and rock to lock everything, bone and matrix alike, tightly into place. Just as emergency-room doctors use plaster and gauze to set a broken bone, locking it in place to allow healing in the correct position, so do we apply a form of first aid to fossil bones. Perhaps "last-ditch assistance" is a more appropriate metaphor.

The product is a block of bone and rock tightly bound together in a plaster jacket. Properly applied, the jacket encases the block on all sides. If padded and secured, blocks can be transported from an excavation site to a museum or laboratory with little fear of damage. As the plaster-and-burlap bandages do not seal the contents from air, the bone and rock dry gradually, thus preventing the growth of mildew and fungus. Wellexcavated bones locked in plaster jackets can be stored in museums indefinitely, awaiting laboratory preparation. Some plaster jackets are never opened, the energy and optimism of the excavation team having been captured by more important discoveries. Some museums store hundreds of unopened plaster jackets because excavation support is much easier to secure than laboratory support.

Orientation of skeleton. Wooden models (background) mark the position of the original eight tail vertebrae excavated in 1985. On discovering the continuation of the vertebral column where these excavators are working, we established the orientation and trend of the tail, lying on its right side. The left bend in the trend of the tail (the rigor mortis arch) is barely evident at this stage of the excavation. Wilson Bechtel, left; unidentified volunteer, center; Peggy Bechtel, right.

Because the bones of Sam's vertebral column were articulated, excavation was a special challenge. Articulation is, of course, the optimal condition of preservation from the standpoint of descriptive anatomy, but it makes excavation far more problematic. Keeping the vertebrae in articulation became an important goal. We wanted to study the position of the bones in their actual orientation, connected at the joints as they were in life. Many of our crew of volunteers had participated in the excavation of the eight-ton block of Coelophysis skeletons from the Ghost Ranch quarry, so removing Sam's bones in one-ton or two-ton blocks was well within their range. One ton is still a formidable size; such a block is as heavy as a small car, half the weight of a small elephant.

Wilson Bechtel excavating a block for removal. The 1,500 pound block of four vertebrae had to be undercut to free it from the sandstone and then encased in plaster-soaked burlap reinforced with lumber. The operation is as dangerous as it appears in this photograph and should be undertaken only when supervised by an expert.

When we uncovered the top surface of the first four caudals (numbers 4, 5, 6, and 7 as we later counted back from the hips), I decided we should take these out in a single block, keeping them together end-to-end for laboratory preparation later. The centrum (or main body) of each vertebra lay in sequential contact with the centra of adjoining vertebrae--like a series of tin cans on their sides, aligned top-to-bottom. Quite remarkably, the contacts had remained in perfect orientation: each bone had not settled separately in the sand before burial. I expected that we would be able to remove them as a single and manageable block within a week or two after commencing the downward exposure. The block would weigh only about three hundred pounds, I thought. It would be easy to handle and move to the laboratory.

I was wrong. This experience first brought home to us the immensity of Sam's skeleton. I had calculated the projected length of the entire body from the original eight caudal vertebrae, but I couldn't tell from those figures just how large the other bones would be. We dug deeper and deeper, seeking the lower limits of these four newly exposed vertebrae. We aimed to leave them perched on a pedestal of underlying rock. This is the trenching phase of excavation: finding the edges of a bone and digging straight down, to a level beneath the lowermost level of bone, in preparation for undercutting. Although I had reckoned the trench would need to be no deeper than one foot, we instead had to trench nearly three feet alongside these four bones. Their surprising dimensions more than doubled my original estimate of their diameter.

That realization was exciting. But the practical consequences were harrowing. The projected weight of the plaster jacket that could contain these four bones quadrupled: it would now weigh more than a half ton.

This change in dimensions of the block prompted us to reconsider the initial plan. Should we split the block into two manageable sections? By so doing, we could more easily handle the jackets in the quarry. We could expect to turn them over more easily in order to plaster their bottoms, and then it would also be easier to lift them out and haul them to Albuquerque. But splitting the block would surely induce more fractures through the bone. Should we dare to keep them intact?

It would have been impossible in the field to separate the vertebrae exactly along their natural joints; they were solidly fused by the sandstone, with the hardness and durability of concrete. Moreover, the bone in the joints was nearly indistinguishable from the rock. The only way to separate the vertebrae would be to drive a chisel into a crack and allow fractures to open. These fractures, however, would surely pass through  bone rather than between  bone. We had already used this technique of breaking through bone to remove the eight tail vertebrae from the initial discovery phase of excavation. We didn't want to break the bones just for convenience. We decided to reinforce the entire, four-bone block with lumber and steel and to find a way to lift it from the quarry intact.

Exposure of remaining caudal vertebrae. The vertebrae (caudals 4 through 7) beneath the square-meter scale became Block A. The continuation of the vertebral column (caudals 1 through 3) in the foreground became Block B. The dimensions of the tail vertebrae closest to the pelvis surprised us: the side-to-side dimension of each vertebra was nearly a meter, and the height exceeded a meter.

Simultaneous with the trenching around these four vertebrae ("Block A") other crew members continued work on exposing three more in succession. We learned later they were the first three (and largest) of the tail. Caudal vertebrae numbers 1, 2, and 3 became "Block B." Still in articulation and lying on their right side, they curved upward in a tight rigor-mortis arch and connected directly to the vertebrae of the sacrum. The next decision involved a knotty problem: how to separate Block A from vertebra no. 3 in Block B?

We considered cutting through the rock with a chainlike device equipped with sharpened teeth, similar to a chain saw used for cutting trees. This option offered control over the separation between the two blocks, and the edges could be matched upon completion of laboratory preparation. However, sawing through the bones would leave a half-inch gap that could never be replaced. As attractive as this idea was for reasons of controlling the separation of the two blocks, I couldn't tolerate even a half-inch gap. Instead, we devised a more complicated procedure that would generate breaks in the bones preferentially along already existing natural fractures. These could then be matched and repaired in the laboratory. Block A would weigh at least half a ton.

We would begin this multi-step procedure by deepening the trenches around Block A on three sides, leaving its forward-most centrum (and the overlapping spines that interconnect above each centrum) attached to the centrum of the first vertebra (no. 3) in adjoining Block B. This would produce a pedestal beneath the bones that we could then slowly chip away. By adding props (firewood turned on edge, with shims), we could support the block underneath, without moving it, as we carefully undercut it. We knew this would be a rather hazardous operation because we had to be partly under the block as the undercutting progressed. With every several inches of undercut, we would add new layers of burlap and plaster to hold the bone and rock in place, add new props, and then resume chipping away at the pedestal.

We custom-designed an A-frame support for a hoist, to be constructed from six-inch diameter drilling pipe used for oil wells (cost for materials and welding, about $100). With a chain hoist mounted from the cross beam situated over the quarry, we would lash the block with a chain for support in case the block should fall during the undercutting stage. This would be for safety, and for stability later, if we should have to break it away from Block B.

Thus far, the plan was working beautifully. Trenching and undercutting progressed exactly as we envisioned, albeit slowly, taking the better part of two months to complete. Once the last of the pedestal was removed and the plaster jacket completely encased the block except for its contact with Block B, it was ready to move. Tightening the chain supports from the hoist, we began to remove the props one by one, expecting the block to settle slightly in the process and a break to develop naturally along the contact between Block A and Block B. Our strategy was to control the break by controlling the fall of Block A, and then to immediately plaster the broken ends of both blocks to lock broken bone into position.

Tension mounted as we removed the props. Wilson Bechtel checked the hoist and cinched it up a notch, expecting it to cushion the fall and allow us to control the direction of movement when it finally gave way. But there had been no dress rehearsal for our new and untried technique. The block was supposed to settle and break away. We were astonished when the last prop came out, and the block remained suspended, supported entirely from its side-connection with Block B. On reflection we now realize this phenomenal support of a half-ton lever system was possible because of the exceptionally strong and continuous cementation of the bones, which acted like steel reinforcement bars.

This turn of events and the profound silence it produced weren't in the plan. We replaced two props for safety and our group of fifteen excavators held a powwow. We faced the same problem as before, how to separate the two blocks, only now the risks were higher.

No saws, I insisted. We had to use breaks that could be perfectly matched and repaired in the lab. Wilson resumed control of the hoist, and I climbed atop the overhanging block. It didn't budge. I jumped a little, then higher. Still it didn't budge.

With a small sledge hammer, I pounded a small, cold chisel into the position where we wanted the blocks to separate. I could feel the resonant vibrations in the block with each blow of the hammer. With all my senses intent on Sam's tail, I struck the chisel again and again, driving it deeper into the rock and bone. Throughout, I was poised to jump to safety if the block should suddenly give way. Wilson strained, more from anxiety than labor, and I deliberately drove the eight-inch chisel deeper and deeper.

Suddenly, it happened. A vertical crack opened and I jumped by reflex more than thought. Wilson tugged on the chain hoist. The block fell slightly, a perfect few inches.

Wilson let the chain off a notch, so that the block's weight could fully separate the contact with Block B. It settled slightly, and then he gave a mighty heave on the chain. The block fell a little more, and twisted away from the break. We found props to steady it, to hold it still for plastering. Peggy and volunteers had paper, plaster, and burlap ready for the break, and we immediately covered both broken ends to lock the bone and rock exactly in position. Block A, suspended from the A-frame, held steady as we finished closing the jacket.

The seat-of-the-pants engineering had worked. Later we moved Block A from the quarry to the New Mexico Museum of Natural History. We began trenching Block B, which contained the three tail vertebrae closest to the pelvis. Then we began work to free the sacrum, a fusion of five enormous vertebrae, each about four feet in diameter. This became Block C. Both blocks came out of the quarry with the same procedure. Block B was slightly larger than A; Block C was the biggest. When weighed on public scales used by freight trucks, it measured 3,200 pounds. Block C was not, however, to be our largest block.

Block B hanging from the hoist managed by Wilson Bechtel. The A-frame and hoist allowed us to use a controlled fall of the block to free it from the end of what would later become Block C. The vertebral bones are not evident here, but the shape of the vertebrae in cross section is approximated by the exposed side of the block facing the camera. This face had to be covered with plaster and burlap, and reinforced with lumber, before it could be hauled away from the site.

All three blocks (A, B, and C) we removed from the quarry using a truck-mounted winch provided by the Bureau of Land Management. We winched each block out on skids, then lifted them to a truck bed with the A-frame hoist. Properly jacketed and secured, these enormous blocks of bone and rock can be transported without injury even over rough jeep trails. In days gone by, horses pulled blocks like these on wagons, or the workers crafted rails and rail-cars for transportation. Our techniques, in truth, seemed no more advanced than theirs.

Hauling the sacrum (Block C) from the site. This specially designed dinosaur truck was lent to us by the Earth Science Museum of Brigham Young University so that we could haul the sacrum from New Mexico to Utah. We were given laboratory space at BYU in which to prepare the bones so that we could make direct comparisons with the BYU collection of giant sacral bones from Dry Mesa Quarry that probably belong to Supersaurus.

As the excavation proceeded to the sacrum, the line of curvature of the vertebral column turned directly into the hill. Over the sacrum lay about eight feet of hard sandstone, and the overburden layer became higher in the direction we projected for the rib-bearing vertebrae--the dorsal vertebrae or, simply, the dorsals. By 1990, under the Bechtels' continuing supervision, the cadre of volunteers had removed enough overlying rock to follow the dorsals into the mesa. A total of seven vertebrae still in articulation were uncovered, out of a possible ten that the original carcass would have contained. Some of these dorsal vertebrae had ribs still attached in living position; other ribs had detached and collapsed to ground level before burial.

This position of the dorsals was the sole target identified by the seismic remote sensing that proved true. The Oak Ridge Team only a few months earlier had predicted that at this spot lay the best possibility for bone in the whole area. The success was rewarding.

We had proved that the skeleton did indeed trend into the hill, and with every new bone we exposed its size became more and more impressive. We proceeded to expose the dorsal vertebrae along their sides. In so doing, we began to encounter dozens of gastroliths, or stomach stones. These stones, generally the size of a plum, were purportedly swallowed by Sam to enhance the muscular grinding necessary to digest coarse plant materials. Some gastroliths were in direct contact with ribs and others were scattered away from the skeleton.

This important discovery slowed the excavation considerably, because we had to expose each gastrolith in turn, plot its position on the quarry map, label it, and take photographs. Peggy Bechtel took charge of the gastrolith excavations, and Wilson supervised and engineered the exposure of the dorsals in a huge block. At this point, about four or five workers, mostly volunteers, were on-site on a typical day. From my new post as state paleontologist of Utah I visited whenever I could on weekends and holidays, but the Bechtels had assumed the day-to-day management of the excavation.

Upon completing the trenching around the dorsals, to our great disappointment we came to the end of the articulated portion of the vertebral column. From the sacrum forward we had seven (possibly eight) vertebrae intact and still joined. These particular dorsals are sometimes called the presacral vertebrae, if one counts them from the sacrum forward. Counting from the base of the neck rearward, however, is the usual manner. Sauropods generally have ten vertebrae between the base of the neck and the sacrum. The first presacral is therefore the tenth dorsal, the second presacral is the ninth dorsal, and so on.

In mammals the vertebrae between the neck and sacrum (or sacro-iliac joint) are differentiated into those with ribs (the thoracic vertebrae) and the lower (or rear) vertebrae, which lack ribs (the lumbar vertebrae). Dinosaurs do not have this differentiation; all vertebrae between the neck and the sacrum have ribs. Thus there is no need to distinguish a lumbar and a thoracic region--hence the generic term dorsal vertebrae , a confusing terminology at best.

The line of Sam's vertebrae abruptly ended near the front of the rib cage, at the seventh or eighth presacral (or fourth or third dorsal) vertebrae. This disappointing termination of the series became a convenient boundary for establishing Block D: it would contain the seven presacrals and their ribs in various states of attachment. Later Peggy and Wilson discovered the next dorsal vertebra, just forward from the seven dorsals. It had fallen on its joint face and lay over a set of gastroliths.

With the discovery of that vertebra we made one more connection with the remote sensing experiments. It had to do with the drill hole no. 2 that had filled in when no bone was discovered.

On removing sandstone around the front part of Sam's torso, Peggy and Wilson came across the second core hole, filled with debris. They removed the debris, but didn't think more of it as they dug deeper around the dorsal vertebrae nearby. The right sides of these vertebrae stretched more than a meter down from the level at which they first appeared. This portion of the vertebral column, containing seven dorsal vertebrae, took on a monolithic aspect, shaped like a giant loaf of bread carved out of the rock.

Murphy's hole. This was hole no. 2 in our initial selection of core-hole positions. The choice of this hole marked our best guess for the occurrence of subsurface bones, as indicated from combined data from radar and magnetometry. The hole, however, almost perfectly bisected the notch in the neural spine of a hidden vertebra, the third dorsal. That vertebra, in turn, covered a large suite of gastroliths embedded in the rock below the bottom of the hole.

One morning as they started digging, with the sun at a perfect angle, Peggy peered into the cleaned-out hole no. 2. She saw bone. In disbelief she looked more closely. The hole had indeed hit bone at the lowest bone level in the quarry, about nine feet beneath the top of the mesa. That bone was a total surprise, because we thought the vertebral column had been broken apart and the forward bones had been carried away by stream action before the skeleton was buried in the sand bar.

On excavating the surprise bone, they discovered that it was the next dorsal vertebra, isolated and detached from the succeeding vertebra to the rear, and offset from its living position by a distance of four feet. The core hole itself had barely scraped the edge of the bone, and since the core sample was a cylindrical plug created and extracted by the inner surfaces of the pipelike drilling device, the scrapings of the drill did not come to the surface, and we did not see bone fragments in the spoils. More spectacular, the hole was a perfect bull's-eye: it fit almost exactly between the two projections of the neural spine, the V-shaped notch that gives these vertebrae a slingshot shape.

This vertebra, the third beyond the base of the neck and the furthest forward of all the dorsal vertebrae we would find, had fallen forward and come to rest on a pile of gastroliths. The crew carefully exposed the edges of the vertebra and the two dozen gastroliths surrounding it. Each day Wilson plotted the positions of newly exposed gastroliths, labeled them with specimen numbers, and took documentary photographs. The excavation of this isolated bone required removal of considerable rock beneath the vertebra. Much later, during laboratory preparation of the block containing the surprise vertebra, Peggy and Wilson would discover two dozen more gastroliths they had not seen in the quarry. The vertebra had protected these gastroliths from scavenger and stream action prior to burial.

This was the last dorsal vertebra found in the quarry. We found no other bones in the immediate vicinity, but later we found neck vertebrae, downstream from the torso and in the access road far from the quarry. Core hole no. 2, now renamed Murphy's Hole, and Murphy's vertebra (technically, the third dorsal vertebra or the seventh presacral vertebra) took on new significance. Our original decision to position the second core hole in a likely spot for bone, in order to calibrate the remote sensing, was a good selection after all. The underground truth came three years after the coring.

What did those sensing techniques, the ground-penetrating radar and the proton free-precession magnetometry, "see" in the subsurface when they came up with positive readings? Bone? The dense cluster of gastroliths? Both? Or something else, in which case our positioning was purely accidental? I have ruled out the last idea; our data were real and probably related to the bone and gastroliths, but whether one or the other or both, remains problematic.

We faced the same problem we had with the seven presacrals that had stayed together: how to plan, undercut, and jacket a block for removal from the quarry. The same goals applied: minimize fracturing, leave sufficient rock to hold the bones tightly in place, keep everything as intact as possible. With our success in removing Blocks A, B, and C, we were confident we could handle an even bigger block. Block D would weigh several tons, at least double the weight of the sacrum (Block C).

The A-frame and hoist were not designed to carry such heavy loads, however. We had to change tactics. This time we relied on the experience gained with the eight-ton Coelophysis block at Ghost Ranch. We would use the same technique for Sam's dorsals.

The first part of the procedure was exactly the same. Using small jackhammers and chisels we outlined the limits of the bones, dug down several feet below the bones, and produced a trench on each side that was about eight feet long, four feet wide, and three feet deep. Then we did something new. In the trenches, the crew gently chiseled tunnels every two feet or so beneath the enormous block. We plastered underneath the block as each tunnel was expanded, placing props to support the block's underside.

Eventually we expanded the tunnels until the entire underside was free from the ground and the block was supported only by props of firewood and shims. We then fixed hefty green timbers, each about a foot in cross-section, parallel to the block along its sides. Above these we set cross timbers through the tunnels. These would become skids. Then we set new props and shims between the cross timbers and the underside of the block to lock everything into place. Now the block rested entirely on the timbers, designed like a makeshift sled. To prevent flexing, the timbers were locked together with steel bolts and braces.

Preparing Blocks A and B for removal had taken about three months each, Block C about a year. Block D required nearly two years, as the quarry was expanded and the gastroliths uncovered and mapped. Fortunately, we were able to acquire additional grants to keep the excavation going without a break and long after the initial one-year grant had been spent. Late in 1991 we hired a wrecker to pull the block containing the dorsal vertebrae from the quarry to a holding place nearby. The skids worked beautifully, and nothing in the block shifted. Finally, in 1992, Block D was moved by truck to the Museum of Natural History in Albuquerque. The museum had dedicated sufficient laboratory space for the crew to begin the long work of removing and preparing the bones.

At the public scales used by truckers to weigh freight, we weighed the hauling truck before and after delivery. This cargo must have been one of the most unusual loads ever to cross those scales. The block weighed five tons.

Reconstruction of the skeleton of Seismosaurus based on the known elements. Positions and anatomy of the four neck vertebrae that we collected are conjectural; these bones were isolated in the quarry and heavily eroded. Not shown in this perspective are the ribs and lower pelvic bones of the opposite (right) side, which were recovered in the excavation. The peculiar kink in the tail is based on the anatomy of the vertebrae at the beginning of the downward bend. This kink appears to be unique to Seismosaurus.

The articulated part of the skeleton had thus been successfully excavated by 1992, but there was plenty more to do on-site. Because gastroliths were scattered over a broad area, we had to continue to expand the quarry floor to map their occurrence and collect them for the study. This slow and tedious process required back-breaking manual labor and meticulous handwork in order to expose them without disturbing their positions.

Four isolated bones were discovered late in 1991 and 1992. All were heavily eroded cervical (neck) vertebrae that had been displaced downstream from the main part of the skeleton. Excavation of these bones was relatively routine, after the experience of the truly big blocks. Those were the last of the bones we were able to locate. All the other neck bones and all those of the head were missing, as were the smallest vertebrae of the tail. Nevertheless, we were pleased. We had found all of the vertebrae from the shoulders to the middle of the tail, all of the ribs, the complete sacrum and pelvis, and some of the chevrons.

To our great dismay and puzzlement, however, no leg bones were found with the skeleton. Trying to explain the loss of leg bones has been difficult, but the taphonomic history of the skeleton, the subject of chapter 7, offers some clues. First, let's take a closer look at the several hundred gastroliths--and what they imply about Sam's digestive system.