Roy sidled closer and studied the dog for a moment. "Cherokee, show mort," he said. The dog stood up, slowly sniffed his way around the tulip poplar, and then sat again, in almost the same spot as before. This time, he bent down and touched his nose to the ground at the base of the tree. "Good boy! What a good boy!" The dog leapt to his feet and whirled, just in time to catch a knotted-up towel Roy had pulled from a pocket and tossed in his direction. With the force of a bear trap snapping, the dog's jaws closed around the fabric, and he began biting and thrashing his head, as if he were trying to dismember a rat. With one paw, he held the end of the bundled fabric on the ground and shredded it with meticulous savagery.

"Glad that's not my throat he's got ahold of," commented Dewar.

After the towel was reduced to bits, Roy led the dog back to our waiting group. "It looks like maybe there's something near the base of that tulip poplar," he said.

"No kidding," said Arpad. "I guess it's my turn."

He opened the back door of the Subaru and brought out the TopGun Freon detector. It squealed when he switched it on, then the noise died down to an occasional chirp as Arpad walked toward the base of the fallen oak. We followed, since the gadget-unlike the dog-wasn't p.r.o.ne to distraction by people or extraneous smells.

Stopping midway between the dead tree and the live one, Arpad bent down and eased the tip of the wand through the leaves and into the soil. The detector continued to chirp at the same slow rate. Stepping closer to the tulip poplar, he repeated the maneuver, with no discernible change. Next he positioned himself right where the dog had indicated and took another reading. The chirping might have sped up slightly, or I might simply have imagined that it did. Arpad frowned, looking puzzled and slightly embarra.s.sed. "As cold as it's been, it could be that the Freon compounds just aren't volatilizing," he said. "Or maybe they're long gone, if we're looking for something sixty years old."

"Or maybe the dog's just smarter," said Emert, earning a scowl from Arpad.

He took the Freon detector back to his car, swapping it for his prototype sniffer. As the gizmo fired up, I noticed how much I preferred its understated clicking to the Freon detector's electronic squeal. As before, Arpad stopped short of the target area, gently working the instrument's probe into the top of the soil. It continued to click quietly, almost like a clock ticking. Despite the chill of the day, I thought I saw glimmers of sweat on Arpad's brow, and I realized that he had a lot riding on this field test. If the dog gave a positive alert but Arpad's sophisticated instrument did not, should we excavate anyway? I thought we definitely should; after all, the dog had an impressive track record in other searches, and he seemed to show no hesitation or doubt once he started zeroing in on the tulip poplar. There was no guarantee we'd dig up anything, but it seemed only fair to give the dog the benefit of the doubt-after all, if we didn't trust the dog, we shouldn't have enlisted him in the search.

But would Arpad-a former student and now a valued colleague-take offense if we seemed to trust the dog more than the gizmo? I hoped not, but I knew scientists could be sensitive if it appeared their work was being questioned.

As I was turning over the alternatives in my mind, trying to settle on the most diplomatic way of handling the dilemma, I became aware of a quietly insistent sound. Arpad now stood at the center of the circle with the gizmo's probe in the ground, and the slow, steady ticking had given way to a sound almost like muted machine-gun fire. A smile spread across Arpad's face. "Eureka!"

"Cool," said Miranda.

Thornton reached into his pocket and pulled out a folded handkerchief, which he knotted into a ball. I was puzzled, until he said "Good boy" and tossed the handkerchief onto the ground in the direction of the gadget's probe. Suddenly I glimpsed a streak of movement at the edge of my vision. Moving at lightning speed, Cherokee swooped in, grabbed the handkerchief, and began ripping it to bits.

Miranda burst out laughing. "Holy c.r.a.p, that was fast," she said. "Serves you right for being a smart-a.s.s." Thornton just flashed her that grin again, bigger and more sheepish than ever. Miranda turned to me; it might have just been the effect of the chilly breeze, but her cheeks looked pink. "Does this mean it's our turn to look now?"

"I think it does," I said. I tapped the two Oak Ridge detectives on the shoulder as Miranda and I started toward the truck. "You guys mind giving us a hand?"

They followed us to the back of the truck and I handed a rake and galvanized-metal bucket to each of them. Miranda grabbed the two shovels, and I carried a large plastic bin containing smaller items: evidence bags, trowels, rubber gloves, a tape measure, a compa.s.s, a handheld GPS unit, a topographical map, my digital camera, a clipboard, pens and Sharpie markers, and a blue plastic tarp. I spread the tarp near the area we were about to excavate, and we laid the rest of the gear on it.

I began, as always, by taking pictures-several wide shots at first, showing the entire area, the vehicles, and the group of people. Then an inspiration hit me, and I took several shots of the fallen tree, the small valley, and the concrete silo, reproducing Novak's perspective as closely as possible. The comparison photos would be an interesting addition to the file, I thought. An interesting footnote to Oak Ridge history. An interesting thing to show Isabella over pizza. Next I took tighter shots of the fallen tree, the area near the base of its trunk, and the orange survey flag Arpad had stuck in the ground. Miranda switched on the GPS unit, held it over the flag, and pressed a b.u.t.ton to save the lat.i.tude and longitude coordinates. I found it amazing that a three-hundred-dollar gadget, about the size and shape of a calculator, could home in on satellites hovering thousands of miles overhead, pinpointing and remembering this precise location on an isolated hilltop: an electronic X marking a tiny spot on a big planet. I marveled at the technology, though I still didn't entirely trust it. That's why we had the compa.s.s and tape measure: in addition to marking the site on the topo map, Miranda would draw a more detailed sketch of the search area, showing the dirt road, the fallen tree, and the excavation, with compa.s.s directions and measurements-the diameter of the excavation, for instance, and how many feet west of the trunk the dog and the sniffer had alerted.

I had tried to talk Miranda into letting me bring one of the other graduate students in her place-I worried that the burns on her fingers hurt, and I feared she might damage them-but she insisted on coming. "I'll wear an extra pair of gloves," she said, "and it'll be fine." I hoped she was right.

After I'd taken a dozen or so photos and Miranda had sketched the key landmarks of the site, we began to rake the leaf litter off the soil. When the big tree had been ripped from the earth, long ago, its roots had torn a crater in the ground, six or eight feet in diameter and several feet deep. Gradually, though, the crater had filled as dirt fell from the edges, rainwater trickled down the sides, and decades of leaves swirled into the hole and crumbled into dust. By now all that remained was a slight, subtle hollow-with a sixty-foot tulip poplar growing from it. If not for the ma.s.sive oak trunk touching one edge of the rim, the low spot would have seemed simply a slight, random variation in the surface of the ground. By excavating carefully, I hoped Miranda and I could work our way back to the original, deeper contour of the hole, as a starting point in our quest for whatever might lie at its center. It wouldn't be easy, though.

"So," said Miranda, "that tree sure is in the way. Wonder what we could do about that pesky tree?"

"Just a thought," said Emert, picking up on the sly tone and the elbow-in-the-ribs emphasis, "but I'm thinking chainsaw. If only we had a chainsaw right about now."

They laughed; Roy and Arpad and the ORNL guard looked puzzled, so Emert told the chainsaw story. "Go ahead," I said. "Rub it in. But next time your heart is breaking, don't expect sympathy from me."

Roy spoke up. "I feel your pain, Doc. I'm pretty attached to my Husquevarna. Matter of fact, it's in the back of the truck. If you promise not to steal it, I might be willing to share the love."

The Husquevarna wasn't as nice as the Stihl-it didn't feel quite as solid, somehow-but it sliced through the eight-inch trunk in a couple of minutes. I cut the tree at about waist level, first, then-once it was down-cut the stump almost flush with the ground. I thanked Roy for the saw, handed it back, and then picked up the three-foot length of trunk and carried it to my truck. Emert asked, "You running low on firewood?"

"Souvenir," I said.

The edge of what had once been the crater in the ground-the border between "hole" and "not hole"-wasn't at the surface, so I used a shovel to remove a thin layer of topsoil, beginning within the slight depression and skimming outward, beyond the rim. The shovel slid easily at first, which told me that the soil here was loose; after about a foot, though, I encountered more resistance: the resistance of packed, undisturbed earth. I lifted the shovel and looked at the swath I'd just sliced. Sure enough, closer to me, the soil appeared lighter, fluffier, and more crumbly; then-across a faint and irregular but unmistakable line-the soil was denser and darker, infused with rocks and clay that appeared to have lain undisturbed since the dawn of time.

"Okay," I said to Miranda, "here's the rim. How about we excavate about halfway around the circ.u.mference, then work in from the edge?"

"Whatever you say, Kemo Sabe," she said.

"Excuse me, Dr. Kemo Sabe," said Thornton. "Can I ask a dumb question?"

"No such thing as a dumb question," I said.

"That's not what the instructors at the Academy used to tell me," he said. "Why start excavating at the edge? Why not just aim right for the bull's-eye, which seems to be somewhere around that stump you just made?"

"If we dig straight down and there is something there, we'll keep knocking dirt down onto it," I said. "The sides of the hole will keep collapsing. Plus we'd be on top of the bones; we might end up breaking some of them. Coming in from the side means a little more digging-but a lot more control."

"Ah," he said. "Anything we can do to help you?"

"Sure," I said. "If you don't mind lifting buckets of dirt, you guys could haul out dirt as we excavate."

"Sounds like something we might be able to handle," he said.

"Arpad," I said, "how long since you've used a trowel?"

"To dig up bones, or to plant tulips?"

"To dig up bones."

"Not so long ago that I've forgotten the backaches," he said. "Ten, twelve years, maybe."

"About time you brushed up," I said, handing him a trowel.

I was about two feet in from the rim of the crater, and about eighteen inches below the level of the leaves and twigs Miranda and I had raked off the surface, when my trowel hit something hard. Using its triangular tip, I flicked at the soil underneath what I'd hit, and as the dirt fell away from the object, I gradually made out the distal end-the elbow end-of a humerus, an upper arm bone. "Eureka!" I said, echoing Arpad's earlier exclamation. Burrowing a bit farther, I unearthed the medial ends of the radius and ulna, the bones of the forearm. From the angle at the elbow, I could tell the arm was slightly flexed, with the hand probably somewhere in the vicinity of the hip. "This is the right arm," I said. "He's lying facedown. a.s.suming it's a male." I troweled away more soil, exposing the distal end of the forearm, the loose, pebbly bones of the wrist, and the carpals and metacarpals of the hand.

Emert leaned in and squinted at the stained bones. "You're sure it's human," he said, "not a bear? I saw the bones of a bear's paw once, and I'd have sworn it was a hand or a foot."

"Well, unless these Oak Ridge bears are smart enough to tell time, I'm pretty sure it's human," I said, "because it's wearing a man's wrist.w.a.tch." With the tip of my trowel, I pointed to a disk of corroded metal hidden beneath the wrist.

"Eureka indeed," said Thornton.

Before I even had a chance to ask her, Miranda left the spot where she'd been working and came to kneel beside me. We'd done this so many times, our teamwork was seamless, wordless, and almost telepathic. I shifted to the upper arm and began excavating toward the shoulder and head; Miranda began working her way along the hand and then down the right leg.

As I troweled my way along the shoulder and toward the area of the head, the dirt began to drop away, revealing the rounded surface of a skull. Working with only the tip of the trowel, I started teasing the soil free. Occasionally I was forced to trade the trowel for small gardening shears, so I could snip away roots that clutched at the bones.

As the back of the skull came into view, I saw a prominent b.u.mp at its base. The b.u.mp-the external occipital protuberance-had once served as an attachment point for muscles at the back of the neck. The b.u.mp's presence and prominence told me that the skeleton was definitely male, and a robust male, at that. I'd been fairly certain of the s.e.x just from the size and muscle markings on the humerus, not to mention the wrist.w.a.tch, but the external occipital protuberance confirmed it.

The head was rotated, so that instead of facing straight down, it was turned toward the left shoulder; it was tilted slightly backward at an odd angle as well. For a moment I wondered if the neck had been broken-hard to tell, with all the soft tissue gone-but I quickly rejected that theory in favor of another, simpler explanation: the body had simply been rolled down into the hole, and had come to rest slightly askew.

With three of us excavating, the work moved fairly quickly, but even so, it was midafternoon before we had worked our way around the entire skeleton. Rather than removing bones one at a time, we left the skeleton in place until we had exposed it completely, digging down on all sides so that the bones lay on a raised platform of earth-a technique called "pedestaling." The soft tissue had decayed completely, as had all the clothing, except for thin, crumbling remnants of the leather soles of the shoes.

One by one, Miranda and I snipped the tulip poplar's remaining roots, freeing the bones from their grasp. By the time the roots were cut and the stump pulled from above the torso, the stump itself looked skeletal and dismembered.

The torso posed a challenge. Normally a body in a shallow grave would gradually collapse, the vaulted rib cage flattening as the cartilage decayed and the ribs detached from the spine and sternum. In this case, though, a latticework of tree roots supported the ribs.

By this point I'd been on my hands and knees for the better part of four hours, so I groaned my way to my feet and clambered out of the hole we'd dug. Excusing myself from the group, I wandered into the woods, ducked behind a large tree, and took a much-needed bathroom break. Arpad and Miranda headed off in other directions to do likewise. In years past, I'd had female graduate students for whom the lack of bathroom facilities in the field posed problems, ranging from minor inconvenience to full-blown crisis, but Miranda had long since jettisoned most of her modesty about such matters. "Oh, good grief," I'd once heard her chide a squirming female colleague, "we're out here scooping up some dead guy's rotten guts, and you're too refined to tinkle in the bushes? Get over it already." From the direction of a pine thicket about fifty yards away, I heard a yelp. "Jeepers," Miranda shouted, "you guys have no idea how cold it is out here."

"Next time I'll bring a propane bun-warmer just for you," called Arpad.

Once we regrouped, I photographed the skeleton from every angle, including wide shots and close-ups, and then prepared to remove the bones from the pedestaled grave. I asked Miranda to record the inventory of the skeletal elements-the listing of every bone-and Arpad to bag them in evidence bags.

I began with the skull. As I eased it from the ground, lifting and rotating it, I got my first glimpse of the right temporal bone, the oval bone just above the ear. A small, neat hole pierced the bone. The location coincided exactly with the dark circle on the head of the dead man in Leonard Novak's photographs. "It's you," I said to the skull. "It really is you."

The hole was about a quarter inch in diameter at the outer surface, but it flared wider as it bored through the bone. The beveling was the unmistakable signature of a bullet blasting its way through the skull. Any kid who's ever shot a BB gun through a plate-gla.s.s window has seen the same physics, on a smaller and less lethal scale: as the BB enters the gla.s.s, it creates a shock wave that fans out like a cone, fracturing a steadily wider cross-section until it emerges on the other side of the window amid a shower of tiny shards.

The entry wound was about an inch above the opening for the right ear, and judging by its perfect roundness, the bullet had been fired directly toward the center of the cranial vault, since an angled trajectory would have caused an oval hole. There was no exit wound on the left side of the skull. I gave the skull a vigorous shake and was rewarded with a clattering inside. "I think we've still got the bullet," I said. "Probably a .22. The entry wound's small, and the bullet didn't have enough oomph to punch out the other side."

"Had enough oomph to do the job, though," said one of the detectives.

"Funny thing about a .22," I said. "Seems like a sissy gun, but the bullets tend to ricochet around inside the skull and really chew up the brain. Sometimes a .22 does more damage than a larger-caliber bullet that just blasts right through."

"Reckon what he was doing," said Emert, "when that bullet hit him?"

"Trying to steal atomic secrets," said Thornton. "Or trying to keep them from being stolen."

"Or making a pa.s.s at the wrong guy's wife or girlfriend," I said.

"Pleading for his life," said Miranda.

We had not unearthed any artifacts besides the watch in the process of pedastaling the skeleton. Now, though, as we removed and bagged the bones, I came across seven small objects embedded in the soil. Six were metal b.u.t.tons-one in the region of the chest, where a left shirt pocket would have been; three along the midline of the body, s.p.a.ced between the chest and the pelvis; and one at each ankle. The seventh object, at the waist, was a rectangular plastic buckle, olive green, with a rotting bit of canvas webbing still threaded through it. As I handed each object to Arpad, as carefully as if it were a precious gem recovered from a pharaoh's tomb, the law enforcement officers crowded around to inspect them. At the sight of the buckle, Emert voiced what I'd been thinking. "This guy was wearing army coveralls," he said. "I've still got my dad's in a chest in the attic." There were no coins or keys in the grave, which led me to believe that the pockets had been emptied. I was therefore not surprised, though I was disappointed, that the grave contained no dog tags.

"So we've got a dead G.I. from World War II here," said Emert. "Swell. There were only, what, ten thousand of those here in Oak Ridge?"

I thought we were finished-through the bones, down to the dirt-when the tip of my trowel snagged on a clump of clay. But it wasn't clay. A chunk of it broke off, and when it did, it revealed odd striations within the soil. Looking closer, I began to discern a lump, a shape, about a foot long and slightly narrower, somewhat paler than the rest of the red clay lining the grave. I probed gently at the edge I had exposed. The striations were quite thin-paper-thin, I realized, as the proportions of the rectangle registered in my brain. "I don't know what this guy was doing when he died," I said, "but it seems to have involved a mighty thick stack of papers."

CHAPTER 29.

I HOPED THE BONES MIGHT TELL US MORE THAN THE papers did about the dead soldier. Officially he was case 09-02, the second forensic case of 2009, but a number was a poor subst.i.tute for a name.

One of my UT colleagues in the College of Agriculture-a scientist in the Forest Products Laboratory-had confirmed that the rectangular lump we dug from the grave was indeed a stack of paper. From the thickness, he estimated it to be somewhere between 400 and 500 pages, and he said it appeared to be a low grade of typing paper-long on wood pulp, short on linen fibers. Because it was cheap and pulpy, it tended to crumble into chunks, rather than peeling apart into individual sheets. "I managed to pry apart a few fragments," he told me, "but I'm afraid there's not much there. Ink smears and mold. Whatever's written on those pages, it hasn't stood the test of time."

The bones, on the other hand, had held up well. After a day of simmering in hot water, Biz, and Downy fabric softener, followed by some gentle scrubbing with a toothbrush, Miranda had laid the clean, caramel-colored bones of G.I. Doe-that's what she'd dubbed 09-02-in anatomical order on a table in the osteology lab. She had also taken skeletal measurements with a 3D digitizing probe. After entering the measurements in the Forensic Data Bank, she plugged them into ForDisc, the software developed by one of my computer-savvy colleagues at UT. According to ForDisc's a.n.a.lysis of the data-the size of the skull, s.p.a.cing of the eye orbits, width of the nasal opening, and the length and diameter of various bones, among others-G.I. Doe was a white male of about 180 centimeters, or five feet eleven inches, in stature. None of that surprised me; after all, ForDisc had been programmed to make, quickly and automatically, the kinds of calculations and a.n.a.lyses physical anthropologists had spent years learning how to make with calipers, and slide rules and calculators.

ForDisc was not, on the other hand, programmed to estimate age. Estimates of age required looking at multiple features of the skeleton and making judgments, sometimes complicated or subjective ones, about the degree of development or maturity in the bones. Those weren't the kinds of automatic calculations a computer program could perform.

It was my custom, when doing a forensic examination of a skeleton, to keep quiet until my students had examined the bones and offered their opinions. Miranda was used to this, and she required no prompting, beyond a tilt of my head and an inquiring lift of my eyebrows. She began by setting the skull upside down in a doughnut-shaped cushion, exposing the upper teeth and the roof of the mouth. Then she picked up the lower jaw in her left hand, pointing with the little finger of her right hand at the teeth. "So. Both third molars in the mandible are fully erupted," she said, "which would indicate an adult." Still holding the mandible in her left hand, she touched her pinkie to the wisdom teeth in the upper jaw, which were small and well below the level of the second molars. "The third molars haven't erupted in the maxilla," she said, "but these appear to be impacted, unlikely ever to erupt through the gums. So, his teeth say that he was probably at least eighteen years of age."

She laid the mandible down and lifted the skull from the cushion that cradled it. Cupping it in her left hand, she used the tip of a probe to trace the pattern of the four sutures, or seams, in the roof of the mouth. One of these, the palatomaxillary suture, ran from one side of the palate to the other, like a line drawn between the second molars. Another, the incisive suture, also ran sideways, just behind the four incisors at the front of the jaw. Two of the sutures ran along the midline of the roof of the mouth: the intermaxillary suture extended from the front of the mouth to its intersection with the palatomaxillary suture, and the interpalatine suture ran from that intersection to the back of the palate. In most subadults-people under eighteen-these four sutures were not fully closed; the joints were still in the process of being filled with new, growing bone. By eighteen, though, they tended to be fused, and during the decades of adulthood, the suture lines gradually smoothed and faded, or obliterated, sometimes disappearing altogether. In 09-02, the maxillary sutures were fully fused, but their lines remained vividly drawn. "The maxillary sutures are fully fused, so we know he was an adult," Miranda said, "but probably a young adult. Not a geezer, for sure." I smiled at the way Miranda bounced back and forth between scientific formality and slang.

"I'm gonna save us both some time here," I said. "I know you know the basics. You've probably got the whole osteology handbook memorized by now, right?"

"I'm a little fuzzy on some of the specifics on page two," she said.

"What's on page two?"

"All that Library of Congress copyright stuff," she said.

"I'd be worried if you were wasting brain cells on that. Okay, let's skip ahead. Instead of talking me through the whole skeleton, show me what you think can pin down his age more precisely."

"Three things," she said. "First, the anterior iliac crest." She pointed to the large, curving edge of the hipbone and used her finger to trace a line near the edge. A faint seam there marked a joint in the broad bone, as if the Creator had decided the hips were a touch too narrow and had gone back and tacked another sliver of bone along the outer edge. It wasn't actually an afterthought, of course, but an epiphysis, a joint that had remained open while the bones were still growing, then closed when the final growth of adolescence was done. "The epiphysis is completely united, so that suggests he was in his early twenties, at least, and maybe mid-twenties or later. The prime of life, in other words." She wiggled her eyebrows, and I smiled; Miranda was poised between her mid-and late twenties.

"I'm following you so far," I said, "even though my brain is well past its prime. Second?"

"Second, the pubic symphysis." She picked up the two halves of the pubic bone and showed me the face where they met at the midline of the body. "The symphyseal face shows a lot of beveling in the ventral area," she said, pointing to the rear portion of the joint. "That suggests late twenties or beyond."

"Are you basing that on the work of Todd, or McKern and Stewart, or Suchey?"

"All of the above," she said.

I smiled. "Good answer. Third?"

"Third is always most important," she said. "Third, the clavicle." She picked up the left collarbone, which was nearer the edge of the table, and indicated a faint, smooth seam near the end that joined the shoulder. "The lateral epiphysis is fully fused, which you'd expect, since the dude's a grown-up. But the medial epiphysis"-she pointed at a ragged, incomplete seam near the bone's other end-"isn't completely united yet; it's still undergoing terminal union."

"Leading you," I said, "to conclude what?"

"To conclude that G.I. Doe was thirty. Plus or minus a year or two."

"Bravo," I said. "I agree. Now let's look at trauma. Did you see any skeletal trauma other than the wound to the head?"

"Nothing," she said. "There's a small amount of osteoarthritic lipping on some of the vertebrae, but that's just the beginning of age-related wear-and-tear, not trauma. Nope, I think one shot to the head did it."

I picked up the skull and, using a pair of calipers, measured the diameter of the entry wound in the right side of the cranium. The hole was almost perfectly circular, but not quite. At its widest, it measured nearly a third of an inch-about the size of a .32-caliber bullet. At its narrowest, though, which was the crucial dimension, the hole measured less than a quarter of an inch. That made it too small to be caused by anything larger than a .22. Short of cutting off the top of the skull, there was no way to get the calipers inside the cranium to measure the hole's diameter as it broke through the bone and entered the brain, but by shining my key chain flashlight into the hole, I guessed the inner diameter to be nearly half an inch, because of the conical beveling gunshots always produced. The force of the bullet had also caused three small fractures, each about an inch long, to radiate outward from the hole.

A small, ragged blob of metal lay on a tray beside the skull. I laid the skull down and picked up the blob. Although it was small, it felt heavy and soft. "You got the bullet out," I said.

"I did," she said. "I managed to shake it out the foramen magnum," the opening at the base of the skull through which the spinal cord exited. "Took me back to my childhood days, when I shook coins out of my piggy bank."

As I studied the deformed bullet, I was struck by its shape. "Does this remind you of anything?"

"Reminds me not to get shot in the head," she said.

"No, I mean the shape."

She plucked it from my hand and held it in her fingertips, and the gesture clutched at my heart: It was the same way Garcia had held and studied the iridium in the morgue. It was the same way Miranda had plucked the deadly pellet from his grasp with these very fingertips. Now, though, they were tipped with white gauze.

"Well, I'll be," Miranda said. "This bullet is a dead ringer for a mushroom cloud."