Blind Man's Bluff: The Untold Story Of American Submarine Espionage

Five — Death Of A Submarine

It was May 27, 1968, and the end of a long day. John Craven was driving along the Potomac, on his way home, when the news ninety-nine men were missing. L came over the radio: the USS Scorpion (SSN-589) was missing;

Barely two months had passed since U.S. intelligence had realized the Soviets had lost their Golf submarine. And Craven was still helping Bradley figure out where it had gone down when this latest news came. Craven listened hard for details about Scorpion, but there weren't any.

Nobody had any idea where Scorpion was or what had happened to her. All they knew was that the 3,500-ton nuclear attack submarine was due back in Norfolk, Virginia, and had failed to arrive. She hadn't been slinking off Soviet shores or even plumbing new depths, as the USS Thresher had been doing when she was lost five years earlier. Scorpion had simply been cruising through the Atlantic Ocean on a straight track for home. Just like the World War II submarine she was named for, Scorpion had vanished without a trace and seemingly without reason.

Craven slowed his car at the next exit and turned for the Pentagon. As Craven stepped into the controlled pandemonium of the War Room, all he knew was that, as the Navy's top deep-water scientist, he would be needed. A submarine was missing; ninety-nine men were missing.

Surveying the crowd of captains and admirals and other officers already there, Craven sensed something he had never encountered in a room full of top-ranking military men: abject fear.

The fear could be seen in the tensed faces of the men who stood scrutinizing a huge wall chart mapping Scorpion's assigned track, and it could be heard in the shaken tones of others who were intently studying navigational charts strewn all over the room. Men were laying out hypotheses and search patterns. They were plotting Scorpion's track, creating a path for search planes above and looking for the sparse undersea mountains below. Just a few months before, the USS Scamp (SSN-588) was nearly lost when she rammed into an undersea mountain in the Pacific in her race to go monitor a Soviet missile test. A similar accident, and Scorpion might be lost forever. Then again, those mountains might also be the only places along her path where she and her crew could have sunk without meeting instant, crushing death.

Other officers were studying the positions of nearby Soviet ships and submarines, wondering whether any had crossed Scorpion's path. People all over the room were trying to weigh the possibilities, wanting to believe that Scorpion was still intact, her crew stranded but alive.

"What can my organization do to help?" Craven said over the worried voices, the roar of competing conversations, and the rustle of the charts. Nobody looked up or even seemed to notice him speaking from the doorway. Most of these officers knew nothing of Halibut, of Craven's role in preparing her for deep-sea searches, or even of his success in pinpointing the atomic bomb the Air Force lost in the deep Atlantic near Palomares, Spain. To most of the rank and file here, Craven was just another skinny engineer. Those few who did know him well found him to be a man full of odd ideas and strange search methods that didn't sound like anything ever penned in a Navy manual. Few of the officers in the War Room that day would have believed that Craven might he their best and perhaps only chance of finding Scorpion.

Craven repeated his question. This time, someone answered: "We haven't been able to find Scorpion on the acoustic nets. We don't know where it is. If there's anything you can do with respect to that, do it."

With that, Craven was left on his own, left to try to figure out why and where Scorpion had vanished. Odds were worse than a million to one against anyone finding the boat. She could have been anywhere on a track that covered 3,000 miles of the Atlantic.

The families of the Scorpion crew had begun to worry as early as February 15, 1968, three months before Craven heard the news on the radio, three months before rumors began swirling through the sub force that the Soviets might have sunk her.

There, standing on the dock tossing the final mooring line to the crew as Scorpion departed, was Dan Rogers, an electrician's mate who had risked his career by demanding to he transferred off the boat, writing to his captain, Lieutenant Commander Francis A. Slattery, that everyone on board was "in danger." The Navy had always portrayed the 252-foot-long sub as a gleaming showpiece, but Rogers said Scorpion was so overdue for a thorough overhaul that the crew had taken to calling her the "USS Scrap Iron." There were oil leaks in the hydraulic systems and seawater seeping in through the propeller shaft seals. Her emergency ballast systems weren't working, and the Navy had restricted her depth to 300 feet, less than one-third of the operational depth of other boats of her class.

There had also been a frightening incident three months earlier when Scorpion had vibrated so violently during high-speed maneuvers that she seemed to corkscrew through the water, sending huge pieces of equipment swaying on their rubber mountings. The cause was never diagnosed. Rogers and other crewmen feared that the problem could reappear at any time.

Most of the submarine fleet had undergone massive safety overhauls after Thresher was lost. The bulk of the work on Scorpion, however, had been postponed due to tight budgets and the relentless pace of intelligence operations, which were growing rapidly toward a peak never before seen during the cold war. As she set out, Scorpion was one of only four of the Atlantic Fleet's submarines that was still waiting to be refitted with post-Thresher safety features.

Rogers and his mates complained to Slattery that he and his officers weren't taking their concerns seriously. Rogers wasn't even released from the boat until he agreed to Slattery's demand that he erase the Cassandra-esque warning of "danger" from his request for transfer.

One month later, Scorpion was assigned to join in NATO exercises in the Mediterranean. She was sent there only because the Navy needed a last-minute replacement for Seawolf, the same submarine that Craven had bypassed in favor of Halibut when it came time to pick a special projects boat. Seawol f had knocked herself out of the fleet rotation by ramming an undersea mountain in the Gulf of Maine, badly crushing her stern.

The Mediterranean had become the latest cold war arena. Since the Arab-Israeli War in 1967, the Soviet Union had been sending growing numbers of attack subs armed with nuclear cruise missiles to stalk U.S. aircraft carriers and to try to trail U.S. missile subs roaming from a base in Rota, Spain. U.S. surveillance subs were also watching ports in Egypt, where some of the Soviet vessels stopped. The traffic was so thick that by December the Med saw its first underwater collision, between the USS George C. Marshall (SSBN-654) and a Soviet attack sub.

Most of the time, however, the problem was simply detecting the Soviet subs. SOSUS listening nets, which helped in other areas of the world, didn't reach into the Med-or, for that matter, down the west coast of Europe, a key Soviet route to the area. The Med itself has horrible sonar conditions, with saltwater meeting fresh water, warm meeting cold, all of it sending sonar bouncing in unpredictable directions. Besides all that, nobody on the U.S. side really understood how the Soviets operated in the Med, or how many subs they were sending. Indeed, submarine analysts in London and their counterparts in Norfolk, Virginia, were having long analytic arguments about Soviet operations, arguments that went on all the longer because there were so few facts to hack them up.

On the assumption that sheer numbers would fill in the gap left by expertise, the United States began trying to train its allies' sub forcesthose in southern Europe and the Middle East-in the art of sub-chasing. Scorpion, in fact, was sent to the Med to play rabbit, to be hunted by foreign forces as part of their training. For Scorpion's men, this should have been a plum assignment, one with the rare perk of port stops in sun-sprayed Spain, Italy, and Sicily. But many would have preferred to have stayed onshore with Rogers, at least judging from their letters home.

"We have repaired, replaced, or jury-rigged every piece of equipment," twenty-four-year-old Machinist's Mate Second-class David Burton Stone wrote to his parents on April 12. Stone sent his letter two months into the trip, just before Scorpion got caught in a dangerous game of chicken with a Soviet destroyer. The incident was typical of operations in the Med: both sides had taken to harassing the other at sea. When Scorpion surfaced to exchange messages with the USS Cutlass (SS-478), the destroyer raced forward as if to ram the submarine. With a crash seemingly moments away, the Soviet ship backed off.

"It did it three or four times," says Herbert E. Tibbets, the commanding officer of Cutlass, who watched the incident from his bridge. "I kept sweating, thinking, `I hope those guys hack it down this time."'

Reports of that incident and rumors of another mission have left many of the families convinced that the Soviets were the likely cause of Scorpion's destruction. According to the most virulent story, Scorpion supposedly was hit by a Soviet torpedo during a final mission in which she tried to chase a Soviet attack submarine away from a U.S. Polaris boat out in the Atlantic.

There was, in fact, a final mission, but it had nothing to do with chasing Soviet attack subs. It began in late April. Scorpion was on her final port visit, this one to Naples, Italy. From there, her men expected to be going home. Instead, they were told they were being sent to monitor strange Soviet activity. U.S. satellites had photographed a group of Soviet surface ships, just outside the Med, flying balloons about the size of weather balloons. The Soviet ships had been engaged in this baffling behavior for nearly a month. In the Pacific, the Soviets had been known to launch balloons equipped with electronic sensors in the vicinity of U.S. nuclear tests. Perhaps this was a new application of that spying technique.

Figuring that Scorpion was going to pass near the area on her way home anyway, Captain James Bradley, still the Navy's top submarine intelligence officer, ordered the sub to swing by and take a look. Slattery and the other Scorpion officers were distressed. After more than two months at sea, the officers wanted to go straight home, and they made that clear at a farewell cocktail party in Naples, cornering Bradley with their concerns. He was sympathetic, but the orders stood.

Scorpion set out toward the Soviet ships on April 28. Slattery stopped outside the breakwater at Rota, Spain, to drop off a crewman and a spook who had become ill, then went on. Scorpion lurked near the Soviet ships for two or three days before Slattery turned his sub for home. When he reached a safe distance from the Soviets, he radioed a message that he had collected a few photographs but little insight about the Soviet exercise.

It is not entirely clear where this group of ships was working, but declassified Navy documents cite one possibility. Air-reconnaissance planes had spotted two Soviet hydrographic survey ships-a submarine rescue ship and an Echo 11-class nuclear attack submarine-conducting an unspecified "hydro-acoustic operation" southwest of the Canary Islands, which lie about 300 miles off northwest Africa. Air reconnaissance was cut off on May 19 and resumed on May 21, just about the time Scorpion would have left the area.

"There were no observed changes in the pattern of operations of the Soviet ships, either before or after Scorpion's loss, that were evaluated as indicating involvement or interest in any way," the Navy would later report in a document prepared in 1969 by a court of inquiry into the Scorpion disaster and kept classified for years.

On the evening of May 21, the Scorpion's crew radioed in their location and reported that they had embarked upon their assigned route home, the "Great Circle Track" through the North Atlantic. Ordered to transit at 18 knots, they said they expected to arrive in Norfolk at 1:00 P.M., Eastern Standard Time, on May 27.

Admiral Thomas Moorer, the CNO, and Vice Admiral Arnold E Schade, commander of submarines in the Atlantic, began to worry when Scorpion failed to answer messages on May 23 as well as repeat messages over the next two days. They quietly asked a few Navy ships and planes to scan for signs of the submarine. No general alarm was raised. After all, Slattery and his men could be racing home underwater and out of radio contact.

Concern turned to fear on May 27, at 12:20 P.M. It was twenty minutes before Scorpion was supposed to arrive at Norfolk. By now, she should have been on the surface, her crew talking with the base. Schade initiated an intensive communications check. Ships and planes flooded the air with Scorpion's call name.

"Brandywine…. "

"Brandywine…. "

"Brandywine!"

There was no answer.

At 3:15 P.M., Scorpion was declared missing.

Back at the dock, the crewmen's families waited, waited for their husbands, sons, and fathers to come back from sea, waited in a spring rain that washed the dock clean. They knew nothing about the frantic radio messages tearing through the air around them. Then the Navy told them to go home, told them that Scorpion had been delayed. It was only when news reporters started calling that the families learned that their sons, husbands, and fathers were missing.

By the time Craven turned for the Pentagon, intelligence officers had already been frantically scrambling for acoustic evidence or other signs of an accident, a collision, or a battle. Reconnaissance pilots placed all known Soviet and Eastern Bloc surface warships, merchant ships, and submarines at least 50 miles away from any point Scorpion was expected to pass. The Navy would later report that there was "no evidence of any Soviet preparations for hostilities or a crisis situation such as would he expected in the event of a premeditated attack on Scorpion." Indeed, by the time Craven walked into the War Room, the Navy basically had ruled out Soviet involvement in Scorpion's loss.

Vice Admiral Schade set out himself to join the search on the USS Pargo (SSN-650). Rogers, the former crewman, went out looking as well, aboard his new submarine, the USS Lapon (SSN-661).

There was a moment when everyone on Lapon believed that Scorpion had been found. Lapon's radiomen picked up an SOS from "Brandywine." But soon it became sickeningly apparent that the message was a fake, a sadistic joke from merchant seamen or pleasure boaters.

Meanwhile, Craven launched a search that would take so many twists, and leave him so at odds with the rest of the Navy, that he himself would begin to wonder whether he had indeed gone mad. He began routinely enough, thinking of ways to acoustically delve the ocean depths. It was clear that the SOSUS listening nets were going to be useless. While the listening system in the Pacific had picked up that one pop, the only sign of the Soviet Golf's loss, the extensive SOSUS arrays in the Atlantic could not do the same thing. The Atlantic SOSUS system was designed to filter incoming noise, allowing the sonar nets to record the consistent clatter of machinery, the whir of submarine screws, and all the other music made by submarines as they move underwater, while muffling the blasts of oil exploration, undersea earthquakes, and the calls of whales. That sane filtering system would have eliminated any evidence if Scorpion had fallen to the ocean bottom, would have broken apart the terrible cries of a submarine imploding, rendering them nearly indistinct from the normal ocean din.

"How the hell are we going to find these poor bastards?" Craven muttered to himself. Within days, he would be named chairman of a technical advisory group convened to help find Scorpion by Robert A. Frosch, the assistant secretary of the Navy for research and development. Craven and the other group members were to report directly to the CNO and the commander of the Atlantic Fleet.

He began calling upon the small oceanographic research stations that dotted the Atlantic. Top on his list was Gordon Hamilton, a friend who ran an oceanographic laboratory in Bermuda that was funded by the Office of Naval Research.

"Hey, Gordon, do you have any hydrophones in the water that could have heard the Scorpion?" Craven asked without bothering to offer a greeting.

"Well, I don't, but part of my laboratory in the Canary Islands has a hydrophone in the water all the time," Hamilton answered.

The hydrophones generated mounds of scrawled paper, those peaks and blips that accumulated as pens moved over continuously rotating drums. There was a problem, though. Six days had passed since Scorpion's last message to shore, and laboratory workers were supposed to clean up and toss the records after two or three days. Any scrawls that could have registered a final tragedy aboard Scorpion should have gone out with the trash.

Still, Craven firmly believed that people rarely do what they are supposed to do. Housekeeping, he reasoned, is usually the first thing to go. Within a couple of hours, Hamilton called back. Craven was right. There were piles of paper all over the lab, and buried within those piles were two weeks of acoustic records-including eight separate ocean explosions or severe disturbances during the six days Scorpion had been out of contact. But the disturbances could have been caused by almost anything, including blasts from illegal oil explorations, a fairly routine sound ringing through the North Atlantic. And they could have come from almost anywhere, and from any direction.

With only one set of records, Craven had no way to come up with a geographic fix on any of the blasts. To do that, he would need to triangulate three separate recordings from three different hydrophones set up in three different points. Since he didn't have the data to come up with a precise fix, Craven worked backward, charting the times of the explosions against Scorpion's known path and speed. He came up with eight mid-ocean locations where he assumed the sub would have been at the time of any of the disturbances. Bathymetric charts showed all eight sites to be in waters deeper than 2,000 feet, deeper than the crush depth of a submarine.

Acting on Craven's data, the Navy sent planes to all eight spots. The pilots were looking for floating wreckage and oil slicks. They found none. The lack of debris was far from conclusive, given that the water was so deep. But Craven needed more to go on. The hunt for sonic evidence continued.

Independent of Craven's efforts, Wilton Hardy, the chief scientist of an elite acoustic team at the Naval Research Laboratory, the Navy's primary underwater testing facility in Washington, D.C., came up with the next clue. He knew that the Air Force kept two hydrophones near Newfoundland to track underwater shocks from Soviet nuclear tests. One was right off the peninsula of Argentia. The other was about 200 miles from there.

Hardy sent for the records, knowing he was playing a long shot. Both Air Force hydrophones were about as far from the Azores, and Scorpion's last-known position, as any listening devices could be and still be in the North Atlantic. And sitting right between the hydrophones and Scorpion's track was the largest chain of mountains on earth, the undersea Mid-Atlantic Ridge. The mountains were enough to block most sounds from the Azores.

Indeed, at first glance the Air Force records looked useless. There were none of the dramatic peaks that had been registered by the Canary Islands lab. But, to Hardy, it seemed that if he looked real hard, maybe squinted a bit, he could just possibly see something. He laid the Canary Island recordings directly on top of the Argentia recordings.

There they were, almost entirely buried in local noise, slight blips that seemed to match the more dramatic peaks picked up by Hamilton's lab. Hardy called Craven, who was by now coordinating the Navy's entire acoustic search effort. Craven decided to convince himself that the Argentia recordings were neither coincidence nor phantoms.

If the Argentia blips were worthless noise, then the plots would probably fall hundreds of miles or thousands of miles from the relatively tiny line of ocean that made up Scorpion's track. But if the new data pinpointed any one of the eight events picked up in the Canary Islands on that tiny line, the acoustic matches would almost certainly have to be valid.

Hardy found it first. There, right on Scorpion's track, was an explosion strong enough to tear through a steel hull and send a submarine, flooded, toward the ocean bottom.

There was no telling what caused the explosion. But 91 seconds later, there were a series of much louder blasts and there was no mistaking what caused those. Craven and Hardy were convinced that they had to be implosions, the agonized shouts of a submarine collapsing in on itself, compartment by compartment breaking down with the force of nearly 500 pounds of TNT.

The men on the submarine could have survived the initial explosion, if that sound was indeed from Scorpion. They might have lived long enough to see her walls begin to quaver inward, but that would have been all. Nobody could have lived through the first implosion. That shock would have sent the tail section and the bow section plowing into the center of the submarine, like a papier-mache model crushed in front and in back with a single, violent clap. The cataclysmic heat and the shock of that would have killed everyone on board in less than onehundredth of a second. The men would all be dead even as the ocean pressures continued to pummel Scorpion: a second implosion four seconds after the first, then another five seconds later, then two seconds, then three seconds, then seven seconds, then another and another and another. Three minutes and ten seconds after the first explosion, it would have been all over. Three minutes and ten seconds of destruction before the ocean went suddenly quiet.

Recorded only eighteen hours after Scorpion's crew had sent word they were heading home, the blasts meant that the sub had managed to travel less than 400 miles toward Norfolk.

It was now four days after Scorpion had been declared missing. Craven called the chief of Naval Operations to tell him that Scorpion was probably lost forever. Moorer wasn't ready to hear that. He wasn't about to tell the crewmen's families and the nation that there was no hope based on a bunch of tiny, almost indiscernible blips on paper. The fact that they occurred at a point right on Scorpion's track, at a moment when she was expected to be there, was enough to convince him only to declare the spot "an area of special interest." Then he waited to see whether any of the planes, ships, and submarines turned up anything else.

Rear Admiral Beshany, commander of the submarine force, began funneling all press inquiries to Craven. But the scientist remained under strict orders to avoid the word lost and even the suggestion of death. It wasn't until another six days had passed with no sign of Scorpion that Beshany and Moorer were forced to accept that Craven and Hardy were right. On June 5, Moorer announced that the Scorpion was "presumed lost." Hours later, the secretary of the Navy formally declared Captain Slattery and his ninety-eight other officers and crewmen legally dead.

But Scorpion was still missing. Without examining the remains of the sub, the Navy would never know what had gone wrong. Without that understanding, the nuclear submarine fleet would forever operate with the fear that a fatal flaw, somehow overlooked, could cause another catastrophe. Absent proof the crewmen were dead, their families might never be able to shake the thought, against all logic and against all available information, that the men might have been captured and were alive somewhere, perhaps in a Soviet prison.

And so began the second phase of the search. Now it was up to Craven and his team to find Scorpion and to find out what killed her. He turned his attention back to the acoustic echoes.

The site of the first explosion-now being called "Point Oscar" marked where his search would begin. But that still left him far from finding the sub. Thermal layers in the water could have distorted the sounds of Scorpion's loss as they traveled to the Canary Islands and the Argentia hydrophones. Craven calculated that there could be ten miles of error for any of the spots mapped by the triangulated data.

Also, the water at Point Oscar was 2 miles deep. The Scorpion would have stopped imploding about 7,000 feet before she hit bottom, cutting off the acoustic trail. Depending on how fast she had been traveling, and in what direction, and depending on the force of implosion and the position of her stern planes as she fell, she could have been thrown miles further.

All that meant that the submarine could be anywhere within a 20mile-wide circle, leaving a vast, unknown universe to search. And the art of deep-sea search was still in its infancy.

In starting the Scorpion search, Craven had far less data than he had when searching for the Soviet Golf in the Pacific. The Navy decided to send a surface ship to comb the area surrounding Point Oscar. There was no thought of sending Halibut on this search; Halibut was a boat designed for secrecy, and there was little need to shroud the fact that the search was going on since the Soviets could easily read about the missing submarine in American newspapers.

Instead, the ship the Navy employed was the USNS Mizar, an oceanographic survey vessel. She was a 266-foot-long former polar supply ship that had been converted to research at the start of the Navy's post-Thresher scramble to the deep. For this mission, she would be under the direction of Hardy's team at the Naval Research Laboratory, where she was based.

Mizar carried towed cameras, less-advanced versions of Halibut's fish, and with those she would start the slow, painstaking survey of the ocean bottom. The search would be led by Chester "Buck" Buchanan, a civilian oceanographer and senior NRL scientist.

As Buchanan set out, he knew he was in for a long haul. Crawling at two knots, it would take Mizar months to cover the area. But the captain was a tracker by nature, short, stocky, and good-naturedly pugnacious. He began to grow a heard the day Mizar left port, a Vandyke, declaring that he would shave only when he found his quarry.

Staying in constant contact with Hardy and Craven as they sorted through the acoustic crumbs, Buchanan began moving Mizar in circles over Point Oscar, finding little more than what seemed to be iron-rich meteorites. Following the Navy's lead, Mizar then began scouring the area west of Point Oscar. The Navy reasoned that since Scorpions had been heading west toward Norfolk, that was the best direction to search.

Meanwhile, Craven began digging for more evidence, anything that could help direct Mizar from shore. He set about trying to map each implosion in the hope that he could figure out how far Scorpion had traveled before the final sounds of her loss subsided.

He found much more.

Craven's map showed that Scorpion had not been traveling west toward Norfolk during her final moments. Instead, Craven's calculations surprisingly showed that the submarine had been moving east, back toward the Mediterranean. Perhaps a submarine could turn if it were fleeing from another boat, but intelligence officials had already told Craven that they were all but certain that the Soviets were not involved. It had to be something else.

The scientist went straight to Beshany's submarine command. He had one question. "What could make a submarine go in the wrong direction?"

Craven asked the same question of several captains and admirals. Each time he got the same answer.

A submarine turns around 180 degrees when a torpedo activates while it is still on board, an event submariners call a "hot run." The boat turns because that triggers fail-safe devices on a torpedo, shutting it down. The same safety devices keep the weapons from turning and blowing up the submarines they are fired from.

Scorpion carried a load of torpedoes, armed and ready for the worst, as did all cold war attack submarines. There were fourteen Mark 37 torpedoes, seven Mark 14s, and two nuclear-tipped Mark 45 Astor torpedoes. Hot runs were particularly common with the Mark 37s, and if there had been a hot run, Slattery would have called "right full rudder," ordering a 180-degree turn the moment the torpedo room reported the problem. Any captain would have-the maneuver is one of those things that are drilled into submariners until the reaction becomes simple reflex. In fact, Scorpion had recovered from a hot run in December 1967, six months before she was lost, precisely because Slattery had followed the standard procedure.

That had to be it, Craven reasoned. Scorpion was traveling west, and that had to mean that something had gone wrong with one of the sub's torpedoes. Somehow it had activated. And somehow it had exploded.

Craven began to dig around. He learned that there was a flaw in the onboard testing equipment that could easily have triggered a hot run. And he learned that torpedoes, along with almost every other piece of equipment on board, are routinely tested as submarines make way for home.

One of Craven's favorite maxims was, "If something can be installed backward, it will be." And in this case, it was true. Several submarines had reported hot runs as a result of electric leads on the test equipment being installed backward. The problem had become common enough that the commander of the Atlantic Fleet issued warnings.

With that known flaw and the acoustic data, it seemed to Craven that Scorpion's fate had been determined. Scorpion had been battling a hot-running torpedo, probably created when somebody mistakenly reversed the leads during a test. Only her turn to the east had been too late. The logic, the evidence-it all fit. Craven was convinced.

There was only one problem: almost nobody else agreed with him. The sonic experts, the torpedo experts, the submarine commanders, all listened as Craven held forth with his theories, his evidence, and his logic, his voice rising and falling as if offering a Shakespearean soliloquy, albeit one punctuated with his own trademark maxims of the deep sea. But nobody of any rank, from the chief of Naval Operations on down, thought Craven could be right.

Hardy, the acoustic expert at the Naval Research Lab, was convinced that Craven was reading way too much into the acoustic data and was chasing ghosts. The only thing that turned east toward the Med, Hardy believed, was Craven's phantom trail. His arguments instilled some doubts within Craven. Besides, it was Hardy's lab that was guiding the Mizar, and Craven needed his support if the ship was going to turn around and start searching to the east. Craven's own relationship with the lab was shaky. As director of the Deep-Submergence Systems Project, he had basically stolen one of the lab's prize possessions, the bathyscaphe Trieste 11, to assist in working out features for Rickover's NR-1 mini-sub.

The officers in charge of torpedo safety at the Ordnance Systems Command soon joined the group of naysayers. They insisted that it was impossible for a hot-running torpedo to detonate inside a submarine. For detonation to occur, the command insisted, a warhead would have to run into an object at top speed and stop moving only as it hit. Then and only then would it go off. The Ordnance Systems commanders were hacked up by the Bureau of Ships. Walter N. "Buck" Dietzen Jr., a top submarine official, was also firmly in doubt. As the debate raged on, none of the men forgot that they were looking for their own dead.

Still at one point, in an effort to lighten things up a bit, Dietzen wagered Craven a bottle of Chivas Regal scotch whiskey that he would turn out to be wrong. Operational commanders were betting with Dietzen. Mizar had already dug up some tantalizing clues on the Norfolk side of Point Oscar. There were three items found that could have fallen from Scorpion: a piece of elbow pipe, what seemed to he a woman's umbrella, and a rope tied in a "monkey's fist," the ballshaped knot that sailors tie at the hitter end of a mooring line to make it easier to catch when it is tossed onto a pier.

There was some argument within the Navy about whether the monkey's fist Mizar found was tied in the U.S. style or in the style favored by the Italian Navy, but the umbrella, the operations officers believed, had to have come from Scorpion's crew. They had made port stops, hadn't they? This could have been someone's souvenir or gift for a woman back home. Months would pass before Navy biologists declared that what looked like an umbrella was actually alive, one of the many odd creatures that live on the ocean floor.

Still, given the Mizar evidence and the strong opinions around him, even Craven began to wonder whether he was wrong, just "smoking opium," as he liked to say. But then again, maybe he was the only one who was right. Craven had no trouble believing either possibility, so he kept digging. He arranged to have a ship drop small explosive charges at Point Oscar. By comparing the acoustic signatures taken at the site with the signals that reached Norfolk, he would be able to figure out once and for all whether an explosion in the area would create echoes-sonic ghosts-as others had contended.

Gordon Hamilton flew into Norfolk from the Canary Islands for the occasion. The two men camped out in a bare cinder-block room in a Norfolk communications station. There they would wait, all day and all night and all the next day, until the calibration charges rang through to shore.

On the first and second tries, the charges were too weak, and none of the acoustic signals made it back to Norfolk at all. By now, Hamilton and Craven were tired of eating cold sandwiches, tired of the bare walls and bare room, tired of sleeping on the blockhouse floor, and more tired of one another. They had exhausted their repertoire of shop talk. Craven had even run out of maxims of the sea.

Craven began doing push-ups. He had already taken to filling in the time left over from his two submarine searches, the design of NR-1, and his running of the SeaLab program by putting himself through the Royal Canadian Air Force exercise program. By now, he could do eighty push-ups at a set. He proved that several times over before the explosives finally signaled through to Norfolk.

They came through with no echoes. And when Craven and Hamilton recalibrated the Scorpion signals with the new data, they realized that not only had Scorpion been traveling east, but she was traveling east even faster than Craven had thought.

Craven was back to his torpedo theory. But he wanted more evidence.

With typical dramatic flair, Craven arranged a reenactment of the tragedy. He needed a submarine simulator, and he needed Lieutenant Commander Robert R. Fountain Jr., the former Scorpion XO who had been detached from the submarine just before she embarked on her final mission.

Fountain was put at the simulator's helm, and a computer was programmed to factor in the orders he gave as the simulator reenacted various possible causes of Scorpion's loss. Ten different scenarios were tested this way, and ten failed to create a match with the acoustic evidence. Then, Craven's team asked Fountain to try one last time. They said nothing about a possible torpedo explosion, they simply told Fountain that he was heading home at 18 knots, leaving it to him to choose a depth. Craven then asked him to test his torpedoes. The team waited ten or fifteen minutes, giving Fountain a chance to stand calm. Then they rang an alert. "Hot-running torpedo in the torpedo room."

Without missing a beat, without waiting, without asking questions, Fountain ordered, "Right full rudder."

There it was. The turn that Craven believed had been executed on Scorpion.

When Fountain's simulated turn was almost complete-maybe half a minute or so after he had called out "right full rudder"-the staff called into the simulator: "Explosion in the forward torpedo room."

The same information was fed into the computer, which began to register extensive flooding in the submarine.

Fountain answered with a seemingly endless stream of orders: blow ballast, initiate watertight security, speed the boat. He did everything a submarine captain should do. Still, the mythical submarine continued to flood, and it continued to head toward the bottom. Exactly 90 seconds after Craven announced the explosion, it passed 2,000 feetpassed right through collapse depth-and the computer registered an implosion. Someone on the staff announced the event with one word: "bang."

The simulated implosion occurred just one second off the 91-second time recorded between Scorpion's explosion at Point Oscar and the first implosion under crushing ocean pressures.

Chills shot through Craven when he saw the results. By now, he and several others attending this test were nearly certain they had replicated Scorpion's loss. No one told that to Fountain. No one told him he had just possibly enacted the circumstances that led to the deaths of the men he had once helped to command. Maybe nobody had to tell him. He left the simulator without asking any questions, without saying a word.

Craven's compassion for Fountain and for the crew of Scorpion couldn't squelch the exuberance he felt. As a detective, he had come up with two new important pieces of evidence, and he now raced them to Admirals Schade and Bernard A. Clarey, the vice chief of Naval Operations. By now, even they were becoming intrigued by Craven's detective work, but they remained unconvinced. As did the Ordnance Systems Command, which continued to insist that there was no way a torpedo could explode on board a submarine.

Nobody was ready to face the specter that the Navy itself was responsible for the deaths of those ninety-nine men. Craven understood their reluctance, understood how difficult it was for the admirals to believe they might have been somehow responsible for a mistake that had caused the loss of so many people's lives. Both admirals had lived through a time when death aboard submarines was common. Both were veterans of the World War II diesel boats, but then death had come from the enemy and not from their own boats. Schade was probably the more hard-nosed of the two, and no wonder. As a young executive officer on the USS Growler (SS-215), Schade had gotten his first taste of command while his skipper lay wounded on the bridge of the sub. Commander Howard Gilmore shouted a final order to young Schade, an order to take Growler into a desperation (live to escape a Japanese gunboat while leaving Gilmore wounded up top. Schade did as he was directed.

Despite the admirals' reluctance, Craven wasn't about to give up, not now that he was convinced he had enough information to find Scorpion and prove what killed her. He began to mathematically construct a map of the ocean bottom, using Bayes' theorem of subjective probability, the same algebraic formula he had employed during the search for the H-bomb.

Few of the officers involved in the search for Scorpion had taken much note of Palornares. And by the time Craven was finished explaining that he was going to use a system of Las Vegas-style bets to factor the value of a hunch into his data, some of the operational commanders were convinced that he had gone completely over the edge. To them, it sounded like he was talking about ESP. Craven once again tried to explain that Bayes seemed to draw on the knowledge that even experts are not always consciously aware they have. The commanders remained highly skeptical.

Still, Craven pushed on, asking a group of submarine and salvage experts to bet on the probability of each of the different scenarios being considered to explain Scorpion's loss. To keep the process interesting, and in line with previous wagers, the men bet bottles of Chivas Regal.

Scorpion could have glided down to the ocean bottom at speeds between 30 and 60 knots. His submarine experts bet that Scorpion had glided downward at between 40 and 45 knots.

Next, the experts were asked to bet on whether they believed Scorpion was trying to shut down a hot-running torpedo and was therefore traveling east. About 60 percent of the bets favored the torpedo theory. Craven, it seems, was winning some converts.

In a third round of betting, the experts picked a glide path. At the most, Scorpion could have moved 7 feet forward for every foot she descended; at the least, she could have nosedived straight down. The bets favored a glide path of about 3 or 4 feet forward for every foot down. That meant Scorpion would have traveled 6 to 8 miles after the first explosion.

By the time the bets were finished and Craven sat down to draw a probability map, the calculations had become so complicated that he had to rehire the group of mathematicians who had helped him with the H-bomb. They concluded that Scorpion was east of Point Oscar, 400 miles from the Azores, on the edge of the Sargasso Sea.

Years later, the mathematicians would write a book based on their work with Craven, entitled Theory of Optimal Search. The U.S. Coast Guard would adopt the method for search and rescue, and the Navy would use Craven's interpretation of Bayes to help Egypt clear sunken ordnance from the Suez Canal. But in the Scorpion search, naval officers just shook their heads at Craven's acoustic evidence and his probability map. The scientist may have been convinced that Scorpion lay further east, but the Mizar had found the three scraps of debris to the west, and that's where the Navy wanted to keep searching.

Weeks passed. Craven waited, trading messages nearly every night with Buchanan. By late August, nothing new had been found and the jubilation within the Navy that had accompanied the Mizar's find of the supposed umbrella and the monkey's fist knot diminished. By September, all of the likely spots between Point Oscar and Norfolk were almost ruled out. By October, the weather was getting so bad that the Navy decided it would end the search by the end of the month.

But Mizar still hadn't really searched east. And it had never searched the site Craven had pinpointed. By now, Buchanan was sporting a fully grown Vandyke beard and was willing to point Mizar east one last time.

Almost as soon as Mizar passed east of Point Oscar, its long-range sonar registered iron, and lots of it. Mizar steamed ahead full speed, right past Craven's point of highest probability, and then lowered its cameras for another look. All it found was iron ore-filled rock.

That was it. The end. Schade and Clarey had had about all the disappointment they could take. The decision was made. It was time to give up. Time to call Buck Buchanan and Mizar home.

Buchanan, pugnacious and stubborn as ever, refused to accept their decision. He flashed a message to Craven.

"Can't you get the Navy to let us stay out another month, or a week or two weeks? Tell them I need to calibrate the area for future operations."

Craven knew that there was nothing left to "calibrate." But Craven also knew that if Buck Buchanan wanted to stay out, it could mean only one thing. The oceanographer was going to take Mizar to the spot Craven and his team had pinpointed. Craven went to the admirals and began mixing his rapid-fire logic with pleas. By the time he was finished, he had won two more weeks.

Exactly one week later, Craven received a one-line missive from the survey ship: "Buchanan shaved his beard."

Craven didn't need any translation. Scorpion had been found. It was October 29, almost five months to the day that she had been declared missing.

Mizar found Scorpion within 220 yards, one-eighth of a mile, of where Craven, his mathematicians, and a group of experts betting for bottles of scotch had said she would be. The sub was 11,000 feet underwater.

Dangling cameras, Mizar took photographs showing Scorpion halfburied in silt and sand and separated into two pieces that were barely held together by a small hinge of metal. The forward part of the engine room had imploded and, in a fraction of a second, collapsed like a telescope into the auxiliary machine space.

The propeller and the propeller shaft were separated from the hull altogether. So was the submarine's sail. Lying near the submarine was Scorpion's sextant-an age-old symbol of navigation. No navigator, officer, or crewman was anywhere in sight. It was impossible to see inside the submarine or even the outside in much detail. Although Mizar's cameras dangled only between 10 and 50 feet over Scorpion, the overhead pictures looked as though they'd been photographed through a deep fog.

A court of inquiry looking into the disaster was made up of seven naval officers and chaired by retired Vice Admiral Bernard L. "Count" Austin, who had also led the inquiry into Thresher's sinking. In a January 1969 press release, the Navy told the public that the court of inquiry, after a six-month investigation, had concluded that the Scorpion disaster remained a mystery, that the cause could not be "ascertained from any evidence now available," and that "no incontrovertible proof of the exact cause" could be found.

Indeed, the Navy publicly appeared to rule out a torpedo disaster of any sort, saying, "Procedures used in handling ordnance on board were consistent with established safety precautions." Then it went on to boast that testimony "also established a long history of safety in submarine torpedoes."

Technically, the Navy told the truth, but in such a limited form that the result was a massive evasion that bordered on an outright lie. In fact, when the court's more detailed findings were finally released in 1993, they showed that it had concluded that the top three probable causes of Scorpion's loss all involved torpedo accidents.

Leading the list was Craven's theory that there was a hot-running torpedo on board Scorpion, perhaps caused when crewmen tested torpedoes in preparation for their arrival home. But then the court veered from Craven's theory that the torpedo exploded on hoard Scorpion. Instead, it speculated, "Acting on impulse, and perhaps influenced by successful ejection of a Mark 37 exercise shot which was running hot in the tube in December 1967, the torpedo was released from the tube, became fully armed, and sought its nearest target, SCORPION."

The court acknowledged there was no evidence of an external torpedo hit but reasoned that there was also a lack of any visible torpedo-room debris near Scorpion, which would prove that the explosion occurred inside the sub.

Former submarine torpedomen say it is almost unthinkable that Scorpion's crew would have panicked and jettisoned a warshot torpedo. The 1967 incident involved a torpedo meant for practice shots that carried only a dummy warhead and no live explosives.

The court seems to have crafted a compromise for its classified findings. Citing Craven and his acoustic evidence, the court concluded that a torpedo was at fault. But the contention of an outside explosion seemed patterned on the Ordnance Systems Command's insistence that it was impossible for a hot run to lead to an onboard explosion.

Also in the report was a list of possible submarine accidents of all types, prepared by the Bureau of Ships. The list included gas leaks, broken hydraulic lines, fires, and more. But only one item on the list showed catastrophic results: a weapons accident. That, the bureau said, would result in "loss of ship."

In mid-1969 the Navy directed a top-secret effort to try to examine more closely the submarine's wreckage and unravel the mystery. It was most interested in the torpedo room and the torpedo doors. The Trieste II was sent down for a closer look. 'The first dive was made on July 16, only days before the Apollo 11 astronauts made the first manned landing on the moon.

"My God, what a crazy world we live in," Craven muttered to himself as he stood on the floating dock that had launched the Trieste. "We think we're doing a technological feat which is every bit as difficult and every bit as meaningful to humans as this man-on-the-moon thing, and we're the only ones who will get the chance to savor this operation."

Trieste made nine dives that year. Watching the first from monitors aboard the floating dry dock were Craven and Captain Harry Jackson, the engineer who had helped test Thresher and never stopped being haunted by his near-miss on that sub. They could see that there was no evidence of attack, and no evidence of an external torpedo hit. But there was also no conclusive evidence to show just what had sent Scorpion to the ocean bottom.

Craven would always struggle with the last piece of the puzzle. He was nearly certain a torpedo blew up inside Scorpion. But how)

It all seemed to end there, with the question left unanswered, with the families of the Scorpion men left to wander in nightmares of explosions and phantom battles and disbelief.

"All we ever wanted was an explanation," said Barbara Baar Gillum, who lost her twenty-one-year-old brother, Joseph Anthony Baar Jr. "After the disaster everything was covered up."

The Scorpion disaster quickly faded from the greater public conscience, which was already being battered by nightly images of bullets flying, soldiers bleeding, and a seemingly endless line of body bags in Vietnam. The Scorpion families might have been left alone to forever struggle with their own investigations had the Navy not decided to mark the grisly quarter-century anniversary of Scorpion's loss by releasing the court of inquiry's conclusions and videotapes of her sunken shell.

By then, Craven was sixty-nine years old and long gone from the Navy. He was, instead, intensely involved in developing a new form of agriculture in Hawaii. The Chicago Tribune printed a story about the documents and his role in using his torpedo theory to find Scorpion. It was only when that story ran that Craven was handed what he is convinced is the last piece of the puzzle.

It all played out in a scene reminiscent of the final chapter of a detective novel. The Tribune article reached the desk of Charles M. Thorne, who had been technical director of the Weapons Quality Engineering Center at the Naval Torpedo Station in Keyport, Washington. Seeing Craven's name in print, Thorne picked up the telephone and dialed.

The two men had never met. Neither had known anything about the other during the long Scorpion search and all the years after. Still, they had much in common. Thorne, too, had long had reason to fear that a torpedo had been the cause of Scorpion's death. Back in the summer of 1968, he had been a top engineer in the Keyport lab responsible for testing torpedoes and their components. He worked there for twenty-five years, and by the time he called Craven, lie had been retired for twelve years. All that time he had held information about Scorpion that he felt barred by classification rules from telling anyone. Now the engineer reached out to the scientist.

Thorne asked Craven whether he had seen a classified alert that had been mailed in mid-May 1968 to the department that had been renamed the Naval Ordnance Command. The letter described a test failure of an MK-46 battery that was designed to power the Mark 37 torpedo, a fastmoving warshot that had been deemed the primary weapon for use against Soviet subs. He was referring to an alert that the testing lab had sent to Rear Admiral Arthur Gralla who headed that command. Then Thorne went on to describe its contents. He knew them well since he had written the alert himself, although it had been reviewed and signed by Captain James L. Hunnicutt, the CO of the station and a decorated World War II submarine skipper, who has since died.

In the alert, the lab reported that a torpedo battery had exploded in flames during a vibration test because a tiny foil diaphragm, a part worth pennies, had failed. As Thorne related this to Craven, the news seemed to parallel the discovery that the failure of inexpensive rubber 0-rings had caused the space shuttle Challenger to explode. About a yard long and 17 inches wide, the batteries on Mark 37s were bolted within about an inch of the torpedo warheads. And each warhead carried 330 pounds of HBX explosive.

The lab's alert had recommended that all batteries from that production lot "be withdrawn from service at the earliest opportunity," and it said that sufficient heat was generated in the test sample "to risk warhead cook-off and loss of a submarine."

This alert was the strongest of any ever issued by the testing lab. It was the only time in the lab's twelve years of operation that it had ever warned of the possibility of a failure that could have life-threatening consequences. It was because the engineers were so deeply concerned, that they had their commanding officer sign the alert. They wanted the added emphasis.

Scorpion was carrying fourteen Mark 37 torpedoes, and she was lost days after the letter was sent. Horrified at the possible connection, lab engineers specifically asked the Naval Ordnance Command about the torpedoes Scorpion was carrying. The Navy keeps records and serial numbers of all torpedo components and where they are issued. One of the lab's engineers remembers being told verbally that one of the batteries powering a torpedo on Scorpion did indeed come from the same production lot as the torpedo battery that had exploded at Keyport. (Other former engineers there said they did not remember hearing this.)

Over the past several years, one of the engineers made requests for the battery records under the Freedom of Information Act, hoping they might answer that question once and for all. But the answer came back twice that no such records could be found.

Still, Thorne believes that a warhead cook-off initiated by a battery fire was the likely cause of Scorpion's loss, and he became all the more convinced of that when he read that Craven had concluded that Scorpion had suffered an internal torpedo explosion. He was stunned that Craven had never seen the lab's alert. Thorne had always assumed that the secret missive had been shown to the people involved in trying to make sense of the Scorpion disaster. Now it seemed that vital information had been withheld from Craven and the court of inquiry.

Thorne asked Craven for a copy of the videocassette of Scorpion's wreckage and the court of inquiry report. After viewing those, he wrote to Craven with his analysis.

"I have agonized for years over what more we might have done to have averted that tragedy," Thorne said in the letter. "The people that did the testing, the workman and other engineers, we all wondered. We asked questions."

Thorne then went on to tell Craven that his worst fears were borne out by Scorpion's wreckage. The videotape clearly shows that the upper hatch covers of the torpedo-loading hatch and the escape trunk hatch are gone. Both hatches lead into the torpedo room. Both, Thorne wrote, could have been blown away as a result of a violent explosion inside the torpedo room, and that could have resulted in massive and uncontrollable flooding of the submarine.

The battery failure that prompted Thorne to pen the alert was easily the most severe the lab had ever experienced. The test failure happened on a Saturday afternoon as three engineers-John Holman, John Grobler, and Robert Trieschel-subjected one of the 250-pound batteries to strong vibration. They had just walked out of the room where the tests were conducted when there was a tremendous explosion, strong enough to rattle the 2-inch solid wooden door. Holman threw the door open and ran in. The mechanism meant to shake and vibrate the battery was all but obscured by blue-green flames shooting 10 feet to the ceiling.

"Fire!" he yelled as he grabbed an extinguisher. The room began to fill with black smoke and flames. Two technicians were missing. Holman got on his hands and knees and began looking for them as fire trucks screamed up to the laboratory.

Chemical extinguishers failed to put the fire out. The men slipped rags over their faces and began unbolting the still-burning battery from the shaker. The battery exploded a second time, drenching them in the potassium hydroxide solution that served as the batteries' electrolyte. Shrapnel was embedded in the ceiling and walls.

The engineers took the burning battery out of the building. Its 16gauge steel case was peeled open like foil and the silver plating on the battery was partially melted. As soon as they could they raced hack inside to drench themselves in the laboratory's emergency showers. Then the three lab employees and three firemen were rushed to the hospital to be treated for smoke inhalation and chemical burns. The lab called for the recall within two or three days of the incident.

A similar battery failure on Scorpion could have been enough to cause a warhead to explode, but the alert was sent too late to save the sub and crew. The phrase "withdrawn from service at the earliest opportunity" was usually construed to mean that a recall should be conducted as each boat reached port. When the recommendation reached the ordnance command, Scorpion was either already lost or still en route to Norfolk, where the recall would have been implemented.

Had the alert been made available to Craven right after the accident, months might have been shaved off the search for Scorpion. But instead of sharing the information, the ordnance command continued to insist that such an explosion was impossible. Had the court of inquiry been given the information, it might have done much more to solve the mystery. But the court in its report clearly relied heavily on Naval Ordnance's statements about the impossibility of an onboard detonation.

That the lab's alert made it to the ordnance command seems certain. It was specially coded to be routed from the mail room straight to the commander's desk. In addition, some weeks after Scorpion's wreckage was found, a representative of the ordnance command showed up at the Keyport lab, called Thorne into a private office in another building, and castigated him for including warnings of warhead cook-off and loss of a submarine in the alert.

The command had good reason to he deeply concerned by what Thorne wrote-Naval Ordnance had created the potential for catastrophe by bypassing its own safety procedures. In its effort to keep up with the sub fleet's demands for the torpedoes, it had rushed the weapons into production. The fleet desperately needed torpedoes that could go fast enough to catch the new classes of Soviet nuclear subs, but manufacturers were having a terrible time building components that could pass safety tests. There was such a backlog caused by repeated battery failures that the Keyport lab was at least two months behind in its quality assurance tests. Rather than slow down production, the ordnance command had been issuing torpedoes with components from lots that had never been safety tested at all. That was a clear violation of regulations that required that three samples from every lot of one hundred batteries be tested before any battery from the lot was issued to the fleet. The samples were supposed to be subjected to two or three weeks of tests measuring their resilience to shock, heat, vibration, and any other condition that could be expected to occur on a submarine. Only after the samples passed muster was the ordnance command supposed to allow any of the components on hoard any submarine.

Two companies had originally contracted to manufacture the batteries and suffered so many failures that the Navy brought in a third company to try to make up for the production problems. That company never managed to produce any batteries that passed quality assurance tests, but because of the shortages, the contractor was allowed to ship as many as 250 of its batteries to the fleets. It was one of the third company's batteries that exploded in the lab.

All three companies were having problems because the basic design of the batteries was dangerously flawed. Engineers had warned of this from the first failures back in 1966, well over a year before the catastrophic explosion that resulted in that last and strongest alert. Throughout that time, they had said the batteries had no margin of safety and recommended a redesign. The ordnance command was unwilling to do that.

The problem existed in how the batteries were activated. A sliver of foil that governed the flow of electrolyte into the power cells was etched to be only one seven-thousandth of an inch thick. That was because it was supposed to break with pressure, allowing the battery to power the torpedo's propulsion motor when the weapon was activated.

In a typical hot run, the kind Craven imagined had occurred, a torpedo receives an inadvertent start-up charge, fully activating the battery and also turning on the motor. That condition is easily detected as a spinning propeller on the torpedo alerts crews to the need for an immediate 180-degree turn.

The kind of failure experienced in the lab was far more insidious because there never would have been enough power in the battery to turn on a motor, or cause the torpedo propeller to spin. Instead, what happened to the batteries in the testing lab, and what caused the explosion of the test sample, was more difficult to detect. The lab discovered that when the batteries were subjected to vibrations, the electrolyte was pushed against that thin diaphragm with enough force to only partially rupture the foil. This allowed just enough electrolyte to slowly leak into the batteries' power cells to cause them to begin to spark and overheat. It was precisely because the diaphragms broke so easily, and the overheating condition could remain hidden until a fire or explosion that the lab determined that the design had almost no margin of safety.

During the vibration test that Thorne described in his letter, there was no hint of a problem until the battery exploded into flames. If the same thing had happened on a submarine, it was entirely possible that no one would have noticed anything was wrong unless they smelled insulation burning or touched the torpedo and felt heat. By then, the battery could have been only minutes away from exploding.

"If the hot torpedo shell is not discovered until the paint begins to blister or burn," Thorne wrote to Craven, "there may not be time to move the torpedo from a stowage rack and load it into a tube for jettison before warhead cook-off."

Such a torpedo accident could have occurred in any of the fourteen Mark 37 torpedoes on Scorpion, or, for that matter, on any of the submarines equipped with those torpedoes. Carrying a battery from a defective lot would have heightened the risk, but the risk existed in the design nonetheless. The diaphragm on one or several of the batteries could have ruptured as the weapons lay in torpedo tubes or in their racks. Men didn't have to be testing them or handling them in any way. Shipboard vibration could have been enough.

Scorpion may have been more at risk than any other boat. The vibration tests in the lab that led to the explosion were supposed to emulate the usual vibrations that could be expected on board a sub and when transporting torpedoes to a submarine-vibrations that were far less severe than the shaking Scorpion had suffered in the 1967 incident when she was sent corkscrewing through the water. If Scorpion suffered a repeat of that incident, any one of the batteries might have failed. Weapons engineers say that judging from the crew's descriptions, the vibrations created during the 1967 mishap far exceeded military specifications for battery safety. Scorpion, in fact, had been failed twice. The potential for a repeat of the vibration incident existed because she had never received the overhaul she was due for. And she had been sent to sea carrying weapons Naval Ordnance knew suffered from a critical defect.

Nevertheless, Naval Ordnance has never acknowledged that Scorpion could have been at risk for torpedo detonation, or even that Scorpion's torpedoes were powered by batteries with a defective design. In fact, after the sub was lost, the Naval Underwater Systems Center in Newport, Rhode Island, vigorously argued against the test lab's conclusions.

Naval Ordnance also withheld the information about the flawed battery design even after another torpedo battery began to heat up on board a submarine in the western Pacific months after Scorpion was lost. The crew of the second boat reported that their torpedo battery reached temperatures so high they had to spray the torpedo constantly with water to cool it. The water turned to steam. They had no choice but to continue spraying until the torpedo could be loaded into a tube and jettisoned.

Finally, about a year after Scorpion went down, Naval Ordnance did order a new battery design. The new system replaced the thin foil diaphragm with two stronger ones. In this new design, both diaphragms could be broken only when they were mechanically punctured with a cookie-cutter device, eliminating the danger that shipboard vibration could lead to a battery fire and set off an explosion.

Any written record of the Keyport lab's alert, and the alert itself, appear to be missing. There should be copies of the engineers' alert at the main administrative offices of the Naval Undersea Warfare Engineering Center, formerly the Naval Torpedo Station at Keyport, and at the ordnance command, but another recent request made under the Freedom of Information Act for a copy of the alert at both sites came back with the answer that there was no record of it, or even of its destruction-something that also should have been logged if the alert was removed from the files.

Still, after hearing Thorne's story, Craven and some sub commanders and weapons experts are looking back at the Scorpion disaster. Craven is angry that the ordnance command failed to disclose the fires itself. "The public and the press and a lot of other people feel an organization is engaged in a cover-up when they so stoutly deny this kind of thing," he says.

Taking the new evidence into account, Craven theorizes that the cry that led Scorpion's skipper to make that final turn could have been "Hot torpedo" instead of "Hot-running torpedo."

Chester M. Mack, who skippered Lapon when she searched for the downed Scorpion, insists that no captain would ever take the time to ask for more information before executing an immediate 180degree turn. "`Hot torpedo'-there's only one thing it could mean. The damn thing is running in the torpedo room," Mack says.

After years of being told over and over again that he was wrong, that it was impossible for a torpedo to detonate on board Scorpion, Craven is now convinced that he has what is very possibly the final piece to the mystery he set out to solve more than a quarter of a century ago.

It is a mystery that continues to unfold. In 1998, almost five years after the Navy released the court of inquiry report, almost five years after Craven first spoke to Thorne, the Navy released a 1970 report by another technical advisory group-this one convened shortly after Craven retired from the Navy. It was set up to review pictures and data collected by Trieste during her nine dives down to Scorpion. The report had been completed only a year after the court of inquiry had finished its work, but the document had been held back from the public, from the Scorpion families, even after the court of inquiry reportwas released, and even though it specifically discounts many of the conclusions reached by the court of inquiry.[6]

This advisory group throws out the court of inquiry's conclusion that Scorpion was likely felled by the external explosion of an ejected torpedo that turned back on her. It also tosses out the possibility that Scorpion was destroyed by an internal torpedo explosion. Still, the authors of the report clearly did not have the information about the torpedo battery failures in the Keyport lab. Indeed, Craven, Thorne, and some submarine captains believe that much of the evidence used to refute the torpedo theory actually supports it.

The group's report also makes no attempt at all to explain why Scorpion was found just where Craven said she would be if her captain had turned to battle what he believed was a hot running torpedo. Instead the Navy had gone back to debating the meaning of the acoustic trail Craven and his team followed to Scorpion's grave. Although the details of the Scorpion search and the crucial role Craven played were retold in a second report the Navy declassified at the same time, the 1970 analysis instead relies heavily on Naval Ordnance assurances that Scorpion could not have suffered a torpedo accident. The group, in short, based its conclusions on statements made by the same Navy department that was withholding critical information that had been kept from the court of inquiry and the search teams.

By the time the second report was written, Naval Ordnance's argument had changed. Instead of insisting that a torpedo could never explode on board a sub, Naval Ordnance had focused on the visual evidence collected by Trieste: from the outside, the hull of the torpedo room looked basically intact, while the ship's battery well was largely destroyed. The Trieste pictures did show that all three hatches leading from the torpedo room through the pressure hull-the forward escape trunk access, the escape trunk hatch, and torpedo loading hatch-were all dislodged. (Trieste could not get cameras inside the torpedo room to check for internal damage.[7]

As the report states:

Because of that argument, the technical advisory group also discounted the simulated reenactment of Scorpion's loss staged by Craven and Fountain.

Still, Craven and several munitions experts say that the ordnance command's argument is deeply flawed and that if the command had told investigators about the failures of the batteries on the Mark 37 torpedos, the analysis would have been changed considerably.

The kind of external hull damage the ordnance command insisted would have followed a torpedo explosion, weapons experts say, may not have followed a torpedo explosion caused by a fire. Rather, the damage the command was describing would likely occur during, a fulltriggered explosion, an explosion in which a torpedo is set off just the way it is designed to be set off, with the power of 330 pounds of HBX explosives unleashed all at once in a massive and directed forward thrust. That kind of detonation would, the experts agree, likely detonate other torpedoes. And a multiple detonation would, as naval ordnance said, probably crack or at least buckle the submarine's hull in a way that would be visible from the outside.

But the fact that there was a possibility of a torpedo detonation caused by fire in the torpedo battery profoundly changes the equation. A weapons explosion caused by fire will almost certainly not be the same power or predictable shape of a neatly triggered explosion. In fact, there is no way to predict the size, the shape, the properties of a blast caused by fire. Such blasts just don't follow the usual rules. It is, weapons experts say, entirely possible, even probable, that a torpedo warhead set off by a fire, could go up in what is known as a low-order detonation.

A low-order detonation could be strong enough to kill anyone nearby and could be strong enough to blow the hatches in a torpedo room without setting off other torpedos, especially if those other torpedos were not laying directly against the torpedo that exploded. Submarines often went out without their torpedo racks full. (That's why men often bunked in the torpedo rooms. Any torpedo rack without a torpedo laying on top of it made a reasonable place for a mattress.) One low-order detonation, without subsequent detonations of other torpedoes nearby, could easily occur without the kind of external hull damage the men diving on Trieste had been told to look for. The Navy itself acknowledged that to be entirely possible in the 1969 court of inquiry report into Scorpion's loss. The report cites a 1960 incident on the USS Sargo (SSN-583) in which an oxygen fire in the after-torpedo room spread and caused two Mark 37 torpedo warheads to detonate "low-order." The report states: "The pressure hull of the Sargo was not ruptured." Sargo was pier-side and on the surface at Pearl Harbor at the time.

Indeed, the fact that Scorpion's torpedo room is intact raises the probability that she was lost to a torpedo casualty, say sub commanders and Craven. Scorpion's torpedo room did not implode, which makes it very likely that it was flooded before she fell to crush depth. Since a flooded room is exposed to equal ocean pressures both inside and out, it does not collapse at crush depth and it does not implode. It is left intact.

The technical advisory group does say that the torpedo room hatches "probably failed" when "pressure inside the torpedo room increased," or when the bulkhead leading to the operations compartment gave way. The idea seems to be that the hatches were forced open by the violent implosions going on right outside the torpedo room. The advisory group offers no theory about why the torpedo room would have lost only its hatches, when the group's own experts say compartments just outside were completely destroyed in the same violent instant. Trieste's pictures show that Scorpion's operations compartment, which is right next to the torpedo room, is squashed flat and that just beyond operations, Scorpion's tail has telescoped completely into her auxilliary machine space.

The Trieste photos also show that the massive battery that powered Scorpion was torn apart. The advisory group theorizes that this is what destroyed Scorpion-echoing the theory about what destroyed the Soviet Golf. The sub's battery could have exploded as it was being charged if ventilators failed and the concentration of highly combustible hydrogen gas was allowed to build up. The battery, however, could also have been torn apart by the same forces that destroyed the rest of the submarine.

Admiral Schade, Fountain, and others have guessed that perhaps Scorpion's trash disposal unit failed, allowing tons of seawater to pour into the sub and into the battery well. Seawater can cause a battery to emit a number of gases, including hydrogen gas. The trash disposal theory, however, is based both on the lack of any other apparent reason and the fact that a trash disposal unit failed on the USS Shark (SSN — 591), Scorpion's sister ship. (Shark survived.) Again, many of these observers had discounted that the first flooding could have been caused by a torpedo, because they were told that there was no way for a torpedo to explode on board Scorpion, certainly not without blowing up the rest of the weapons as well.

"I think we are all guessing," says Ross E. Saxon, who went down on Trieste and took some of the photographs studied by the technical advisory group. "We who were out there, who dove down on the thing, are guessing."[8]

Offered the new information about the flawed torpedo batteries, some people close to the investigation who had discounted a torpedo explosion after the 1970 report now say that a torpedo explosion has to be put back on the list of possible causes of Scorpion's loss.

"If a room blows up and there was a hand grenade there, but then I call up and say I took the hand grenade out of the room, you would discount the hand grenade," says an active-duty Navy official familiar with the case through its latest developments. "If I didn't tell you there were two hand grenades though, if someone was being less than fully truthful, providing less than all of the information, maybe there would be cause to go back and look at it again. Based on the information on file now, the two most likely causes are a ship's battery explosion and a weapons cook-off. Based on the information we had, I'd say battery explosion. Now there is a good way for a weapon to cook-off. Any information about specific engineering problems in a weapon ought to be tossed into the fore, ought to be discussed."

The officer, as well as Craven and many others agree that there needs to be more investigation, perhaps another effort to take a look inside Scorpion's torpedo room. For now, Craven remains convinced that a torpedo was the most likely cause of the sub's loss. He is not alone. In June 1998, Craven stood before a throng of Navy officers when he became the first man to be given the distinguished civilian service award by the Naval Submarine League for his work on Scorpion, Polaris, and other projects. When the ceremony was over, an officer approached him. His voice lowered so it wouldn't carry through the crowded room, the officer began talking about Scorpion and told Craven that he had been convinced for years that she was lost after a torpedo accident.

Without knowing about the alert sent from Keyport, without knowing that there were known problems with the batteries powering the Mark 37 torpedos, the officer told Craven: "I know it was a torpedo because I had a torpedo battery cook-off on me."