The War Below Read online

  The value of this newfound intelligence crystallized in the spring of 1930 when almost the entire Japanese navy put to sea for fleet maneuvers. In a break from previous protocol, Japan shrouded the operation in secrecy. An American listening post on Guam, however, intercepted the fleet’s radio communications. The decrypts stunned American analysts. The five-week operation had revealed Japan’s war plan against the United States, a plan that called for the capture of Guam and the Philippines followed by the destruction of the Pacific Fleet. Japan’s simulation had even prepared for the possibility of an American carrier raid against Tokyo, just like the one Jimmy Doolittle would lead in April 1942. But the decrypts revealed yet another shock. Japanese intelligence had successfully mapped out America’s updated contingency plan for a possible conflict in the Pacific—dubbed War Plan Orange—that strategists in the nation’s military colleges had rehearsed that same year.

  The intelligence haul America bagged from the 1930 fleet maneuvers led to an even more ambitious experiment three years later when the Japanese navy again put to sea for exercises. Rather than collect communications and forward them for decryption—a process that could take weeks—the Navy wanted to test what could be learned short of cryptanalysis from a real-time examination of message traffic. The test demanded analysts look at message lengths, addresses, call signs, and the frequencies used. The 1933 test proved a great success. Brute grunt work that involved charting the time, date, and sender of each transmission—among other details—as well card indexing every call sign, helped analysts map the entire organization of Japan’s naval forces. More than just illuminating Japan’s war plans, the test confirmed the value of radio traffic analysis, a technique that would prove invaluable in the years to come as Japan’s cryptosystems grew more sophisticated.

  Japan changed its naval codes several times in the decade leading up to the war, replacing the so-called Red Code with Blue then Black before finally settling in 1939 on JN-25, the fleet operational code. This complex new cryptosystem would serve as the workhorse of the Japanese navy, handling as much as 70 percent of all message traffic. Compared to previous naval codes, JN-25 represented a radical departure. Just to operate this cryptologic nightmare required several books, including an instruction manual. An operator preparing to send an encrypted message would first consult a codebook filled with some 30,000 five-digit numbers, representing everything from place names and numbers to Japanese Kana particles. This allowed the operator to convert each word or phrase of a plaintext message into a five-digit numeric code. The Japanese designed the code with a safety mechanism to guard against garbles: the sum of the five numbers in each group was divisible by three.

  The operator would next consult a 300-page additive book comprised of random numbers: 100 five-digit groups per page. The operator would add these numbers to the original code groups, but with a catch: no numbers could be carried over. This erroneous arithmetic served to further encipher the original code. Keys hidden inside each message instructed operators on the receiving end the precise page and line to turn to in the additive book to subtract the numbers and reveal the original code. Operators then only needed to look up the corresponding plaintext meaning. The cryptosystem’s use of varying additives meant that the one word could be encoded thousands of different ways. American naval cryptanalysts had never seen such a sophisticated system. Japan further exacerbated America’s code-breaking challenge, rolling out a new and more expansive version of the code eighteen months later and by issuing eight new additive books before the attack on Pearl Harbor.

  America failed to grasp the importance of this new cryptosystem, wasting thousands of hours trying to crack a seldom used flag officers’ code. The pursuit of that cryptologic fool’s gold along with pressure to read diplomatic messages and a shortage of code breakers sidelined efforts to break the sophisticated system. Largely unable to read messages transmitted in the complex naval code, America depended on traffic analysis to monitor Japan on the eve of the war. But those efforts soon came under attack when Japan changed its naval call signs on November 1, 1941. In what analysts viewed as an ominous sign, Japan changed them again one month later. Traffic analysis ground to a halt for a couple of days until operatives figured out 200 of the most frequently used call signs. Analysts made an alarming observation: none of the recovered radio signs were for Japanese aircraft carriers. No known carrier since November 25, analysts realized, had either sent or been addressed a message.

  Where were Japan’s flattops?

  Pacific Fleet intelligence officer Lieutenant Commander Edwin Layton had struggled to answer that question when he prepared the December 2 intelligence sheet. But the reality proved far worse than Layton or his colleagues could imagine. The Japanese transmitted bogus messages to disguise the departures of almost three dozen carriers, cruisers, and destroyers, most slipping port one by one so as not to arouse suspicion. This massive task force built around six flattops had secretly rendezvoused in a snowy harbor in the Kuril Islands before departing at dawn on November 26 to steam east toward Hawaii. America had actually lost track of Japan’s flattops as far back as mid-November. Tasked to keep Pacific Fleet commander Admiral Husband Kimmel apprised of when and where a potential enemy might strike, Layton could only speculate as to the location of Japan’s main aircraft carriers, writing simply: “Homeland waters?”

  “What?” Kimmel had barked when he read the intelligence sheet. “You don’t know where the carriers are?”

  “No, sir,” Layton had answered, pointing out that he had only guessed homeland waters.

  “You mean to say that you, the intelligence officer, don’t know where the carriers are?” the admiral pressed.

  “No, sir,” Layton replied. “I don’t.”

  “You mean they could be coming around Diamond Head, and you wouldn’t know it?”

  “Yes, sir,” the intelligence officer answered, “but I hope they’d have been sighted before now.”

  The irony of that conversation would haunt Layton.

  Cryptanalysts dropped the flag officers’ code after the Pearl Harbor attack and focused on cracking JN-25, prying the sophisticated system open in time to decipher details of Japan’s plans to capture Port Moresby and Midway. This intelligence tap spilled over to the undersea service, thanks in part to Lieutenant Commander Jasper Holmes. A Naval Academy graduate and former skipper of the submarine S-30, Holmes had waged a battle against spinal arthritis that forced him to retire in 1936, ending thirteen years of commissioned service. The Navy recalled him on the eve of the war, but this time as an intelligence officer based at Pearl Harbor. The intelligence community fiercely guarded information derived from cryptology—dubbed Ultra—for fear that leaks or unexpected successes might arouse Japanese suspicions and trigger a code change. But Holmes argued that the submarine force could benefit from America’s cryptologic success. “Intelligence, like money, may be secure when it is unused and locked up in a safe,” he would write. “It yields no dividends until it is invested.”

  The veteran submariner succeeded in convincing the head of Station Hypo, Oahu’s communications intelligence unit. Armed with decrypted information on aircraft carriers or enemy task force movements, Holmes would pen the longitude and latitude on his palm and walk it over to Vice Admiral Lockwood’s chief of staff, careful to scrub his hands of the evidence afterward. Despite Holmes’s best intentions, the effort netted little of value, largely because the enemy targets were all well-protected naval ships. Aircraft carriers tended to steam far faster than American submarines, making it difficult for the undersea boats to get into an attack position. Radio intelligence further revealed that when submarines did attack, torpedo and exploder malfunctions often led to failure. “Whatever the causes,” Holmes would later write, “in the first two years of the war American submarines sank no aircraft carriers, and the Ultra information on carrier movements that we gave to submarines was all risk for no gain.”

  But that changed in early 1943 when cryptan
alysts cracked what many dubbed the “maru code,” the Japanese cryptosystem used to route merchant ships throughout empire waters. Solution of the four-digit code with a superimposed cipher promised to unlock the submarine force’s full potential to destroy Japan’s economy. The daily noon position reports of each convoy combined with the 8 a.m. and 8 p.m. reports of individual skippers allowed American analysts to plot Japan’s freighters, tankers, and transports in real time. This made manageable the undersea service’s otherwise impossible task of patrolling eight million square miles. A submarine skipper alerted of a convoy’s course could plan to intercept, shadow the ships, and set up the perfect attack, stretching the assault out over hours or even days if necessary. These same decrypts would allow Lockwood and his staff to gauge the success of such attacks, often listing the names of enemy ships sunk and types of cargo lost to the number of troops killed.

  Commander Richard Voge served at the time as Lockwood’s operations officer. The Chicago native had commanded the first submarine lost in the war: Sealion, bombed during an overhaul at Cavite just three days after the attack on Pearl Harbor. Voge helped make the Japanese pay for that loss when he skippered the submarine Sailfish, making four patrols into enemy waters in those dark early days of the war. Voge understood the value of the intelligence his old friend Holmes possessed. The two men met at 9 a.m. daily; Voge armed with a chart of the western Pacific that showed the estimated positions of all submarines on patrol. A review of the latest decrypts allowed the officers to update the thin-paper plot with projected convoy routes. Any late-breaking information could be shared via a private phone line that bypassed the switchboard and with hand-crank magneto ringers. The submarine force then used special internal codes—unknown to even American surface ships—to relay target info to submarines at sea.

  Code breakers developed fifty-seven possible targets in January 1943. Submarines missed forty of those for various reasons, ranging from failure to receive the message to bad weather or the submarine hunting another target at the time. Of the fifteen targets spotted, submarines attacked eleven, damaging seven and putting four on the bottom. Silversides had been one of the first submarines to benefit from this breakthrough in intelligence. On his fourth patrol off the Caroline Islands—as torpedoman George Platter convalesced from his appendectomy—Burlingame received one of his first Ultra messages: “1 freighter to arrive at 07-04N, 151-12E on 4 January 1943 approaching from position 8-01N, 149-29E—speed 10 knots.” The brief message had given Burly all he needed: the precise arrival date and position as well as the speed and course of the target. Burlingame peered through the scope at 11:28 that morning. Just as the Ultra predicted, the target arrived on time. The freighter Burlingame hoped to sink, however, turned out to be a marked hospital ship, off limits to attack.

  Another Ultra soon reached Silversides, alerting Burlingame that the 10,023-ton tanker Toei Maru would reach Truk at 10 a.m. on January 18. The message revealed that a single escort guarded this prized target. Burlingame picked up the Toei Maru hours ahead of its scheduled arrival, firing four stern tubes at 2:55. The skipper watched as multiple fish crashed into the tanker. “Tremendous explosion and pillar of black smoke 200' high,” he recorded in his report, “flame and sparks at its base and out stack.” Burlingame’s Ultra fortune continued. Intelligence alerted the skipper days later of a four-ship convoy that carried more than 4,000 enemy soldiers, horses, and supplies. Burlingame picked up the freighters on January 20 and fired six torpedoes, hitting the 5,154-ton passenger cargo ship Sonedono Maru, the 4,391-ton freighter Surabaya Maru, and the 8,230-ton freighter Meiu Maru. Code breakers had helped Burly put 17,775 tons of enemy ships on the bottom and kill more than 800 sailors.

  But Silversides wasn’t alone. Submarine successes had continued month after month as code breakers decrypted Japanese intercepts. From January to October 1943 code breakers developed 810 possible targets. Submarines spotted 354 of those, chased down and attacked 120, damaging fifty-six and destroying another thirty-three. “There were nights,” Holmes observed, “when nearly every American submarine on patrol in the central Pacific was working on the basis of information derived from cryptanalysis.” Drum was one of those. Armed with a cup of hot soup from the galley, Rindskopf liked to volunteer for the midnight to 4 a.m. radio room watch, decoding messages. The wiry lieutenant would slip on a garish yellow Aloha shirt when an Ultra arrived. Drum would go to battle stations within a few hours and attack a ship. After a few sinkings, the next time Drum went to battle stations, Rindskopf looked up to find most of the crew dressed in yellow Aloha shirts.

  • • •

  This November morning Drum lookouts had spotted the convoy at 6:05 a.m. Within three minutes, Drum dove to attack. Williamson counted three ships guarded by three escorts. One of those that appeared in the skipper’s periscope was the Hie Maru, an 11,621-ton behemoth. The 535-foot-long vessel had begun its life thirteen years earlier as a transpacific liner, shuttling passengers and cargo between Yokohama and Seattle. One of Hie Maru’s earliest voyages had been a mission of peace. The city of Yokohama in 1930 presented Seattle with an eight-ton traditional Japanese stone lantern in appreciation for the city’s help after the Great Kanto Earthquake a few years earlier. Seattle reciprocated with a gift of 2,000 rosebushes, many transported in the cargo hold of Hie Maru. Requisitioned by the military like so many other ships, Hie Maru served as a transport and submarine tender. In place of goodwill and friendship, Williamson saw only his first target—and a very valuable one.

  Williamson wasted no time. He may have been a new wartime skipper, but Williamson knew which ship to target: the largest, the Hie Maru. The former ocean liner steamed in the middle of the convoy, guarded by an escort off the starboard beam. Williamson watched the ships zigzag across the glassy sea toward him. The skipper fired all six bow torpedoes at 7:10 a.m. Twenty-two seconds later, the first exploded prematurely. With the element of surprise gone Williamson ordered the Drum deep. He heard two torpedo explosions, but the run times showed the fish had detonated at about 1,500 yards—250 yards short of Hie Maru. Not only did the run times not match, but the explosions sounded reminiscent of the earlier premature detonation. Williamson waited out four depth charges before he started up. Maybe he had damaged the ship and could hunt it down. He surfaced to find only debris and scrum from the depth charges, but no oil or wreckage. His prized target had slipped away.

  It would take almost a week of false starts, but Drum lookouts finally spotted smoke again six days later on November 17, almost fourteen miles away. The three-ship convoy, guarded by two submarine chasers, had departed Rabaul two days earlier, bound for Truk. Tamashima Maru and Nagoya Maru accompanied Hie Maru through the choppy seas. Williamson had a rare second chance at this massive target. Hie Maru had not only escaped Drum a week earlier, but had also survived an attack by B-24 bombers that had damaged the former ocean liner. Williamson didn’t want to let the ship escape again. The skipper and his men tracked the convoy on the surface, submerging more than eleven miles ahead. Williamson fired four bowshots at 2:40 p.m. and then ordered Drum deep. Hie Maru’s goodluck had finally run out. The skipper listened beneath the waves as a one-and-a-half-ton torpedo ripped open the Hie Maru’s cargo hold that had once carried an American peace offering of roses.

  • • •

  Rindskopf was impressed. He had begun the patrol worried that Williamson, handicapped by age and lack of wartime experience, would be timid. Diesel Dan had so far surprised him, sinking a significant target that would no doubt sting the enemy. Maybe you could teach an old seadog a new trick. Williamson demonstrated his aggressiveness again just five days later when lookouts spotted smoke on the horizon on the morning of November 22. The faint puff of smoke at seventeen miles developed into a four-ship convoy guarded by two escorts. Williamson, once again, didn’t hesitate. He ordered Drum down and his men to battle stations. Williamson picked the lead freighter in the starboard column as his victim and fired four fish at 11:37 a.m. He di
dn’t stick around, but ordered Drum deep. Williamson heard a string of heavy explosions that didn’t match the run times. He suspected the blasts weren’t hits, but end-of-run detonations.

  The skipper ordered Drum up to periscope depth more than half an hour later. He spotted the four freighters lying to in the opposite direction in the rippled sea. He couldn’t see any visible damage, but the freighters had not scattered. Maybe he had hit one after all. The escorts hunted Drum. Williamson had found few density layers in the waters between Bismarck Archipelago and the Caroline Islands that would shield him from the enemy’s sonar. Today was no different: he was naked and exposed. One of the escorts turned toward Drum. Williamson ordered his boat back down to 300 feet, but it was too late. The escort zeroed in on Drum. The first charge detonated at 12:43 p.m., followed by three more. The violent explosions—the worst Rindskopf had experienced in his eight patrols—shook the submarine. The escort passed directly overhead. The men inside the cramped conning tower hundreds of feet below could hear the rhythmic thump of its propellers through the hull.