The Secret War

by Max Hastings

  Early in May, Rochefort informed Nimitz that it was plain the Japanese were planning a major new initiative, though he was unsure what. Among his virtues was a fabulous memory for places, words, callsigns. The Navy Department decreed that the three codebreaking groups in Washington, Melbourne and Hawaii should each address enemy messages relating to designated geographical areas. Rochefort ignored this clumsy constraint, and strove to grasp the strategic big picture. Hypo was now receiving between five hundred and a thousand intercepts a day, about 60 per cent of all Japanese transmissions, of which its officers managed to read fragments of some 40 per cent. By 9 May, Rochefort was able to tell Nimitz that the Japanese fleet would sail for a major operation on the 21st, but added: ‘Destination of the above force is unknown.’

  The C-in-C speculated that the Japanese might intend a new assault on Pearl, or even against the US West Coast. Hypo, aware that Yamamoto was interested in the US base in the Aleutians, pondered the possibility that he might launch not just one major operation, but two. 13 May was a critical day: decrypts made plain that Pearl and the Aleutians were not the foremost Japanese priority. Instead, this was to be Midway, the most forward of all American Pacific bases, 1,200 miles north-west of Hawaii. An intercepted message instructed the supply ship Goshu Maru to load stores at Saipan, then proceed to ‘Affirm Fox’ – AF. Rochefort recalled signals back in March which identified AF as Midway. He picked up the secure phone to Layton, telling him, ‘It’s not cut and dried, but it’s hot!’ The intelligence officer said, ‘The man with the blue eyes will want to know your opinion of it.’

  Nimitz was assuredly interested, but preoccupied with other business. On the morning of 14 May he sent Captain Lynde McCormick, his new war plans officer, to discuss the possible threat to Midway. On ply sheets laid upon trestles in the Dungeon, Rochefort and his team set out their exhibits – a succession of key intercepts, together with equally important traffic-analysis data, and talked McCormick through them. This proved a fiercely intense, protracted conversation, which continued for most of the day. At the end of it, McCormick returned to Nimitz’s office and reported that he believed Rochefort had got the story right. Though Hypo had no indication of the full Japanese order of battle, it seemed plain that Yamamoto intended to commit four carriers in support of an amphibious assault on Midway.

  This was a debate of supreme importance and delicacy, of which Washington was informed. The US strategic position in the Pacific was still relatively weak, the Japanese fleet very strong. Amid the ocean’s millions of square miles, most warships took twenty-four hours to traverse six or seven hundred miles. With only two, or at best three, operational carriers of his own, Nimitz could not divide his forces. If he wished to engage the enemy, he must bet the ranch on a single rendezvous. A misjudgement about the intended destination of the bulk of Japan’s naval air force would be almost impossible to undo in time to avert a new disaster for American arms. Admiral Ernest King, chief of naval operations, gave his Pacific C-in-C little help with the decision-making. While King professed to favour engaging the Japanese wherever possible, he was wary of any course of action that might inflict further attrition on America’s dangerously small carrier and cruiser force.

  The Office of Naval Intelligence still thought the most likely enemy objective was Johnston Island, an atoll 720 miles west of Pearl; it was a further reflection on the navy’s lamentable coordination that the ONI did not know that Rochefort had already identified Johnston as Japanese designation ‘AG’. Meanwhile the Cast station in Melbourne thought Yamamoto would commit his main force against an island in the Marshalls. Layton clung to a belief that Pearl could be a target. Redman in Washington, no admirer of Rochefort’s view about anything, simply rejected this one. By Saturday, 16 May, however, the man who mattered was increasingly convinced that Rochefort was right. Nimitz accepted that the Japanese were headed for Midway, and his judgement was confirmed by an important new intercept that day, giving the fly-off position for the Japanese carriers. Yet this too failed to convince Washington that Midway was their target. Rochefort, exasperated, dismissed Redman and his comrades as ‘those clowns’ – but they were also his superior officers.

  Early on 19 May an impromptu conference took place in the basement, around the desk of Jasper Holmes. How could the doubts about Midway be dispelled? It was Holmes who conceived a solution, which was immediately adopted: the naval air station on Midway was sent a ciphered message by undersea cable, instructing its operators to send a plain-language wireless signal to Pearl, reporting difficulties with its distillation plant and requesting supplies of fresh water. An American not in on the secret exclaimed furiously, ‘Those stupid bastards on Midway, what do they mean by sending out a message like this in plain language?’ Holmes’s ruse was brilliantly judged, however: the water issue was just trifling enough to be credibly flagged in a plain-language signal, yet its substance was sufficiently interesting to Japanese eyes to merit forwarding to naval headquarters.

  On 20 May the Dungeon’s Red Lasswell broke the Japanese operation order for Midway, though unknown to the Americans, this disclosed only a part of Yamamoto’s plan: they had no hint of the fact that his main force of battleships would trail six hundred miles behind the carrier group, poised to close in and finish off the US Pacific Fleet when this appeared on the Midway battlefield – as Tokyo anticipated that it belatedly would. Nimitz now launched an effective little deception: the seaplane tender Tangier was sent to fly off planes to stage a token air raid on Tulagi; this perfectly served its purpose, convincing the Japanese that a US carrier group must be within range, and thus thousands of miles from Midway.

  Next day the Cast team in Melbourne declared that it was now persuaded Rochefort was right. On 22 May the British Far East Combined Bureau also concluded from its own decrypt activities that Midway was the Japanese objective. But in Washington, Redman and Op-20-G were furious that Rochefort had persuaded Nimitz to undertake the water-signal ruse without reference to them. Stimson, at the War Department, said the US Army still doubted Rochefort’s assessment, and feared Hypo was falling victim to an elaborate Japanese deception. Nimitz wrote uneasily in his own assessment on 26 May: ‘our sole source of information is [sigint] … The enemy may be deceiving us.’

  It is hard to overstate the personal strain on Rochefort in those days. This unloved, awkward man was making a case against the judgement of most of his peers, especially in Washington. Seldom in history has so much hung upon the word of a single junior officer. If he was wrong, the United States could suffer a strategic disaster in the Pacific. On the morning of 27 May, Rochefort donned a clean uniform: he was scheduled to brief Nimitz and his staff. Just as he was about to leave the basement for the C-in-C’s office, Joe Finnegan and Ham Wright broke a signal that identified the dates of the Japanese strikes: 3 June against the Aleutians, the 4th for Midway. This caused Rochefort to arrive half an hour late for Nimitz’s meeting, to a correspondingly stony reception. He was told to describe what he thought he knew, without mentioning in the presence of officers not in on Hypo’s secret the means by which the information had been secured.

  He outlined the Japanese plan for twin strikes, though mistakenly suggesting that the Aleutians thrust was a mere diversion: in reality, it was much more substantial. In consequence of that wrong call, Nimitz sent only cruisers and destroyers northwards. But the Pacific C-in-C made the pivotal decision to commit all his three carriers to meet the enemy at Midway, just before a bitter blow struck the Americans. The Japanese changed their codes, introducing JN-25c. This development had been expected, but the consequence was to slam shut, for a period of several weeks, Hypo’s peephole on the motions of their foes. Rochefort and his team, during the days of electric tension before 4 June, were obliged once more to rely solely upon traffic analysis of enemy transmissions, and there were precious few of these: Yamamoto had imposed wireless silence on his attacking forces.

  American carelessness jeopardised Nimitz’s trap. As his sh
ips set forth to meet the enemy, they talked too much: there was a sharp increase in US Navy wireless traffic, and the Japanese noticed. But Yamamoto scented only a mouse when he should have smelt a giant rat. In one of his major misjudgements of the war, he decided not to break wireless silence to inform Vice-Admiral Chuichi Nagumo, commanding his carrier group, that the Americans might be up to something, perhaps even heading for Midway. Here was a moment when a commander’s fear of the consequences of dispatching a stream of Morse across the ether precipitated a worse outcome than had he done so. At Pearl, tension rose to an almost unbearable level through the long hours of 3 June, as Midway’s reconnaissance aircraft gained no glimpse of the expected enemy flat-tops. Then, at 5.30 a.m. next day, exactly in accordance with Rochefort’s prediction, at last a Catalina flying-boat sent a momentous signal: the enemy’s principal carrier force was in sight.

  The American triumph that followed was anything but ordained. Destruction of Nagumo’s four carriers, the transformation of the balance of the war in the Pacific, was achieved by phenomenal luck as well as the skill and courage of the US Navy’s dive-bomber pilots. While Nimitz had gambled courageously to bring about the clash, the outcome could have gone disastrously the other way. Only on 5 June did Hypo discover Yamamoto’s battleships closing in on the scene, which prompted the American carrier groups to beat a hasty and prudent retreat. But Midway was above all else an intelligence victory, sharing with Bletchley’s breach of the German U-boat codes the status of most influential Western Allied intelligence achievements of the war. Nimitz recognised this when he sent a car to bring Rochefort to attend his own celebration party. The Hypo chief’s luck was as lousy as ever: he arrived only after the guests had dispersed. But Nimitz, in the midst of conducting a staff conference, used the opportunity to pay tribute to the codebreaker: ‘This officer deserves a major share of the credit for the victory at Midway.’

  Those words were to be Rochefort’s only reward. When he was proposed for a Distinguished Service Medal, the citation was quashed by Rear-Admiral Russell Willson, the CNO’s chief of staff: ‘I do not concur in the recommendation … he has merely efficiently used the tools previously prepared for his use. It would be inappropriate to award a medal only to the officer who happened to be in a position to reap the benefits, at a particular time, unless in actual combat with the enemy.’ Jasper Holmes wrote of the post-Midway mood in the Dungeon: ‘there was no great moment of exhilaration’. Rochefort enjoyed one more important success as chief of Hypo: he revealed the Japanese landing on Guadalcanal on 5 July, which precipitated a dramatic and ultimately triumphant American riposte with land, sea and air forces. He was also able to alert MacArthur’s command to the Japanese attempt to cross the Owen Stanley range and fall on Port Moresby.

  So far as Washington was concerned, however, far from Rochefort being dubbed the hero of Midway, he was simply an insubordinate cuss whom nobody liked. On 14 October 1942 he was relieved of his post and assigned to command a floating dry dock in San Francisco. He was succeeded by Captain William Goggins, an officer with no previous experience of cryptanalysis, who was deemed a competent administrator. Although all hierarchies commit some arbitrary injustices, this was an exceptionally brutal and mean-spirited one. In the autumn of 1944 Rochefort gained sufficient rehabilitation to be put in command of the Pacific Strategic Intelligence Unit, but he died undecorated in 1976. Only in 1985 was he posthumously awarded the Distinguished Service Medal that had been denied him in 1942.

  It was a sort of miracle that Rochefort and his team achieved what they did, with the makeshift resources available. The army–navy feud which caused the US for so long to divide its codebreaking operations was worsened by the low priority accorded to intelligence. By 1942, Bletchley and the British service intelligence departments deployed hundreds of the finest civilian brains in the country alongside a handful of career professional soldiers, sailors and airmen, together with technology in advance of anything being used by Hypo or Cast. Rochefort was merely a highly experienced, not personally brilliant crypto-linguist and analyst, and his team was a group of hitherto lowly regarded naval officers.

  The post-war narrative of the US Navy’s Pacific Combat Intelligence Center stated bluntly: ‘In the defensive stages of the war [1941–43] radio intelligence was not only the most important source of intelligence in the Central Pacific, it was practically the only source. There were very few captured documents or prisoners of war. There were no photographs of enemy positions … Excluding the Solomons and New Britain, spies and coast-watchers’ reports never supplied any important intelligence.’ The operational diary of the Japanese navy general staff recorded bitterly after Midway: ‘the enemy had grasped our intentions beforehand’. But not for a moment did Yamamoto or his officers consider the possibility that their ciphers were compromised; they attributed the disaster merely to the mischance that their carriers had been spotted by American reconnaissance aircraft or submarines.

  Joe Rochefort was not personally indispensable. After his departure the US Navy’s intelligence and codebreaking operations became ever more sophisticated and effective, although the difficulties of breaking JN-25’s variants persisted until 1944, and sometimes even beyond. The cottage industry of 1941–42 became FRUPAC – Fleet Intelligence Radio Unit Pacific – a department employing five hundred men, a formidable tool in Nimitz’s hand. But Rochefort deserves to be remembered as a man who changed history, while the honour of the US Navy was tarnished by the scurvy ingratitude with which its chiefs rewarded him.

  6

  Muddling and Groping: The Russians at War

  1 CENTRE MOBILISES

  No one who cherishes illusions about the skill and omniscience of Russia’s secret services could sustain these after studying their wartime record. It was certainly no better, and in most respects worse, than that of the Western democracies. Hitler’s invasion on 22 June 1941 precipitated a crisis for Stalin’s intelligence organisations, which like the Red Army had been crippled by the Purges. Pavel Sudoplatov’s reward for organising Trotsky’s killing in August 1940, together with his unflinching participation in many other liquidations, was an appointment a month after the start of ‘Barbarossa’ to head the NKVD’s ‘Administration for Special Tasks’, officially responsible for ‘sabotage, kidnapping and assassination of enemies’, a job description worthy of Ian Fleming’s novels. Sudoplatov handed Beria a list of 140 intelligence officers confined in prisons or the gulag for political offences whose services were now vitally needed by the state, either to spy or to run spies. He noted that the files showed all those named to have been detained on the personal orders of either Stalin or Molotov. Now, Beria asked no questions about the prisoners’ guilt or innocence, merely demanding, ‘Are you sure we need them?’ The new head of special tasks responded, ‘Yes, I am absolutely certain,’ and was ordered to arrange their release. Unfortunately, as Sudoplatov observed unemotionally in his memoirs, three of the best men proved already to have been executed. The remainder returned to intelligence duties, in varying conditions of relief and trauma.

  If the lives of the NKVD’s officers were precarious, they also enjoyed the perquisites that accrued to favoured servants of the Soviet state – for instance, the children of ‘illegals’ operating abroad were admitted to universities without being required to pass entrance exams. Sudoplatov occupied a relatively spacious apartment above the Dynamo sports store on Gorky Street, in a block exclusively tenanted by the Kremlin’s secret soldiers, including foreign intelligence chief Vsevolod Merkulov. After the June 1941 mass release of political suspects to resume intelligence work, several moved in temporarily with Sudoplatov, their deliverer. One night Merkulov suddenly telephoned to announce that he was coming down to talk. The newly-liberated officers were hastily herded into hiding in the bedroom, lest their presence prompt embarrassment. Sudoplatov had sufficient sense of self-preservation to have taken care not to sign their rehabilitation documents personally. Instead he got Fitin to do so – whi
ch, he said, probably saved his life in 1946, when his own survival hung by a thread.

  The intelligence officers rescued from the gulag rejoiced in their freedom, in some cases too soon. One of them, Ivan Kavinsky, danced about Sudoplatov’s flat in the three-piece suit with which he had been issued in place of prison denims. How wonderful it was, Kavinsky exulted, to be acknowledged at last as a patriot. Hours later, he was dispatched to serve as a stay-behind agent in Zhitomir, deep in Ukraine, which was about to be overrun by the Germans. There, he was almost immediately betrayed by Ukrainians, as many other servants of Moscow were betrayed. Arriving at a rendezvous and sensing a trap, Kavinsky shot himself. The remainder of his cell perished in a subsequent gun battle with the Gestapo. In the same fashion several other important NKVD residents were swiftly eliminated by the Germans. One of Beria’s most cynical ruses was carried out in August 1941: NKVD agents disguised as Nazi parachutists were dropped into the Volga German autonomous region, to test the loyalty of its citizens. Villages where the new arrivals were offered shelter were liquidated wholesale; the entire region’s surviving population was eventually deported to Siberia and Kazakhstan.

  Some intelligence officers liberated for war service had suffered unimaginable horrors in jails of the kind to which they had been accustomed to dispatch others. Dmitri Bystroletov, a pre-war agent-handler in Berlin, was tortured with a ball-bearing swung on a steel cable. After signing a confession, in 1939 he was sentenced to twenty years’ imprisonment. His wife Shelmatova was sent to the gulag, where she slit her throat with a kitchen knife; Bystroletov’s elderly mother meanwhile poisoned herself. It is hard to suppose that rehabilitation now secured him much happiness. Another such figure was Pyotr Zubov, who had been disgraced for failing to carry through an attempted coup in Yugoslavia. Since 1939 Sudoplatov had been urging Zubov’s qualities as an intelligence officer – without mentioning the man’s important role in Beria’s ascent to power. In prison he refused to confess to non-existent crimes against the state, even after his knees were smashed with a hammer, rendering him a lifelong cripple. Zubov was confined in the same Lubyanka cell as Col. Stanislas Sosnowski, former head of Polish intelligence in Berlin, and his compatriot Prince Janusz Radziwiłł. The NKVD set about turning both for their own purposes, and Zubov’s role in achieving this enabled Sudoplatov to secure his release. Zubov became one of his section heads, limping and shuffling around the Lubyanka.