A World at Arms (108 page)

The third, the breaking of codes, remains the most difficult to analyze in spite of the enormous volume of released documents and literature on the subject, because so much remains closed. Much of the German archival material captured by the Americans and British pertaining to German code-breaking is still classified so that the successes and failures of the Germans are still not well understood. The Soviet Union has released next to nothing on the subject, and there are large gaps in the British and American records due to real or imagined concerns about security. In addition, because so much of the cryptographic record was deliberately destroyed during hostilities to prevent leaks, there will always remain major gaps in the record.

The Germans had several agencies working in the field of code-breaking, and while they had a number of major successes, the very dispersion of effort hampered the Third Reich.
⁴⁷
The evidence is good that the Germans read a high proportion of French codes and were assisted by this in their rapid conquest of France. German naval intelligence had begun to break several codes used by the British navy, especially the convoy code, and used this effectively in their submarine campaign in the Atlantic. In spite of periodic alterations which caused them temporary difficulties,
⁴⁸
the Germans were able to read a high proportion of the convoy code messages until the British, finally convinced by the results of their own reading of German naval code messages, realized what had happened and changed to machine cyphers in the summer of 1943.
⁴⁹
Before this change the German superiority in code-breaking had provided them with an enormous advantage; thereafter the advantage in the crucial battle over the oceanic supply routes shifted to the Allies. When the Americans and British landed at Salerno in September 1943, the German naval signals intelligence organization could only report that the landing had not been recognized beforehand; only radio silence had provided a clue that something was afoot.
⁵⁰

Directly, and also on the basis of Italian assistance, the Germans also broke into some American codes, primarily the diplomatic ones.
⁵¹
The American machine code systems were apparently not broken into, but in this regard, as well as what the Germans and others may have learned about American codes in connection with the Tyler Kent incident discussed in
Chapter 3
, there is still a veil of secrecy.

The Germans did have considerable success with lower level Russian
codes though less so with the higher level ones. The Soviets apparently protected their most significant communications by the one code system which, if properly used, cannot be broken–the one-time pad, a system where each message is encoded according to a code used only once and only for that message. This was the system also employed extensively by the British and protected their key messages.

The Germans read several codes of their Italian, Bulgarian, Hungarian and Romanian satellites as well as considerable Turkish and Yugoslav traffic. They utilized this information at times to warn their allies lest the British and Americans read them also.
⁵²
Later in the war, the Germans learned from the intercepts about the efforts of Hungary and Romania to negotiate with the Allies about leaving the war and took precautions accordingly.
⁵³
More needs to be learned concerning German code-breaking activities; but in the interim it can probably be stated that apart from tactical benefits on the Western Front in 1940 and the Eastern Front in 1941 and 1942, the Germans benefited primarily from their reading of British naval traffic in the Battle of the Atlantic in 1942 and the first part of 1943, an asset enhanced by considerable American use of the British convoy code.

Italy had long had a large intelligence and cryptographic establishment which succeeded in reading the major Yugoslav and Turkish codes. The former were especially important to them, and the knowledge was used in 1941 for a successful deception operation which probably saved the Italian army in Albania from disaster.
⁵⁴
The major successes of the Italians, however, were originally based on thefts rather than purely cryptographic methods–but the results were the same. Having acquired important documents from the British embassy in Rome before the war, and other British material early in the conflict, the Italians were able to read a substantial proportion of British diplomatic and naval traffic early in the war.
⁵⁵

Probably one of the most important breaks into Allied code security was the result of the theft in August 1941 of the American military attaché code (called Black Code after its binder) from the office of the United States military attaché in Rome. Used by American officers around the globe, the code system provided both the Italians and the Germans, with whom the Italians shared their find, a vast amount of information of the highest value. That value was at its greatest in the Mediterranean and North Africa, where the almost daily detailed reports of Colonel Bonner Fellers, the military attaché in Cairo, involuntarily provided Rommel and the German and Italian naval and air commanders with the most complete information available to any Axis leaders in World War II. The information was utilized very effectively both by
Rommel and by the Axis commanders of the fight against Malta, and contributed in a major way to the Axis victories and Allied disaster in North Africa in the first half of 1942. Only the capture of German documents in North Africa in July, 1942, which gave away the source of German intelligence and led to a code change, closed this hemorrhage of Allied secrets. It is hardly a coincidence that the tide of fortune in that theater of war also changed days later.
⁵⁶

The Italians not only read the key codes of pre-war Yugoslavia but during the war broke into the code used by Tito’s partisans. This enabled them to confront the Germans with evidence on the negotiations between the latter and Tito.
⁵⁷
When the Germans, therefore, reproached the Italians about the negotiations of Italy with Mihailovic, Rome had an answer ready to hand. The Western Allies, whose code–breakers read radio messages on the contacts of Mihailovic with the Axis but never matched the Italian feat of reading Tito’s messages were left to make their decisions on highly partial evidence.

The Japanese appear to have been the least successful of the major belligerents in the effort to break codes.
⁵⁸
They had had some success against Chinese codes between the beginning of hostilities in 1937 and the expansion of the war in December 1941, and they appear to have broken some Chinese codes thereafter.
⁵⁹
They practically gave up, however, on the major American code systems, never broke any of them, concentrated on low-level codes with minimal success, and relied on traffic analysis for most of their communications intelligence on the United States campaign in the Pacific. The Japanese success in penetrating some of the radio traffic of the guerillas in the Philippines proved of little help to them in maintaining control of the islands. Similarly, the substantial accomplishment of breaking into Allied code messages on the northern and central fronts in Burma in May, june, and July 1944, discovered by the Americans from intercepts in October, appears not to have helped the Imperial army in the abortive invasion of India and its great defeat there.
⁶⁰
What little the Japanese learned about American merchant shipping came to them as a result of the Germans providing them with a codebook captured by one of their raiders; but this information, even on those occasions when it allowed decypherment of messages in a timely fashion, was of little use in Japan’s struggle to retain the empire she had conquered so rapidly.
⁶¹

The issue of timeliness was of central importance to all code-breaking in World War II. German reading of the British convoy code and Allied breaking of German and Japanese codes was often not in time for the information to be useful. Such breaks could sometimes still be used to read back into earlier traffic that had been monitored but could not
previously be read, and such reading back could be helpful for an understanding of the other side’s broader dispositions,
⁶²
but for operational use speed was essential–and frequently not possible. This point must be kept in mind if both the foregoing discussion of Axis and the following review of Allied signals intelligence are to be understood.

One other point frequently overlooked is that even prompt reading of an enemy’s code leads only to the decypherment of the messages sent and intercepted; if radio silence is observed or messages are simply not sent on a particular subject, either for security reasons or because the sending agency has itself been kept uninformed, there will be nothing to decypher. It was both radio silence by Japanese ships–which obviated naval locator intelligence–and the deliberate withholding of information about the planned attack on Pearl Harbor from the Japanese Foreign Ministry and representatives to the United States that made the American success in breaking into the Japanese diplomatic code and the United States navy’s excellent locator intelligence system incapable of providing warning of the Pearl Harbor attack. There was no reference to it in the codes which had been broken, since those Japanese who used them knew nothing of the plan themselves. Furthermore, there could be no locating by radio intelligence of that portion of the Japanese fleet headed for Hawaii because it sent no signals as it crossed the North Pacific.

The Americans had broken the highest level Japanese diplomatic code in 1940 and, since the Japanese continued to utilize it throughout the Pacific War, maintained their reading of it until 1945. This ability provided the Allies with important information on Japanese policy throughout the war. Furthermore, as repeatedly mentioned in this book, it also gave them significant insight into all manner of other matters because of Japanese diplomats’ reporting from the Soviet capital and from Europe on conversations, observations, and developments there.
⁶³
The details which the Allies received from the reading of Japanese diplomatic reports from Europe about new German weapons under development, about the impact of their own bombing effort, and about the political intentions and perceptions of the leaders of the European Axis, were invaluable.
^d
These reports were supplemented by those of Japanese military and naval attaches once their codes were similarly broken by the Americans. In this manner, for example, the detailed report of the Japanese Ambassador to Germany on the German defenses against an Allied invasion
was rounded out by a report on a similar inspection trip by a high-ranking Japanese officer.
⁶⁴

It should be noted that however complete the American mastery of the Japanese diplomatic and attaché codes, many messages were not intercepted at all because of atmospheric conditions, while others could not be promptly decoded or translated.
⁶⁵
On the other hand, the Japanese dismissed as improbable or even impossible every hint they received about Allied breaks into their codes; the assumption clearly was that since the Japanese were unable to break into Allied machine cyphers, the inferior Americans and British were certainly incapable of such a feat.
⁶⁶

The American partial reading of major Japanese naval codes contributed significantly to the victory at Midway, and the success of the American cryptographers in retrieving access into those codes after an interval in July-August 1942 provided the navy with a great advantage for the rest of the Pacific War. Nowhere was this more evident than in the submarine campaign, in which submarines could be directed to Japanese ships in the vast reaches of the Pacific because the shipping code was being read with great regularity.
⁶⁷

The American navy’s handling of signals intelligence was unusually effective, in large part because Admiral Chester Nimitz, its Pacific commander, early developed and consistently maintained a very high regard for the excellent naval intelligence officers in his headquarters. From Midway through Kwajalein in the Marshall Islands
⁶⁸
until the end of the war, he relied heavily and successfully on intelligence from his cryptographers. Such use was always covered by plausible references to other intelligence; thus his approval of the successful attempt to kill Admiral Yamamoto once his route on an inspection trip had been decyphered.
^e
Sadly the navy did not make good use of its intercept of a message from the Japanese submarine which in July 1945 torpedoed the heavy cruiser
Indianapolis;
as a result, hundreds of sailors who might have been saved drowned or were killed by sharks.
⁶⁹

If the American navy generally made wise use of communications intelligence, the army in the Pacific was less adept at it. This was due primarily, it would appear, to MacArthur’s unwillingness to tolerate signals intelligence personnel not under his complete command and his reliance on an intelligence chief, General Charles Willoughby, who had accompanied him from Bataan but whose loyalty was not matched by his ability.
⁷⁰
Japanese army codes were broken later than the navy’s, from the possibility that the breaking of the Japanese naval code might be compromised. Kahn,
Codebreakers,
pp. 595–601; Lewin,
American Magic,
pp. 187–91.
the spring of 1943 and early 1944 (as a result of the capture of Japanese codebooks by the Australians), as compared with the spring of 1942.
⁷¹
Thereafter, the use of signals intelligence by MacArthur’s headquarters was not as fruitful as the navy’s, though the army air force appears to have been more consistent in its use. Certainly in the bitter fighting on Leyte in the Philippines, described in
Chapter 16
, the neglect of signals intelligence by MacArthur’s headquarters was a serious error.
⁷²