4.5. Examining Safford's Version(s) of Events
Safford, after two days of testimony and his two memoranda, failed to convince the Joint Congressional Committee that a Winds Execute message had been intercepted and disseminated within the U.S. government prior to Pearl Harbor. He also failed to persuade the committee members that there had been a cover-up of the event. Instead, his story and evidence were shredded during the Committee's cross-examination of the evidence supporting his allegations. When he had testified before the other panels, aside from some queries to seeking more details, Safford had never been questioned critically nor had his evidence been examined with any rigor. Still, as was reported earlier, even with virtually no skeptical questioning, the majority of the prior Pearl Harbor investigations harbored some reservations about Safford's claim – the major issue being that he was the sole source of the allegations about the intercept and subsequent cover-up.
In front of the Joint Congressional Committee, though, Safford's story was subjected to a thorough and skeptical scrutiny. In trying to defend his version of events, Safford proved to be his own worst witness. He certainly was done in by his lack of tangible evidence. But, more importantly, the changing nature of his narrative finally caught up to him, and the congressional investigators would jump on this. Worst of all for Safford, the Committee had access to his original letters to Kramer and the transcripts of his testimony before the preceding Pearl Harbor inquiries and boards.
What this evidentiary trail revealed was that for the past two years Safford had been changing significant details of his narrative of events at each hearing. More importantly, as we shall see, the cross-examination revealed that he had literally fabricated the text of the purported Winds Execute message of 4 December. He revealed that he had taken the code phrases in the original message of 19 November 1941 and then presented those phrases as the text of the purported Execute message. As he was questioned further, more revelations would emerge that would expose his story as a construct of conjectures, assumptions, and misunderstandings.
In recent decades, some writers have alleged that elements of the U.S. government went through enormous efforts, to include a major search of records, as well as "hostile" questioning, in order to discredit Safford's claims about the intercept and handling of the Winds Execute message.[] Yet this interpretation is simply wrong for two reasons. First, the congressional hearings gave Safford the best platform from which he could make his case publicly. If he passed the cross-examination, then his case was solid. However, his position withered quickly as his testimony and evidence were challenged and found wanting.
Secondly, the writers miss the point that Safford's standing within the American cryptologic community commanded such respect that when he charged that a Winds Execute had been intercepted, it had to be investigated completely. The cryptologic and intelligence offices of both the Army and Navy took Safford's claim seriously and combed all of the relevant records looking for any substantiating evidence. Safford even received help from other navy officers to conduct his own search prior to the congressional hearings.[]
Another aspect of this records search, which is often overlooked, is that both the army and navy already had conducted searches for relevant evidence about the Winds Execute a full year prior [our italics] to the congressional hearings. In September 1944 the navy conducted a search of its records for any material concerning an "'execute' to the so-called "'Winds" message," but found nothing. This search, by the way, was done in response to a memorandum from Captain Safford.[]
The S.I.S. conducted its own search; again it was done more than a year before the congressional hearings. That service organized a thorough review of its records beginning in late September through October 1944 by order of then Colonel Carter W. Clarke as part of his review of classified records handled prior to Pearl Harbor. A team of five people combed all records and found nothing to support the contention that a Winds Execute message had been intercepted and processed. An index of pertinent translations of Japanese was drawn up and studied. Like the navy search, nothing could be found to validate Captain Safford's claim.[]
For this review of Safford's version of what happened, it will be simpler to separate the events into four parts: the intercept of the Winds Execute message; actions taken in the immediate aftermath; who saw the intercept or its translation; and the matter of the missing or destroyed records.
4.5.1. The Intercept of the Winds Execute Message
Before any discussion of whether there was an intercept of the Winds Execute message can begin, there exists the problem with Safford's recollection of which navy monitoring stations had been tasked to listen for the transmissions. In his February 1946 memorandum to the Joint Congressional Committee, Safford stated that he, or Commander George Welker, chief of the section responsible for actual intercept (OP-20-GX), had sent TWX tasking messages to the Navy's monitoring stations at Bainbridge Island, Washington, and Cheltenham, Maryland, to listen for Japanese Morse broadcasts. He added that he might have sent the instructions to other stations, though he does not name any specifically. However, in the same statement, he mentioned that another site, Winter Harbor, Maine, was listening for these broadcasts. He also included citations from those three stations monthly reports as evidence that they were listening to broadcasts. [Exhibit #40, pages 233–4]
The problem with Safford's statement is twofold. First of all, in a statement from 4 December 1945 Safford averred that he had sent tasking by TWX to five navy sites, including the earlier mentioned Bainbridge Island and Cheltenham along with Winter Harbor, Jupiter, Florida, and Amagansett, New York. In his testimony before the congressional hearings, he reaffirmed that these five stations were indeed listening for the broadcasts, though in his statement, he mentions only three sites. The contradiction may seem minor, but, within a span of three months, Safford had offered three different lists of stations.
The second, and much graver problem, was that when the monthly reports of the sites were examined, there was no record of any tasking being received from Washington, D.C., except for Cheltenham and Bainbridge Island. Statements from the radiomen in charge of operations at Cheltenham, Winter Harbor, and Jupiter indicate that they never received any tasking for the Winds message.[] Their position is correct inasmuch as they were never tipped off to the reason for the special tasking – the Winds message. Those monthly reports from the stations at Cheltenham and Jupiter that Safford submitted as evidence of his special tasking, actually reflected the mission tasking of those sites prior to 28 November, such as the Japanese merchant marine broadcast (known as MAM) copied at Cheltenham and the Tokyo and Osaka broadcasts to Europe monitored by Winter Harbor.[] Further, Lieutenant Commander George Welker had told Safford in a letter that he recalled no tasking specific to the Winds message was sent to the stations that OP-20-G controlled.[]
Safford's statement indicated that the Winds Execute message was intercepted by Cheltenham shortly after 8:00 AM (EST) during a broadcast by Japanese station JAP. The intercept was then quickly sent in by teletype to the OP-20-GY watch section within a half hour. [Exhibit #40, pages 229–241] Safford also wrote that an unnamed watch officer first had shown the intercept to Lt. Kramer. According to Safford's account, Kramer underlined the important code phrases – all three were present in the text – and wrote in pencil "free" translations which were "War with England" (including the Netherlands East Indies), "War with the U.S.," and "Peace with Russia." Kramer came into Safford's office and said, "Here it is." [Exhibit #40, page 240]
Safford's account of the time of the intercept contradicted virtually every prior statement he made. In his second letter to Kramer, Safford wrote that the "Weather Report" was broadcast at 4:30 AM (EST) on either 4 or 5 December 1941. In front of both the Hart and Hewitt hearings, he indicated that the intercept occurred on the evening of 3 December and had been sent to Washington that evening when Kramer had verified the text.[] By the time of the congressional hearings, Safford had settled on 8:00 AM, 4 December, as the time of intercept. Why he had done so, as he explained to the Committee, was that just two weeks earlier he had reviewed monthly reports from the sites at Winter Harbor and Cheltenham. He had seen information in the reports which allowed him to postulate when the Winds Execute might have been intercepted and by which station. Safford never detailed what he saw, but information in his statement suggests that he had noted that, according to his calculations, both stations could have heard the Japanese broadcast station with the call letters "JAP," one of the stations he earlier had speculated might broadcast the Winds execute. [Exhibit #40, page 251]
Yet this statement only exacerbated Safford's problem because his postulated time of intercept left him with only the possibility it was heard by either Winter Harbor or Cheltenham. In his statement of 1 February, he had confidently asserted that Cheltenham had intercepted the Winds Execute. But here the weight of his previous testimony bore down on him. In his January letter to Kramer, Safford had written that the message had been heard by both Cheltenham and Winter Harbor. But, later in his statement to the Hart Inquiry, he left out any mention of the intercept site. (Safford claimed that Admiral Hart thought the information "irrelevant.") Before the Naval Court of Inquiry Safford did not name a station.[]In front of the Hewitt Inquiry, Safford said he did not know what station actually intercepted the message, but "guessed" that both sites had the better facilities for monitoring for the broadcast.[]
Yet even Safford's guess could not hold up to scrutiny. In fact, the navy had interviewed the radiomen-in-charge of the Cheltenham and Winter Harbor stations, D.W. Wigle and Max Gunn, and both deposed that their sites had not intercepted such a broadcast as the Winds Execute.[] Finally, during the congressional hearings, only a day after he presented his statement with the confident assertion that Cheltenham had heard the Winds message, Safford was forced to admit to the Committee counsel that there was no evidence that "Cheltenham got that message."[] His Cheltenham claim was based solely on a conjecture that Cheltenham theoretically could have heard a broadcast by station JAP.
As for the action inside the OP-20-GY office spaces when the message supposedly arrived, Safford's version of those events came under considerable correction from the very people he had named as participants. In his 1 February statement, Safford did not name the watch officer who brought the intercept to Kramer. However, from as far back as early 1944, Safford had claimed that at least one of the GY watch officers was a witness to the existence of the Winds Execute message. In his January letter to Kramer, Safford insisted that one Lt. Allan Murray recalled the message, while either Kramer or another watch officer, Lt. Francis M. Brotherhood, brought the message to him. In front of the Hart Inquiry, Safford had stated that Brotherhood was on watch on the evening of 3 December when the Winds execute arrived. Before the Naval Court of Inquiry, Safford said that Lt. Murray or "possibly Kramer" had come in with the yellow teletype sheet and said, "Here it is."[] When the congressional counsel asked Safford about the discrepancy in his story, he stated that he had testified that Brotherhood had brought him the message since that officer had told him the message had come in.[]
However, when Murray and Brotherhood testified to the Hewitt Inquiry, they denied they had delivered such a message. Brotherhood recalled the FCC mistaken intercept of the evening of 4 December and that he had notified Admiral Noyes that same evening. Murray stated that he was the watch officer for the day shift for both 4 and 5 December. His watch ran from 8:00 AM to 4:00 PM. He told the Inquiry that he would have been aware if any such Winds Execute had arrived at the time described by Captain Safford, but he stated no such message came in.[]
Alwin Kramer testified before the congressional committee for almost four days. His version of events differed decidedly from Safford's. He recalled that the incident occurred on 5 December. He had been handed a short piece of teletype paper with about two or three lines of Japanese text, not the two hundred words Safford recalled. He could recall the text in detail, but he said that there was only one phrase on it. He never underlined it or translated the phrase. Nor would he have used the word "War," since the instruction messages never used that word, rather the Kana phrase WAGAHOO NO GAIKOO KANKEI KIKEN NI HINSURU [Exhibit #7], which was translated as "our foreign relations are approaching danger." He added that he had looked at the paper for only about ten to fifteen seconds. He could not recall whether he had entered Captain Safford's office. He never saw that strip of paper again.[]
Kramer's testimony leads into the serious question of exactly what the contents were of the Winds Execute message that Safford believed existed. Recall that Safford stated that all three phrases appeared in the message. (Safford in his statement referred once to codewords – a misleading expression. He later referred to code phrases in his statement and always "phrase" in his testimony.) He said that the phrases occurred during a Morse broadcast. Safford maintained that the Winds Execute message did not have to be sent in a Voice broadcast. [Exhibit #40, page 10] That was true, but it is not the entire story. According to the instructions in Japanese message No. 2354, the format for the Winds Execute in the General Intelligence (Morse) Broadcast was the repetition of a single codeword – HIGASHI, KITA, or NISHI – five times at the beginning and then at the end of the broadcast. The code phrases were to be used in the voice broadcast. Yet, according to Safford, the code phrases, intended for the voice broadcast, appeared in the Morse broadcast. Even more, Safford could not verify if the format was correct and if the phrases had been repeated at the end of the purported broadcast of 4 December.[]
As far back as his second letter to Kramer, Safford had admitted that the format of the Winds Execute was "not right." Yet he could never explain why, after the Japanese Gaimusho had established two discrete formats for warning messages intended for its diplomats, it would then send the warning phrases on the Morse broadcast when it previously had set up a format with single codewords?
Safford also had claimed that the phrase regarding relations with the Soviet Union, KITA NO KAZE KUMORI meant "Peace with Russia" (or "No War"). In testimony to the Congressional Committee, Safford added that it had been believed that "'no war' would be no mention [of the phrase], but they [the Japanese] gave a positive, specific mention as to Russia, but in a negative sense, which we concluded meant peace, or not war as yet."[] This interpretation was totally opposite the meaning set out by the Japanese message No. 2353, in which it is stated explicitly that the phrase meant that relations were approaching a dangerous point. The absence of the phrase from the broadcast was the true "negative" meaning, that is, no danger to relations.
Yet Safford's odd interpretation of the phrase KITA NO KAZE KUMORI was not new. Back in his second letter to Alwin Kramer, Safford had stated that there was a "negative form" of the phrase for Russia, KITA NO KAZ KUMORI (North Wind Cloudy). He said that this "form" was the phrase KISHI NO KAZE HARE. But this "negative form" phrase actually was the warning phrase for relations with Great Britain, which translated to "West Wind Clear." In later testimony, he would change his explanation of the format for his so-called "negative form." Before the Hart Inquiry in April 1944, he stated that this "negative form" was KITA NO KAZE KUMORI, which meant, according to Safford, "Neither North Wind or Cloudy."[] Clearly, this translation of the phrase is not supported by the text. During the June 1945 Hewitt Inquiry, Safford again was challenged on this point, but he managed to avoid a direct answer by just restating that KITA, or "north," was the "negative form."[]
Safford's confusion over the meaning of the code phrase for Russia revealed the fundamental discrepancy at the heart of his claim. It was this: In early 1944, as Safford began to construct his claim about the Winds Execute message, he realized he could not recall any of the text of the purported Winds Execute message of 4 December, including what open code phrase, or phrases, had appeared in it. In fact, it was in Kana and he could not read it. Therefore he had to reconstruct the entire message, specifically the code phrases. To do so, he simply appropriated the three phrases from the 19 November message No. 2353 and then misrepresented those phrases to the Hart Inquiry (April 1944) as the actual text of the Winds Execute message. Exactly what text he recalled and what he added, Safford was unclear. He revealed this ploy before the Army Pearl Harbor Board. On 2 October 1944, during his testimony, this exchange occurred:
General Russell. Now let us turn back to the message. From what source did you obtain these Japanese expressions or words which are found in your evidence given to Admiral Hart?
Captain Safford. I got those from the messages setting up the "Winds" code, plus my recollection of the events: that two came exactly as we expected them, that is one for America and for England, and also the negative form of the Japanese for "North Wind Cloudy." I do not know enough about Japanese to be able to give that from memory. I mean, I remember that it was exactly what we expected to get on those two occasions, and garbled up on the Russian business.
General Russell.Then the memorandum from which you refreshed your recollection at that time you testified before Admiral Hart, as a matter of fact, was the code that you had discovered prior to November 28, 1941, [the release date of the translation of message No. 2353] and that you took that language from that Japanese code and compiled from recollection the message of December 4th and gave that to Admiral Hart as being the message of December 4th; that is the truth?
Captain Safford.That is correct, it being essential or the substance of what we were interested in, because there was a lot more which was just straight Japanese news, and I couldn't make head or tail of it.[]
This exchange also enlightens the origin of the issue of the "negative form" of "North Wind Cloudy." If Safford insisted that all three phrases appeared in the purported Winds execute, then he had a paradox and that was the fact that Japan did not attack the Soviet Union in 1941. Therefore, according to the Japanese format, the phrase KITA NO KAZE KUMORI should not have appeared. So Safford needed to interpret the phrase for the Soviet Union in a completely different manner than the original Japanese meaning in order for it to have appeared in his "reconstructed" Winds Execute message. From then on, in order to portray his artificial Winds Execute message as valid, Safford had to claim that the KITA phrase meant "peace" not war with Russia, which flew in the face of the meaning the Japanese had assigned to it.
In essence, Safford, being unable to recall the full contents of the Winds message he imagined had been sent, simply appropriated the phrases from the set-up message of 19 November and then presented them as the actual Execute message.
4.5.2. Actions Taken in the Aftermath of the Winds Execute Message
Safford, in his prepared memorandum, stated that he had sent the original intercept of the Winds message to Admiral Noyes (Director of Naval Communications and Safford's superior officer) by a courier. Admiral Noyes' office was one floor up and directly above Safford's office. He told the courier to deliver it to Noyes and not to take "no" for an answer. Within a few minutes, Safford said he had received a report that the message had been delivered. Safford also stated that he was satisfied that Noyes had telephoned the "substance" of the message to the War Department, the "Magic" distribution list in the Navy Department, and the Naval Aide to President Roosevelt. Six or seven copies of the translation were sent to the War Department, though a "smooth" (or finished) translation was made in the Navy at that time. Safford added that he believed the Army had distributed the translation. Eventually, a translation was made with the serial number "JD-1 7001." Safford also added that two urgent messages went out from the CNO staff (OPNAV) to various naval facilities in the Pacific, which, in view of the critical situation, ordered the destruction of certain ciphers. [Exhibit #40, pages 242–245]
Again, like much else from Safford's statement, his previous statements and testimony contradicted his testimony before the congressional hearings about the distribution of the translation. In his statement to the Hart Inquiry as well as in testimony to the Naval Court of Inquiry and the Army Pearl Harbor Board, Safford never mentioned sending the intercept to Admiral Noyes. Instead he claimed that OP-20-G, his office, prepared the smooth translations and distributed to the appropriate navy offices such as the CNO, the Director of War Plans, intelligence, communications, etc. Copies were also sent to the State Department, the White House, and the War Department.
Safford stated that this widely disseminated translation had been given the serial of "JD-1, 7001" on 4 December, but when he had tried to locate it he discovered that serial had been cancelled with out any explanation. However, Safford was wrong about the missing serial being assigned to the translation. For one thing, he could not account for the fact that serial numbers subsequent to 7001 had been assigned to messages intercepted prior to 4 December. A list of Navy serial numbers showed, for example, serial number "7017" had been allocated to a Japanese diplomatic message to Washington that was intercepted on 2 December. The translation was issued on 3 December – a full day before the purported Winds intercept. [Exhibit #45][] In fact, when Safford testified to the Army Pearl Harbor Board, he had admitted that he had no direct evidence that JD-7001 was the serialized translation of the 4 December Winds message. At best, he said, there was only "circumstantial evidence." When pressed for a better explanation that serial "7001" had been issued on 3 December, a full day before his purported intercept of the Winds Execute message, Safford could reply only that "things sometimes got a little bit out as far as putting those numbers on was concerned."[]
As for the flurry of warning messages sent out as a result of the arrival of the Winds Execute, Safford had some part in preparing two messages on 4 December sent out to naval bases in the Pacific that ordered the destruction of extraneous ciphers. What is remarkable about Safford's actions after the intercept of the Winds Execute was that he did nothing else. Yet in all of the retellings of his narrative, as far back as his statement to the Hart Inquiry in mid-1994, through to his memorandum to the Congressional Committee in early 1946, Safford emphasized that the Winds message meant war, or that Japan was committed to war. Before the Hart Inquiry he even went as far as to state "We [persons not further identified, but likely Naval Intelligence and OP-20-G] believed that the Japanese would attack by Saturday (December 6), or by Sunday (December 7) at the latest.[] This remarkable sentence is echoed loudly in his 1 February 1946 memorandum that the Winds message "meant war."
For all the urgency that Safford evoked four years after the purported intercept, at the time he did nothing that suggested he saw the immediate danger of war. The two cipher destruct messages he referred to as being transmitted in response to the Winds warning, in reality, were drafted originally by Admiral Noyes' office and sent to the CNO for release. What is more telling, though, is that the CNO had sent out similar messages about code and cipher destruction the day before (3 December) and two days later (6 December). In fact, Safford's selected messages are just part of an ongoing effort by the Navy to remove potential compromises of excess cryptographic material.[] There is no evidence that the alleged Winds Execute had any connection to this series of messages.
As for the warning message that Captain McCollum supposedly was to send to the Pacific commands based on the purported Winds message, the record did not bear out Safford's claim. When the Congressional Committee asked him about the warning message, McCollum explained that on 4 or 5 December, he drafted a message to the Pacific commands that highlighted recent intelligence that suggested, or indicated, that the Japanese might initiate hostilities very soon. He said he took the message to his superior and then on to Admiral Wilkinson, Chief of ONI. Wilkinson said that it had to be approved by Admiral Richmond Turner, the head of the War Plans Division, who was responsible for drawing such conclusions from intelligence. According to McCollum, Turner edited the warning parts of the message and then showed him the warning messages already sent to Admiral Kimmel. McCollum took the edited message back to Wilkinson, who told him to leave it. McCollum added that the message was not sent. But this was not unusual, he added. Many dispatches had gone unsent; that was the prerogative of his senior commanders.[]
When asked directly whether the draft dispatch was related to the Winds Execute, McCollum stated that Safford was misinformed. He added that Safford's claims that the drafted message had a reference to the Winds message and that McCollum had wanted to avoid another Port Arthur (a reference to the surprise attack by Japan on the Russian Pacific Fleet in 1904) were untrue because there was no such Winds message in the first place. When Safford was confronted with McCollum's denial, he insisted that he had been in Admiral Noyes' office when Wilkinson brought in the message. Safford recalled looking its several pages over and seeing the reference to the Winds execute. When told that McCollum had stated that the draft message was about one-half a page, Safford could only claim he had seen a multipage one.[] Safford said that he had phoned McCollum late on 3 December and pointedly asked him if he was going to send a warning to the Pacific Fleet.[] But this exchange occurred a day before the purported Winds Execute intercept. Safford also admitted that he had never spoken to McCollum after that time; he had assumed that McCollum had seen the Winds message.[] Admiral Wilkinson stated that there had been a draft message that both McCollum and Turner had decided it was not necessary to send out. But Wilkinson added one interesting note during his testimony: that another such message was contemplated when word of a Winds message first came in, but was dropped once the report was proven false.[] This placed the incident on 5 December and tied it in with the mistaken FCC intercept.
While Safford was limited in what messages he could send out, certainly he could have drafted some notice to all involved naval monitoring sites, and to those of the army and FCC as well that the Winds message had been intercepted. Yet in his memorandum he records no other action. In an earlier memorandum to the Hewitt Inquiry, (14 July 1945) Safford stated that this very issue of alerting monitoring stations to the intercept had come up. He added that, after discussions with the head of intercept operations, Lieutenant Commander George Welker, it was decided not to order a cessation of the collection of Japanese broadcast because of the chance that the "hidden word/STOP" message might be sent.[] Interestingly, though, Safford did not mention this story in his statement to the Congressional Committee. There is a good reason: in a letter to Safford in January 1946, Welker told him that he could recall nothing of a Winds message ever being intercepted or what was done afterwards with it.[]
Finally, Safford's actions in processing the purported Winds Execute message seem odd in view of the prescribed division of effort between the S.I.S. and OP-20-G. Recall that the S.I.S. had responsibility for processing intercepted messages on even days. Yet on 4 December Safford did not inform the army that the message had been intercepted. In both versions of events that Safford told, whether he informed Admiral Noyes of the intercept or prepared translations of it, in neither case did he pass the intercept to S.I.S. to produce a translation as was required under the standing agreement. He kept it within the Navy offices. This action should be contrasted with that of Alwin Kramer, who, on 6 December, when notified of the arrival of an important Japanese diplomatic message – the fourteen-part message that ended negotiations – proceeded to call back in civilian S.I.S. analysts who had just left to go home for the day, to work on the decryption and translation.
The fact is that Safford, aside from some undetermined role in the preparation of two messages to Pacific installations ordering destruction of cryptomaterial, did nothing else in response when the purported Winds Execute message was intercepted, despite his later claims that he recognized that the appearance of the message "meant war."
4.5.3. Who Saw the Winds Execute Intercept or Translation?
As with all the preceding parts of Safford's story, his various lists of those "who knew" in some way or manner about the Winds Execute was a fluid affair with names on one list disappearing from another, while the nature of an individual's knowledge changed over time.
Safford's most recent list was one he presented to the Congressional Committee on 2 February 1945 that included twenty-six names (see page 65 for list). This list was a copy of the one that had been prepared earlier for the counsel of the Hewitt Inquiry, Commander John Sonnett, on 14 July 1945. He explained that these people "knew in December 1941 that the Winds Execute message had been broadcast from Tokyo on 4 December, although some of them did not learn about it until after the attack on Pearl Harbor."
Yet when the Congressional Committee's counsel pressed Safford for more information about the names, he was less certain about them. For example, during the 2 February 1946 session, immediately after he read the names of the twenty-six people who he claimed knew about the Winds message, Safford was asked who on the list actually saw the message or translation. At first Safford said that the named individuals had "seen or been told about it." A committee member asked him again if he could verify that the people on the list saw the message. Safford backtracked and said that, except for Captain Alwin Kramer, "I have no knowledge that any of these people saw it.[]
Of the twenty-six individuals named by Safford, twenty-two testified or deposed under oath before the many hearings that they had no knowledge of the Winds Execute message being intercepted before 7 December. (One, Colonel John T. Bissell, was mistakenly identified by Safford as General Clayton T. Bissell.) Some recalled that a mistaken or "false" Winds message had come in the week prior to Pearl Harbor.[] Many of the witnesses said that they had learned of the Winds message only recently from reading the papers. No wonder, since prior hearings had been held in camera and many of these individuals had not been asked to testify. At least one later writer tried to transmute these truthful statements into a lie by implying that the common response about the newspapers appeared to have been scripted.[] But this aspersion could not hold. If there had been no Winds Execute, how else could these witnesses learn about it but through the papers or hearsay?
As for the four witnesses that the Congressional Committee did not interview, two, Welker and Chief H.L. Bryant, previously had responded by letter to mailed inquiries from Safford in which he asked them about the Winds message. Both Welker and Bryant wrote back to Safford that they never knew of such a message being intercepted.[] Interestingly, both had replied to Safford before he had supplied their names to the committee. As for Commander Parke, Safford noted before the Hart Inquiry that he had only second-hand knowledge of the message. General Olmstead, at the time the Army's Chief Signal Officer, had been in Panama on an inspection trip from about 2 or 3 December until 20 December.[]
Two days later, Safford was asked about the claim in his letter to Kramer that there were other unnamed people with knowledge of the Winds message. The line "No one in OPNAV can be trusted," was read to him. A senator asked him to supply the names of those he knew in OPNAV with knowledge of the message. Safford refused, announcing, "I would prefer not to answer." The committee then queried him about the line, "Premature action would only tip off the people who framed Admiral Kimmel and General Short." Safford replied that he did not know who framed the two officers. He added that he was "referring to the War and Navy Departments in general, but not to any specific individual I can identify."[]
In his letter to Kramer of 22 January, Safford mentioned he had a list of fifteen reliable witnesses. When the committee asked him to name these people, Safford told them that he had given the list to the Hart Inquiry, but at this moment could not recall one name from the same list. Grilled more about this list, Safford admitted that of the fifteen, eleven would no longer "make the same statements as they did two years ago." Only four – Alwin Kramer, Colonel Moses Pettigrew, Colonel Rufus Bratton, and Colonel Otis Sadtler – could give him (Safford) some support if not complete support. As it turned out none gave him support. By the time of the Hewitt Inquiry, Kramer had already substantially reversed his version of events that once had seemingly supported Safford. (See section about Kramer, page 77.) Pettigrew, who was the executive officer of G-2 at the time of Pearl Harbor, recalled that he had been told on 5 December about a "Winds Code" and that subsequent to this, a message had been sent to the Army G-2 in Hawaii to get in touch with Commander Rochefort about the Winds.[] Bratton and Sadtler testified that they had reacted to the "false" Winds on 5 December. They had received notice of a possible Winds message from Admiral Noyes on the morning of 5 December, but it had turned out to be wrong. Colonel Sadtler could not have known about the purported Winds message until 5 December since was out of his office the day before. He was attending a meeting of the Defense Communications Board.[]
4.5.4. The Matter of Missing or Destroyed Records
Safford had made the charge of missing or destroyed records an important part of his allegation. It already has been demonstrated previously in this chapter that the so-called missing translation of the Winds execute, "JD-1, 7001," was, in fact, assigned a day before the purported intercept of the broadcast. So it was not, as Safford believed, the serial of the translations of the Winds Execute message. The cancellation of the serial was irrelevant.
Another major charge by Safford, which came out in the Hewitt investigation, was the statement that General George Marshall had ordered the destruction of all records related to the Winds message. Safford maintained that William Friedman sometime before the Hewitt Inquiry had told him this story.[] When confronted by the Congressional Committee counsel as to who told him and who else might know about Marshall's order, Safford could tell the counsel only that he had never had any conversation with anyone other than Friedman about the alleged order from Marshall. Safford also was unaware of the findings of the Clarke investigation, which had already reviewed the basis for the charge and had found no evidence supporting it.[]
The Clarke investigation had reviewed this incident in detail and had followed the chain of the hearsay back to its alleged source. When Colonel Clarke asked Friedman from whom he had heard this story, he said Colonel Otis Sadtler had told him this. Sadtler, in his turn before the investigation, said Colonel Ike (Isaac) Spalding, head of Army G-1, or Personnel, at the time had told him[] Spalding told the investigation that he had been told the story about Marshall's order by Colonel John T. Bissell, head of Army counterintelligence at the time. Bissell had said that certain army intelligence papers or files had been destroyed after Pearl Harbor.[] When Bissell testified to the Clarke Investigation, he said that this was not true. He had no access to communications intelligence material from the S.I.S. He did recall that the draft version of the message from G-2 to Hawaii about the possibility of sabotage had been destroyed shortly after Pearl Harbor.[]
As for Safford's claim that all the records of Cheltenham were missing, this, too, was demonstrated to be false. What Safford did not realize was that the navy's standard procedure called for the periodic destruction of outdated or extraneous material at field stations. Far from being the exception, Cheltenham, like all other sites, had burned such records regularly. Cheltenham's records from late 1941 had been destroyed in December 1942. [Exhibit #49][] The copies of intercepted messages, reports, and logs from Cheltenham had been shipped to Washington and were available for review.[]
In fact, Safford was being disingenuous when he insisted the records had been destroyed. A Lt. George W. Linn, who had been one of the OP-20-G watch officers during that period – in fact, he was the senior officer of the watch and spent daytime working hours in the GY office area and was present on 4 and 5 December – assisted Safford in his search for record evidence to support his contention. As Linn recalled, Captain Safford had decided to search station intercept logs for a copy of the execute message. He believed that some station had heard it and this would be reflected in the logs. As Linn recounted, Safford worked out the possible broadcast times and frequencies and the monitoring stations that might have heard them based on his own estimates of the local propagation conditions. Linn would then retrieve the microfilm records and check the station intercept logs [our italics]. He found nothing. Still, Safford believed that Cheltenham had heard the message.[]
4.5.5. Some Observations on Captain Laurance Safford
With all of the skepticism that greeted Safford's claim about the Winds Execute message at the congressional hearings and the reservations expressed about it by some of the preceding inquiries, as well as his continued inability to produce any supporting evidence after a two-year search, it is probably fair to ask why Safford stubbornly persisted in his claim? One observer, George Linn, noted that Safford was not "pleased" with the lack of progress in convincing the various boards and inquiries of his case.[]
There is evidence that Safford believed that Admiral Kimmel was being treated unfairly and blamed totally for the Pearl Harbor disaster. Certainly Safford was not alone in his conviction; many fellow officers believed Kimmel was a scapegoat for the failure in the Roosevelt strategy in preventing Japan's attack. Interestingly, Safford admitted to the Congressional Committee on 6 February 1946, initially he was very "bitter" towards Admiral Kimmel for failing to take measures to alert Pearl Harbor to a Japanese attack, even more so since he believed the 4 December warning message from McCollum had been sent out. But after he learned of the unsent message, the object of his bitterness turned, as he said, to the men in the Navy Department and himself. Now he felt it was important for him to do everything he could to help Kimmel.[] Yet does this turn of heart explain Safford's persistence in the face of continued skepticism or reservations about his allegation or his almost libelous accusation that General Marshall ordered the destruction of relevant records?
Safford's conversion does not explain satisfactorily the lapses in his expertise in areas of radio signal propagation, collection, Japanese communications procedures, and the information available in the Winds "set up" messages. Yet Safford seems to have shrugged off the obvious contradictions and technical errors that permeated his statement and testimony. To those outside the fields of communications and cryptology, Safford's claim may have appeared solid and technically based. Yet when the details of his narrative were examined, many were found to be wrong, or in the case of the "negative form" of the positive phrase for the Soviet Union, to be simply absurd.
It must be pointed out that Safford was not the unambiguously unselfish and solitary hero who struggled alone against a government-wide conspiracy to sacrifice Admiral Kimmel in order to cover up its knowledge of the impending attack on Pearl Harbor. Safford was not above trying to convince other witnesses they were wrong, as in the case of Lieutenant Brotherhood. He may have convinced Kramer against the latter's better judgment that an Execute message had been sent. But when Kramer changed his testimony, Safford portrayed his former colleague as "befuddled."[] Safford also claimed that individuals, such as Chief Bryant and Commander Welker, knew about the Winds intercept, when, in fact, in private correspondence with him they explicitly had denied knowing anything about the message. Also, Safford readily passed along, without any effort to verify it, the charge that General Marshall had ordered the destruction of records dealing with the Winds message. Before and during the congressional hearings Safford had been in close contact with the minority (Republican) members of the Joint Congressional Committee. Admiral Kimmel's counsel had coached Safford on how to answer the committee members, especially the technique of answering any question without giving more information than for which he had been asked. This latter ploy was obvious during his testimony regarding who had seen the Winds message.[] When everything about Safford's role in the Winds controversy is considered, he was, according to Henry Clausen, "a strange duck."[]
The most damaging problem for Safford was that a major portion of his version of events and many of the details of his evidence continued to change over the two and half years from when he began his search in late 1943 through to his testimony before the various hearings on Pearl Harbor from 1944 to 1946. The glaring differences in events and details that marked Safford's testimony at each separate inquiry finally caught up to him when he appeared before the Congressional Committee. The malleable clay that was Safford's evidence was not the stuff upon which a solid case could be built.
In the final analysis, Captain Safford's "evidence" for the existence of a conspiracy to cover up the Winds Execute message simply failed to pass muster. He had not encountered such questioning in any of the previous inquiries or hearings. In those sessions, his testimony and claims were accepted, usually with only queries designed to elicit more detail. Under the cross-examination of the committee's counsels and its members, his case simply disappeared.
After its hearings, and in considering all the evidence from the prior investigations, the Joint Congressional Committee arrived at its conclusion about Safford's story, the existence of a Winds Execute message, and the importance of it all:
....it is concluded that no genuine message in execution of the code and applying to the United States, was received in the War or Navy Departments prior to December 7, 1941...it is believed that Captain Safford is honestly mistaken when he insists that an execute message was received prior to December 7, 1941. Considering the period of time that has elapsed, this mistaken impression is understandable.
Granting for purposes of discussion that a genuine execute message applying to the winds code was intercepted before December 7, it is concluded that such fact would have added nothing to what was already known concerning the critical character of our relations with the Empire of Japan.[]