4.2. Safford Searches for the Missing Winds Execute Message
The story of Safford's search for the missing Winds Execute message begins about two years after the attack on Pearl Harbor. We do not know with much detail what he initially did to locate records, how he went about the search, to whom he first spoke, and when he searched. The exact date of origin and source of his belief that such the message had been sent cannot be precisely placed. What we do know is that sometime in mid- to late 1943, Captain Safford had begun writing a short history of American naval radio intelligence called "The Undeclared War."[] It was completed on 15 November 1943. Interestingly, the last section of the monograph concerned the intercept and handling of the Winds Message instructions, Japanese serial Nos. 2353 and 2354. The section ends on 28 November 1941 with the transmission of the technical message to Hawaii and the Philippines that contained the Japanese broadcast schedule. Next to Safford's signature is a comment: "Not written smooth beyond this point."[] Ordinarily, this comment would be interpreted simply as a comment about the unfinished nature of the history. But in the light of events to come, it now appears to have been an omen.
According to Safford in later testimony, at about the time of his abrupt termination of the history project, he was reading the transcript of the Roberts Commission (18 December 1941 – 23 January 1942) that investigated the Pearl Harbor attack, when he realized that a warning message to Admiral Kimmel from naval intelligence, drafted by Commander Arthur McCollum on either 4 or 5 December, which Safford previously understood had been sent to Pearl Harbor, in fact, had never been transmitted.[] According to Safford, this proposed warning message to Kimmel had resulted from the intercept of a Winds Execute message on the morning of 4 December.
Concerned that no warning message had been sent, Safford undertook his own private investigation to find the translation of the Winds Execute intercept and discover what had happened to McCollum's warning message. His first step was to recover the original Winds intercept and translation: the yellow TWX paper from the intercept station that had copied it and all of the copies of the translation (anywhere from six to over ten) that he believed had been distributed to the S.I.S. and various offices within the Navy. Unable to locate any copies of the documents within the navy, he later informally asked personnel in the Army's S.I.S. and G-2 "on several occasions" to send him a copy of the translation, but he was advised that the document could not be located, though Safford would testify later that he believed that it "was common knowledge [in the Army]" that the translation existed.[] During the search through OP-20-G files that had come up empty, Safford stumbled across a reference to a cancelled navy serial number, JD-1 #7001, which he believed was the serial number OP-20-G had assigned to the translation of the Winds Execute message for which he was searching.[]
No doubt Safford became frustrated over his inability to recover the missing intercept or the translation. He also tried to locate the intercept logs of the East Coast navy monitoring stations he believed had copied the Winds message. But he struck out here, as well. The station logs could not be found. A further inquiry turned up information that at least one station, Winter Harbor, Maine, had destroyed all the logs in mid-1942 with the approval of the Navy.[]
During this time Safford widened his search for evidence and wrote a letter on 22 December 1943 to the one person he believed knew about the Winds Execute message, Commander Alwin Kramer. [Exhibit #41][]In the letter Safford posed a series of vaguely worded and leading questions about the intercept of Japanese diplomatic messages during the week prior to the attack on Pearl Harbor. One question, number 18, concerned the Winds message. "We cannot find the original "Weather Report" (Sent Dec. 5th) and its translation. What became of it?"[]
Figure 4.1. Captain Alwin D. Kramer
Kramer's response, which took some time to arrive since he was stationed at the Joint Intelligence Center, Pacific Ocean Areas, Hawaii, came in a letter of 28 December. It did not help Safford's cause. [Exhibit #42][] In his letter, Kramer noted that "The first (sic) one of the '" "' was not as indicated in parentheses..." It was, as Kramer added, really one of the file of translations turned over at the meeting, which Hull, Stimson, and Knox attended on the morning of 7 December. The folder included the Part 14 of message serial No. 902, the STOP message, and the instructions from Tokyo to the embassy in Washington to deliver the message by 1:00 PM, the latter of which Kramer refers to as "Item 11 (first one) on how the hour tied with the sun, and moves in progress, elsewhere." Kramer then added that he believed that the translations were available in the sections of OP-20-G that handled either collateral intelligence, "GL," or translations and code recovery, "GZ."
This was not the answer Safford expected (or wanted), and he told Kramer so in the next letter he mailed on 22 January 1944. [Exhibit #43][] This letter becomes critical to the story of Safford's later claims for two reasons. First, Safford created a "condensation code" of names, places, dates, and objects that ran two pages. [Exhibit #43, pages 266–7] This code list he appended to the letter. Why he chose to create a "code" for his correspondence was revealed in the letter. He perceived an effort by people in the staff of the CNO to "frame-up" Admiral Kimmel. This leads to the second point in that this letter becomes Safford's initial statement regarding the nature, scope, and gravity of an apparent conspiracy that covered up the record of the events surrounding the Winds Execute message as he recalled them.
Safford began with a warning to Kramer (with the text in place of Safford's code values):
Be prudent and be patient...No one in OPNAV [The Chief of Naval Operations' Staff] can be trusted. Premature action would only tip off the people who framed Admiral Kimmel and General Short. Tell Halsey [Admiral William "Bull" Halsey] that I knew Adm[iral] Kimmel was a scapegoat from the start, but did not suspect that he was a victim of a frameup until about 15 November 1943, and did not have absolute proof until about 18 January 1944. Safford [here he refers to himself in the third person] has overwhelming proof of the guilt of OPNAV and [Army] General Staff, plus a list of about fifteen reliable witnesses.[]
Safford then followed with more questions for Kramer. Then he added a "Comment" section in which he first laid out some of the details of the incident concerning the Winds Execute message and the suppressed follow-up warning message from McCollum:
With regard to the quotes of my item 18 (about the Weather Report") and your items 18 and 10(c) [from Safford's first letter and Kramer's response] you were describing Circular #2492 – the "hidden word" message, of which we have copies of the original and the translation in the GZ files...I was asking about the General Intelligence Broadcast containing [the] false "Weather Report" which was Broadcast at 0430 (EST) on December 4, 1941 or December 5. (Not sure of exact date.) It was heard by "M" [Cheltenham, Maryland] and "W" [Winter Harbor, Maine] and sent in by teletype. It was unheard by "S"....who listened for it. (I have this from the Station "S" files, plus statements of [Lt. Lesley A.] Wright and [Captain Redfield] Mason.) This message (in Morse) included the words – 'Higashi no kazeame.' Nishi no kaze hare (Negative form of kita no kaze kumori.)" The warning was not sent in the manner prescribed by Circular #2353 or #2354, but was a mixture. The [OP-20] GY watch officer was not sure of it so he called you and you came in early and verified it. [Lt. Allan] Murray [a GY watch officer] recalls it so do I. Either you or Brotherhood [another GY watch officer, Lt. Francis.M.] (?) were waiting in my office when I came in that morning and said "Here it is!..."
As a result of the General Broadcast, McCollum [then Commander Arthur] prepared a message – which was a very long message ending up with a translation and significance of the warning in the General Intelligence Broadcast. I read the message in Admiral Noyes' office and was witness to the discussion of it between Noyes and Admiral Wilkinson. I took for granted that the message would be sent and did not know otherwise until 2 December 1943. I believe that I told you about this message and stated that it had been sent. Anyway, I was living in a fool's paradise from 4 December to December 7, 1941. I learned from Wright [Commander Wesley A] that McCollum knew that the message had not been sent (Wright had been informed by McCollum at Pearl Harbor).[]
In the early part of the letter, Safford notes that he did not have "absolute proof" of the "frameup" until about 18 January 1944. Safford does not state explicitly what the source of the proof was, or its nature, but information that came out later that year points probably to one or more conversations he had with William F. Friedman of the S.I.S. as the critical point of origin.
That Safford and Friedman should cross paths in early 1944 was not a surprise. Both had once run their respective services' communications intelligence sections, but had been replaced shortly after the war started. They were familiar with each other's role in establishing the original COMINT programs for both the army and navy. On occasion, they had cooperated by sharing information on code-breaking projects of mutual interest, as with the joint effort against the Japanese Purple machine cipher device. Both now were engaged in issues relating to communications security for the army and navy. The two happened to meet a number of times during 1944.
The discussion of interest was at one of these meetings. When this particular discussion occurred is not clear – perhaps as early as January 1944, but the scant evidence available also suggests sometime later in the year. Friedman remained vague about the exact date and believed it occurred sometime in the eighteen months prior to his testimony before the Clarke Investigations in mid-July 1945, which would place the meeting of interest in the first three months of 1944. Safford in later testimony would not give a date. During the meeting, Safford related to Friedman what he had found so far about the Winds Execute message that he believed had been sent on 4 December and how the intercept subsequently was handled. He also repeated his information about the disappearance from the files of any trace of the translation of the Winds Execute message. Perhaps because he was caught up in the moment of Safford's story, or because he was not aware of the direction or intent of Safford's search, Friedman related a story he had heard regarding the Winds translation. Earlier that year, he had met with Colonel Otis Sadtler, who had served as a communications officer with G-2 at the time of Pearl Harbor. Friedman asked Sadtler if he knew of a Winds Execute message that had been intercepted before 7 December 1941. Sadtler said that he was told that the material had been destroyed. Shocked, Friedman asked who ordered the destruction. Sadtler replied that General George C. Marshall had ordered it. Friedman then told Safford that Sadtler had learned this story from Colonel Isaac (Ike) Spalding, who was head of G-1 Staff (Personnel) at the time. Friedman later testified that he did not give much credence to the story and was surprised that Safford bought into it.[]
Figure 4.2. William F. Friedman
The subject – General Marshall's order to destroy the Winds material – would surface later, first during the Hewitt Inquiry, then the Clarke Investigation, and finally during the Joint Congressional Committee hearings. As for Safford and Friedman, they continued to talk about the missing Winds Execute translation. On one occasion, 17 September 1944, Friedman made notes of an exchange with Safford. [Exhibit #47][] Many of the points that Safford made to Kramer in his letters from nine months earlier were repeated in these notes, notably when the Winds Execute was heard, which station intercepted it, how it was processed and then turned over to ONI for action. Absent, though, were the claims of destruction of intelligence material held by the War Department and that this action had been ordered by General Marshall.