INVESTIGATION OF THE PEARL HARBOR ATTACK -- REPORT OF THE JOINT COMMITTEE ON THE INVESTIGATION OF THE PEARL HARBOR ATTACK

EFFORTS TO MONITOR

The evidence is undisputed that both services extended themselves in an
effort to intercept a message, in execution of the winds code, not only
through their own monitoring stations but through facilities of the
Federal Communications Commission as well. While only fragmentary
evidence of a documentary nature is available to indicate the nature of
instructions to monitor for an implementing or execute message, the
Federal Communications Commission file is complete and, as indicated,
there is no contention that every effort was not made to intercept an
execute message. [9]

CONSIDERATIONS BEARING ON THE POSSIBILITY OF A MESSAGE IN EXECUTION OF  THE "WINDS CODE"  HAVING BEEN RECEIVED PRIOR TO DECEMBER 7, 1941

1. Capt. L. F. Safford in a prepared statement (read before the joint
committee) [10] has set forth a positive assertion that a winds execute
message was received in the Navy Department *on the morning of December
4, 1941,* and has elaborated on the circumstances which serve, in his
opinion, to indicate that a winds execute was dispatched and why such a
message would have been dispatched from Tokyo.

Safford asserted that when he first saw the message it had already been
translated by Kramer; that Kramer had underscored all three "code
phrases" on the original incoming teletype sheet; and that he had
written in pencil or colored crayon the free translation: "War with
England (including NEI, [11] etc.); war with the U. S.; peace with
Russia." Safford has persistently testified that an authentic
implementing message was received.

[8] It is to be noted that this message bears the date December 5, 1941,
whereas the "number group" is 031030, indicating December 3, 1941. From
evidence available (see discussion, infra) it appears this message was
dispatched from Batavia on December 3, 1941, but was not processed in
the Navy Department until December 5, 1941, Inasmuch as the message was
sent "deferred."
[9] See committee record, pp. 9809, 9810.
[10] Id., at pp. 9622-9654.
[11] Netherlands East Indies.

472              PEARL HARBOR ATTACK

2. Capt. A. D. Kramer testified before the committee that *on the
morning of December 5* the GY Watch Officer, thought by him possibly to
be Lieutenant Murray, came to the door of his office and showed him a
message which he, Kramer, regarded as an implementation of the winds
code; that he saw this message only briefly, relying on the evaluation
of the GY watch officer as to the authenticity of the message; that he
had no recollection of writing on the message but that had he written
anything he positively would not have used the word "war"; that he
proceeded to Captain Safford's office with the GY watch officer when the
message was delivered to Safford; that he never saw the message again.
[12]

It should be noted that Kramer testified the message he saw was on a
piece of teletype paper torn off from the machine and was not more than
a line or two, possibly three lines; that in no case did the message
contain some 200 words as alleged by Captain Safford in his statement.
[13] Further, that the message he saw referred to only one country,
which to the best of his belief was England. [14] This testimony must,
of course, be considered along with Kramer's testimony before the Navy
Court of Inquiry. When asked what Japanese language words were used in
the execute message he saw, he replied: [15] "*Higashi No Kazeame*, I am
quite certain. The literal meaning of *Higashi No Kazeame* is East Wind,
Rain. That is plain Japanese language. The sense of that, however, meant
strained relations or a break in relations, possibly even implying war
with a nation to the eastward, the United States."

3. Admiral R. E. Ingersoll testified that during December of 1941 he was
Assistant Chief of Naval Operations; that he saw "messages" which were
supposed to implement the winds code, they being brought to his office;
that he did not recall definitely whether he saw them prior to December
7 or thereafter; that an implementation of the code received prior to
December 7, if genuine, would simply have confirmed what had already
been dispatched to the Fleet regarding destruction of codes by the
Japanese and would have required no action; that he thought the message
he saw referred to all three countries; i. e. England, United States,
and Russia. [16]"

4. Col. Otis K. Sadtler, in charge of the military branch of the Army
Signal Corps in December of 1941, testified that about 9 a. m. or
shortly thereafter on Friday, December 5, Admiral Noyes telephoned him
to the effect that the "message was in" (referring to an implementing
winds message); that Noyes told him "it was the word that implied a
break in relations between Japan and Great Britain"; that he went to
General Miles' office, informing Miles that the "word was in"; that
Miles sent for Colonel Bratton and when Bratton came in, he, Sadtler,
told Bratton word had been received from Admiral Noyes to the effect
that diplomatic relations between Japan and Great Britain were in
danger; that Bratton asked him to verify receipt of the message; that he
called Admiral Noyes again, asking him to verify the "Japanese word" and
Noyes replied that he did not know any Japanese but it was the one that
"meant Japan and Great Britain"; that upon reporting this information to
General Miles'

[12] Committee record, pp. 10481 et seq.
[13] Id., at p. 10491.
[14] Id., at p. 10501.
[15] Navy court of inquiry (top secret) record, p. 957.
[16] Committee record, pp. 11278 et seq.

PEARL HARBOR ATTACK               473

office he did not thereafter get in touch with Admiral Noyes concerning
the message; that he never saw the message Noyes reported to him; and
that insofar as he could ascertain it did "not come over", i. e. to his
office or the Army. [17]

5. Col. Rufus S. Bratton, Chief of the Far Eastern Section of the
Intelligence Branch of the Military Intelligence Division in December of
1941, testified that sometime around 9 or 10 a. m. on the morning of
December 5 he was called to General Miles' office where Sadtler stated
Noyes had just called to say "it is in" (the winds execute message);
that Miles, at his suggestion, requested Sadtler to get from Noyes a
copy either of the Japanese text or of the English translation so a
determination could be made as to whether the message was a genuine
execute or another false alarm; that he did not again see Sadtler
concerning the matter; that he, Bratton, called up the Navy, talking to
either Captain McCollum or Kramer to inquire if they had received a
winds execute message and was advised that no such message had been
received; that he contacted Army SIS [18] and was likewise advised that
no execute had been received; that the Army continued to monitor for an
implementing message up to and after the December 7 attack. [19]

6. Admiral Richmond K. Turner, Chief of War Plans in December of 1941,
testified before the committee as follows:

"On Friday afternoon, I think it was, of December 5, Admiral Noyes
called on the telephone or the interphone, I do not know which, and said
"The weather message", or words to this effect, "the first weather
message has come in" and I said, "What did it say?" And he said, "North
wind clear." And I said, "Well, there is something wrong about that,"
and he said, "I think so, too", and he hung up

"I never saw a draft of that, I do not know from my own knowledge where
he got it from. I assumed until recently that it it was an authentic
message. From what I can determine since coming back here it was
something entirely different, but it was never told to me. If it had
come in and had been authentic I am certain that I would have received a
copy of it."

Turner testified that he did not see an implementation of the winds code
applying to the United States. [20]

7. To complete the picture it would seem apropos to set forth the
testimony of Rear Adm. Leigh Noyes at this point.

Noyes, in December of 1941, Director of Naval Communications, testified
before the committee that prior to December 7, 1941, no genuine winds
execute message was brought to him of to his attention by anyone in the
Navy Department; that prior to the Pearl Harbor attack there were
several instances when messages were brought to him which were first
thought to be winds execute messages but were determined not to be
genuine; that the message described by Captain Safford in his statement,
if received, would not have been regarded as an authentic execute
message since (1) it is alleged to have been in Morse code and not by
voice (2) no provision was made for a negative expression in the winds
code (3) an execute would not have been

[17] Id., at pp. 12357-12363.
[18] Signal Intelligence Service.
[19] Committee record. Pp. 12068-12077:

Colonel Bratton testified: "I can state most positively that no execute
of the winds codes was ever received by me prior to the attack on Pearl
E[arbor. I find it hard to believe that any such execute message could
get into the War Department without passing over my desk.

"It is inconceivable to me. I might have missed it but I had some
assistants who were on the watch for it, and there were some people in
the Army SIS who were also on the watch for it. They couldn't all have
missed it. It is simply inconceivable to me that such a message could
have been in the War Department without some one of us knowing about it
or seeing it." Committee record, p. 12089.

[20] Committee record, p. 5214.

474             PEARL HARBOR ATTACK

interpreted to mean war, and (4) Circular 2353 made no provision for N.
E. I. as stated by Safford.

With respect to Colonel Sadtler's testimony that Noyes called him saying
"The message is in," or words to that effect, Noyes stated he had no
present recollection of having made such a statement although he would
not say it did not occur inasmuch as he talked with the chief signal
officer a number of times each day. [21]

Further, Noyes testified that he was directed to prepare a folder for
the Roberts Commission but that it did not include a winds execute
message and the folder in fact was supposed to contain no magic nor any
reference to it; that the McCollum message, [22] to his knowledge,
contained no reference to a winds execute message. [23]

8. The "Rochefort Message."

On December 5, 1941, a dispatch signed "Miles" was sent by the War
Department to the assistant chief of staff headquarters G-2, Hawaiian
Department, as follows: [24]

"Contact Commander Rochefort immediately thru Commandant Fourteen Naval
District regarding broadcasts from Tokyo reference weather."

At first blush, the foregoing dispatch would suggest, inferentially at
least, the possibility of an execute message having been received.
Colonel Bratton, upon whose recommendation the dispatch was sent,
testified, however: [26]

"I had a discussion with Commander McCollum, now Captain McCollum, as to
the amount of knowledge that the Navy had in Hawaii. He assured me his
man Rochefort there at that time knew practically everything that there
was to be known about the U. S.-Japanese relations through one means or
another. I knew that suitable warning messages had been sent out to
Hawaii and elsewhere. I had not read the messages and did not know their
exact contents. I wanted to make sure that our G-2 in Hawaii got in
touch with the ONI man in Hawaii, to get from him all the intelligence
that he had in his possession, and I knew that if they got together on
the subject of this winds message I did not know, but I felt that they
were going from there, and that there would be a complete exchange of
intelligence and that the Army G-2 would then be in possession of just
as much intelligence as Rochefort, the ONI man, had."

Colonel Bratton's testimony is to the effect that the dispatch of the
message to G-2 to contact Rochefort had nothing whatever to do with
receipt of a message in execution of the winds code. In this regard
Captain McCollum stated: [26]

"I understood that G-2 was very anxious for their G-2 in Hawaii to have
direct access with Commander Rochefort, who had the only agency capable
of intercepting the winds message in Hawaii, sir. The Army, as I
understand it, had no parallel set-up in Hawaii at that time."

[21] In a statement submitted to the committee under date of February 25
1946, in amplification of his testimony Admiral Noyes said: "In reading
over my testimony I noted that I failed to bring out the following
point, which, however, is supported by my previous testimony and by
documentary evidence.

"In connection with the alleged telephone conversation with me on 5
December to which Colonel Sadtler testified and which I did not recall
in that form:

"On 5 December there was received from Colonel Thorpe in Batavia
addressed to General Miles in the War Department. This message was
transmitted by the Naval Attache to Nary Department for delivery to
General Miles. As I have already testified the subject matter was under
discussion between me and the War Department during that day. It is very
probable that I would have called Colonel Sadtler and notified him of
the fact that this message had been received and was being delivered to
the War Department for General Miles on account of its importance. Since
discussion took place between me and the War Department during that day
on the subject matter of this message and the War Department recommended
that we should make no change in our original translation of the set-up
of the Winds Code (see previous testimony), it would appear that any
possible authentic or false execute of the winds message would have also
been discussed and settled during that day."  Committee record  pp.
14101, 14102.

[22] See discussion, infra
[23] Committee record, pp. 12605-12620.
[24] Committee exhibit No. 32 p 20.
[25] Committee record p. 12120, 12121.
[26] Id., at pp. 9271, 9272.

PEARL HARBOR ATTACK            475

CONSIDERATIONS MILITATING AGAINST LIKELIHOOD OF "WINDS CODE" EXECUTE
MESSAGE HAVING BEEN RECEIVED PRIOR TO DECEMBER 7, 1941

1. Examination of Circular 2353 (to which Captain Safford admits the
alleged winds execute was responsive) reflects that an execute warning
would be added in the middle and at the end of the daily Japanese
language short wave news broadcast "in case of emergency (danger of
cutting off our diplomatic relations), and *the cutting off of
international communications.*" When the execute was heard "all code
papers, etc." Were to be destroyed.

A reasonable construction of this circular would indicate that the winds
code was an emergency arrangement designed to be employed in the event
ordinary commercial means of international communications were no longer
available to the Japanese Government. Contemplating that such commercial
means conceivably might not be available to her, it would appear natural
that Japan should devise a means such as the winds code to direct her
diplomatic establishments to destroy their codes and secret papers.
Manifestly and quite naturally the winds code should provide for
destruction of all code papers inasmuch as the necessity for having any
codes whatever of the type outstanding would be precluded by the
cutting-off of international communications.

Ordinary commercial means of communications were available to Japan up
to the December 7 attack on Pearl Harbor and in fact committee exhibit 1
is replete with instructions to Japanese diplomatic establishments with
respect to destruction of codes. [27] Accordingly, it can fairly be
concluded that recourse to the emergency system provided by the winds
code was not necessitated and in consequence was not resorted to prior
to December 7 inasmuch as the contingency contemplating its use (cutting
off of international communications) did not materialize prior to the
Pearl Harbor attack.

2. It is admitted and of course definitely known that a winds execute
message (*Nishi No Kaze Hare* west wind, clear) applying to England was
transmitted from Tokyo stations JLG4 and JZJ between 0002 and 0035 GMT,
December 8, 1941. [28] Such a message was of course reasonable inasmuch
as Japan could very well contemplate that ordinary commercial means of
communications would no longer be available after the Pearl Harbor
attack.

Inasmuch as a genuine winds execute message applying to England was
transmitted after the Pearl Harbor attack, it would appear anomalous
that such a message should also have been sent prior to December 7. [29]

3. The investigation conducted in Japan by headquarters of the supreme
allied commander reflected that a signal implementing Circulars 2353 and
2354 was probably not transmitted prior to December 8, Tokyo time but
was transmitted by radio voice broadcast at some hour after 0230,
December 8, Tokyo time. [30] No evidence could be obtained that an
implementing signal was transmitted by radio telegraph. Significantly,
those who conducted the interrogation

[27] See sections relating to destruction of codes pt. III and IV this
report.
[28] See committee exhibit No. 142.
[29] Admiral Noyes suggested that Japan s sending an execute on December
7 was probably occasioned by reason of the fact that some Japanese
diplomatic establishment had failed to respond to instructions to
destroy their codes which had been dispatched through ordinary channels
of communication.
[30] December 7, Washington time.

476             PEARL HARBOR ATTACK

in Japan had no knowledge prior to the interrogation that the United
States had information that the winds code was used on December 8, Tokyo
time. [31]

Mr. Shinroku Tanomogi was head of the overseas department of the Japan
Radio Broadcasting Corporation in December 1941, and as such was in
charge of programs, including news programs, beamed to foreign
countries. Upon interview he stated he had no recollection at all of any
"east wind rain" report or any similar phrase being broadcast. Prior to
December 8. [32]

4. Inquiry made through the State Department reflects that no winds
execute message was intercepted prior to the Pearl Harbor attack by the
British, Dutch, or Australians. [33]

5. In his statement submitted for the committee's consideration, Captain
Safford definitely states that the alleged implementing winds message
was part of a Japanese overseas "news" broadcast from station JAP
(Tokyo) on 11980 kilcoycles beginning at 1330 Greenwich civil time on
Thursday, December 4, 1941, this time corresponding to 10:30 p. m.,
Tokyo time, and 8:30 a. m., Washington time, December 4, 1941; that the
winds message broadcast was forwarded by teletype from Cheltenham to the
Navy Department shortly before 9 a. m. On December 4, 1941. Further,
that when he first saw the message it had already been translated by
Kramer; that Kramer had underscored all three "code phrases" on the
original incoming teletype sheet; and that he had written in pencil or
colored crayon the following free translations:

"War with England (including NEI, etc.)
"War with the U. S.
"Peace with Russia."

Kramer has testified that had he seen such a message, as alleged by
Safford, he would in no case have interpreted a winds execute to mean
war. [34]

In this regard, the Thorpe and Foote messages, which interpreted the
winds code as meaning war, were not available to the Navy Department
until after the time Safford alleges the winds execute came in and was
interpreted by Kramer to mean war. The Thorpe dispatch, while intended
for General Miles of the War Department, was sent by Naval
Communications and was received at the Navy Department at 1:21 a. m.,
December 4, l941. [35] It was not decoded until 1:45 a. m., December 5,
1941, the delay being occasioned by the fact that the dispatch was sent
"deferred," the lowest priority in handling. [36] The Foote dispatch, it
is to be noted, was not received in the State Department until 9:19 a.
m., December 4. Consequently, as indicated, no information was available
in the Navy Department on the morning of December 4 as alleged by
Safford serving as basis for interpreting a winds execute message to
mean war. Even conceding the availability of the Thorpe and Foote
dispatches, it would scarcely appear likely that the Navy Department
would disregard its own translation of the winds code and be guided
solely by the dispatches from outside sources.

[31] See committee exhibit No. 142
[32] Id., sec 4B.
[33] Committee exhibit No. 142, secs. 4c, 4d, 4e. See also committee
record, p. 11564.
[34] See Navy Court of Inquiry (top secret) record, pp. 968, 969, 975,
987; committee record, 10492.
[35] Committee record, p. 10135.
[36] Id., at pp. 11255,11256.

PEARL HARBOR ATTACK            477

6. The winds execute message Safford alleges he saw on the morning of
December 4, bore the "negative form for war with Russia" and mixed up
the plain language broadcast with the Morse broadcast. [37] It is thus
clear that the alleged winds execute of December 4 was not responsive to
the establishing winds code.

Captain Kramer, it should be noted, testified before the joint committee
that had the "negative form" been employed with respect to Russia, he
would have regarded such fact as nullifying any credence to be placed in
a broadcast purporting to be a winds execute message. It would appear
agreed that the implementation of an establishing code must conform in
meticulous detail to the code as originally established.

7. Referring to Captain Safford's statement, the following matters
appear to be subject to serious question:

A. Safford relies on Cincaf 281430 [38] as basis for evaluation of a
winds execute message to mean war, pointing out that this dispatch
contained the statement "Nishi nishi England *including occupation of
Thai or Invasion of Malay and N. E. I.*"

It should be noted, however, that Cincaf 281430 indicates the winds code
would be employed "if diplomatic relations are on verge of being
severed." In any event the interpretation of Cincaf 281430 as relied
upon by Safford while possibly indicating war with England does not by
any reasonable construction indicate war with the United States. [38a]

B. Safford's reliance in his statement on Cincaf 281430 as providing
basis for evaluating a winds execute as meaning war is in contradiction
of his testimony before the Navy Court of Inquiry where reliance was
placed on the Thorpe and Foote dispatches. [39]

While denied by Safford, the suggestion was made by counsel before the
committee that Safford may have shifted reliance on the Thorpe and Foote
dispatches to Cincaf 281430 by reason of the fact that he had learned
that both the Thorpe and Foote dispatches were not available to the Navy
Department until after the morning of December 4. [40]

C. Safford seeks to bring out that the alleged winds execute was
intended for the Japanese London Embassy inasmuch as the latter had
destroyed its codes 3 days previously and a winds message was the only
way that Tokyo could get news to its London Ambassador secretly. [41]

This statement is not true insofar as it implies that no other means of
communication between Tokyo and London was available. By Circular 2409
of November 27, 1941, [42] the Japanese established the "hidden word"
code and by Circular 2461 [43] instructed that this code be kept

[37] That is, Circular 2363 with Circular 2354.
[38] Set forth, supra.
[38a] See committee record, p. 9670.
[39] Navy Court of Inquiry (top secret) record, p. 748, see also
committee record, p. 9667
[40] Committee record, pp. 9667, 9668.
[41] Id., at p. 9639.
[42] Committee exhibit No. 1, p. 186.
[43] Id., at p. 226.

478              PEARL HARBOR ATTACK

until the last moment. This code system of communication was clearly
available to the Japanese in communicating with their London Ambassador
and was in fact employed on December 7 in Circular 2494. [44] Safford
admitted in his testimony before the joint committee the availability in
the London Embassy of the hidden word code.

Furthermore, in Circular 2443, dated December 1, [45] to London
instructions were issued to discontinue use of the code machine and to
dispose of it immediately. Ostensibly other code systems were still
available after destruction of the code machine and it is known that
coded traffic in the system referred to as PA-K2 passed from the
Japanese London Embassy to Tokyo December 6, 1941. [46]

D. By way of lending credence to his assertion that a winds execute was
received, Safford has testified that McCollum's dispatch of December 4
(not sent) was predicated on such a winds execute and mentioned the
execute in the last portion. [47]

McCollum definitely contradicted this in testifying before the
committee, asserting that his dispatch was based on a memorandum he,
McCollum, had prepared under, date of December 1 [48] and bore no
relationship to a winds execute message; that he neither saw nor
received knowledge of a true winds execute prior to December 7. [49]

E. In further substantiation of his allegation that a winds execute was
received on the morning of December 4, Safford has referred to the fact
that the dispatches from OpNav to our own establishments to destroy
their codes was based on a winds execute.

This assertion is diametrically contrary to testimony of Noyes [50] and
Kramer [51] who declared that OpNav instructions to our establishments
to destroy their codes was based on instructions sent out by the
Japanese [52] to their diplomatic establishments to destroy codes, and
bore no relationship to a winds execute. The testimony of McCollum and
Ingersoll tends to confirm the foregoing.

F. Safford points out that the individual smooth translations of the
alleged winds execute for authorized Navy Department officials and the
White House were distributed at noon on December 4, 1941, in accordance
with standard operating procedure. [53]

Kramer, in testifying before the joint committee, categorically denied
that any copies of a winds execute message were prepared for
distribution by his section, it

[44] Id., at p. 251
[45] Id., at p. 209.
[46] Committee record, p. 9740.
[47] See pt. IV, this report, for discussion of so-called McCollum
dispatch.
[48] Committee exhibit No. 81.
[49] Committee record pp. 9124-9134.
[50] Id., at p. 12623.
[51] Id., at p. 10504.
[52] Committee exhibit No. 1.
[53] Committee record, pp. 9763 et seq.

PEARL HARBOR ATTACK            479

being noted that it was the responsibility of Kramer to prepare and
distribute the smooth translations. [54]

G. Captain Safford has pointed out that a winds-execute was dispatched
in Morse code. Captain Rochefort, who was in charge of the
Communications Intelligence Unit at Pearl Harbor in December of l941,
testified that all of the broadcast schedules giving the various
frequencies furnished by Washington were all voice frequencies; that to
him the very setting up of the winds code implied "voice"; that if an
execute message were sent in Morse code it would have meant that every
Japanese Embassy (and consulate) in every Japanese location throughout
the world for whom the message was intended by the Japanese Government
would "have had to maintain Morse code operators, people capable of
receiving Morse code. I do not think so." [55]

Rochefort further testified that they were monitoring for a winds
execute message at Honolulu and continued to do so until after the
attack; that four of his best language officers were on a 24-hour watch
for an execute; that no winds implementing message was intercepted. [56]

H. Admiral Noyes testified that he would not have regarded the message
which Safford alleges was received as an authentic execute message
inasmuch as (1) Morse code was allegedly used and in consequence not
responsive to Circular 2353; (2) no provision was made in the winds code
for a "negative form" with respect to Russia; (3) an execute message
would not have been interpreted to mean war; and (4) no reference is
made in Circular 2353 to N. E. I., although the alleged execute was
responsive to Circular 2353 and Safford indicates reference was made to.
N. E. I. [57]

8. Safford, in testifying before the joint committee, placed emphasis on
the fact that the winds code provided for destruction of all codes
(Circular 2353) and by reason thereof a winds execute message would have
more significance than the intercepts contained in committee exhibit 1
which gave instructions with respect to destruction of particular codes.
[58]

If a winds execute message was dispatched for the Japanese London
Embassy on December 4, as alleged by Safford, it would necessarily

[54] Committee record, p. 10496.
[55] Id., at p. 12548.
[56] Id., at pp. 12532-12534.
[57] Id., at pp. 12614, 126l5.
[58] When asked what there was in the winds execute message alleged by
him to have been received which indicated *war*, Captain Safford
testified: "For one thing there is instruction to destroy all code
papers. If that is regarded as synonymous with the outbreak of war, as I
have heard testified in this room, that by itself means something more
than the wording of these three paragraphs above * * *. Tokyo had sent
out instructions to various people telling them to burn their most
important codes but to leave two codes open. One was the so-called PA-K2
code and the other was the LA code. Now, with those two exceptions all
codes had been burnt, but this said, '*Please destroy all code papers*,'
and so forth. In other words, there was no exceptions in this one."
Committee record, p. 9778.

In marked contradiction of the foregoing testimony is the explanation of
Captain Safford as to the reason for Japan's London Embassy having the
PA-K2 code system after the alleged winds execute message was received.
He stated: "There were two systems that were exempt from destruction.
One was PA-K2, and the other was LA, neither of which were considered by
ourselves as secret, and we presumed the Japanese did not consider them
secret." Committee record, p. 9741.

It is to be noted, however, that the Honolulu consulate, as well as
Tokyo, used the PA-K2 system for some of the most vital messages shortly
before December 7 (see committee exhibit No. 2). While this was
virtually the only system left after the messages ordering the
destruction of various codes, the PA-K2 system was employed for the
sending of messages which would probably have tipped off the attack on
Pearl Harbor, had it not been for the fact they were not translated
until after the attack.

480             PEARL HARBOR ATTACK

mean that *all* codes were to be destroyed by Japan's London Ambassador.
It is definitely known, as earlier indicated, that London sent a
dispatch to Tokyo in the system known as PA-K2 on December 6, 1941. Such
fact would indicate strongly that no winds execute was dispatched on
December 4 with consequent destruction of all codes. [59]

9. It appears clear that both the Navy and Army were still looking for a
winds execute message after the morning of December 4, based on records
of the Federal Communications Commission. [60]

In this connection at 7:50 p. m. on December 5, 1941, the watch officer
of FCC phoned Colonel Bratton of the Army with respect to a false winds
message received from the FCC Portland monitoring station. The FCC watch
officer submitted the following memorandum for his superior with respect
to Bratton's remarks:

"Remarks by Col. Bratton:

"Results still negative but am pleased to receive the negative results
as it means we have that much more time. The information desired will
occur in the middle of a program and possibly will be repeated at
frequent intervals. (Asked Col. Bratton if I should communicate the
information to Portland concerning the fact that the desired data will
be in the middle of a program.) No, I will have a conference with Lt.
Col. Dusenberg in the morning and will contact Mr. Sterling in that
regard."

The foregoing would indicate that the Army had received no genuine winds
execute message by 7:50 p. m., December 5.

The FCC night watch log for December 4, 1941, [61] contains the notation
that at 9:32 p. m. "Lt. Brotherhood called to inquire if any other
reference to weather was made previously in program intercepted by
Portland. Informed him that no other reference was made." There is
manifested here an interest by the Navy in the nature of a winds message
on the evening of December 4 which is hardly likely if a true execute
was received on the morning of December 4.

Further, it would appear logical that had a true winds execute been
received on the morning of December 4 the FCC would have been requested
to discontinue its monitoring activities. This, however, was not done
and the FCC was still monitoring for a winds execute and actually
intercepted such an execute (with respect to England) after the Pearl
Harbor attack. [62]

10. Collateral considerations tending to minimize likelihood that
implementing winds message was dispatched from Tokyo.

A. Referring to the message telephoned by the FCC to Brotherhood at 9:05
p. m. on December 4,63 Safford testified before Admiral Hewitt [64] that
this was the "false" message which appeared on this surface to use the
"winds" code relating to Russia but which was a genuine weather
broadcast. This message, Safford said, Brotherhood telephoned to Admiral
Noyes and later *Kramer took one look at it and said it was not what was
wanted and threw it into the waste basket*. He testified that this
message was received * * * 12 hours or more after what he referred to as
the "true winds message."

[59] Committee record, p. 9740.
[60] Committee exhibit No. 142-A.
[61] Id.
[62] See also testimony of Colonel Bratton, committee record, p. 12074.
[63] Committee exhibit No. 142, see. 3.
[64] Hewitt inquiry record, p. 113.

PEARL HARBOR ATTACK            481

Query: Why would Kramer be "wanting" a winds execute message 12 hours
after Safford alleges Kramer had an execute message and had noted
thereon "War with England, War with U. S., peace with Russia"?

B. In testifying before the committee, Justice Roberts stated he had no
knowledge of the winds matter and no access to Magic. This would appear
to be partially at least in contradiction of Safford's testimony that he
last saw the winds execute among material assembled for the Roberts
Commission.

Further, Admiral Noyes testified that he was directed to prepare a
folder for the Roberts Commission, but it did not include a winds
execute message and the folder was in fact supposed to contain no magic
nor any reference to it. [65]

C. Safford's detailed recollection of the winds matter, as set forth in
his statement, is in sharp conflict with his indefinite and somewhat
nebulous memory as reflected by his testimony and the letters directed
to Kramer during December 1943, and January 1944.

It should be noted in this connection that Safford testified before
Admiral Hart [66] that the winds implementing message came in on the
evening of December 3 and Kramer went down to get it. From all of the
testimony it appears that Safford's position before the committee was
assumed after a process of elimination of possibilities and
reconstruction of a situation concerning which he had only a partially
independent recollection.

D. Considering the tight reign maintained by the military in Japan and
particularly the desire to clothe the movement against Pearl Harbor with
utmost secrecy, it would seem highly improbable that the Japanese would
tip off her war decision in a news broadcast by advising her London
Ambassador of such decision 3 days before Pearl Harbor.

E. If a true winds execute was received and distributed on December 4 it
would appear reasonable to assume that some record of the message could
be found in the War or Navy Departments. Yet despite repeated searches
there is no record whatever in either department of such a message. In
this connection Safford has suggested that intercept No. JD-7001, marked
"canceled" in the Navy file of intercepts, may have been the missing
winds execute. Such a premise, of course, presupposes a deliberate
abstraction by someone of an official record from the Navy Department.

In evaluation of Safford's suggestion with respect to No. JD-7001, it
should be noted that the file of JD intercepts was maintained by Kramer
who has emphatically testified that no winds execute came into his
section or was distributed by him. Further, Kramer has pointed out that
there are several examples of canceled JD numbers in the file [67] and
presented several reasons in testi-

[65] Committee record. P. 12620.
[66] Hart inquiry record, p. 361.
[67] This appears to be borne out by the record. See committee exhibit
No. 142, sec. 6.

482             PEARL HARBOR ATTACK

fying before the committee why a JD number might be canceled.

Significantly, a check of the Army file of intercepts for the period
December 3-5, 1941, reflected that the Navy file contains all intercepts
that are in the Army file. [68]

Conceding for purposes of discussion that a winds execute message was
received in the form alleged by Safford, it will be noted that such
message would not indicate where or when Japan would strike but merely
her possible purpose to go to war. Bearing in mind the rather frank
admission by Army and Navy officials that they knew war was imminent in
the days before December 7, credence could scarcely be placed in the
theory that the message was deliberately destroyed when it contained no
information that was not admittedly already possessed.

Admiral Ingersoll, for example, testified before the committee that had
a true winds execute message been received it would have been regarded
as merely confirmatory of the implications contained in Japanese
instructions to destroy codes contained in committee exhibit 1, inasmuch
as instructions to destroy codes, particularly in the consulates, meant
war. The testimony of several other witnesses, including Admiral Noyes
and Colonel Bratton, is to the same effect.

11. The testimony of Col. Robert E. Schukraft, assigned to the office of
the chief signal officer at the time of the Pearl Harbor attack, before
the committee on February 19, 1946, is of particular pertinence to the
testimony of Captain Kramer, set forth under section 3, supra. Schukraft
testified that 2 or 3 days prior to Pearl Harbor Col. Rex Minckler
brought to Schukraft's office a piece of yellow teletype paper (the
carbon copy) which contained what appeared to be a winds execute message
but that the message upon examination was obviously not a true winds
execute. Further, Schukraft testified Colonel Minckler had indicated
that the Navy had thought the message a true winds execute, Captain
Kramer having seen the message and so thinking. He stated that he
concluded very positively that the message was not a true execute of the
Winds Code. [69]

12. The following officers have stated they have no knowledge of a
message in execution of the Winds code prior to December 7, 1941:

Navy

Admiral Harold R. Stark, Chief of Naval Operations. [70]
Admiral Leigh Noyes, Director of Naval Communications. [71]
Admiral T. S. Wilkinson, Director of Naval Intelligence. [72]
Capt. Arthur N. McCollum, in charge, Far Eastern Section of Naval
Intelligence. [73]
Admiral Joseph R. Redman, Assistant Director of Naval Communications.
[74]

[68] See Army liaison memorandum dated January 26, 1946. Committee
record, pp. 8965, 8966.
[69] Committee record pp. 13093-13096.
[70] See Navy Court of Inquiry record, pp. 783 872. Confirmed in
testimony before the committee.
[71] Committee record, pp. 12605-12620.
[72] Hewitt inquiry record, pp. 398-401.
[73] Committee record, pp. 9124-9134.
[74] Navy Court of Inquiry record, p. 1103.

PEARL HARBOR ATTACK            483

Lt. Comdr. George W. Linn, GY watch officer. [75]
Lt. Comdr. Alfred V. Pering, GY watch officer. [76]
Lt. Comdr. Allan A. Murray, GY watch officer. [77]
Lt. Frederick L. Freeman, assigned to section disseminating to ONI
intelligence received from radio intelligence units. [78]
Capt. Redfield Mason, fleet intelligence officer, Asiatic Fleet. [79]
Commander Rudolph J. Fabian, Radio Intelligence Unit at Corregidor. [80]
Capt. Edwin T. Layton, Pacific Fleet intelligence officer. [81]
Capt. Joseph John Rochefort, in charge, Communications Intelligence
Unit, Pearl Harbor. [82]

Army

Gen. George C. Marshall, Chief of Staff. [83]
Maj. Gen. Leonard T. Gerow, Chief of War Plans. [84]
Maj. Gen. Sherman Miles, Chief of G-2. [85]
Col. Rufus W. Bratton, Chief, Far Eastern Section of G-2. [86]
Col. Robert E. Schukraft, Chief, Radio Interception for SIS. [87]
Col. Rex W. Minckler, Chief, SIS. [88]
Brig. Gen. Thomas J; Betts, executive assistant to the Chief of
Intelligence Branch MID. [89]
Lt. Col. Frank B. Rowlett, prior to Pearl Harbor attack a civilian
technical assistant to the officer of the Cryptoanalytic unit, SIS. [90]
William F. Friedman, a cryptanalyst of War Department. [91]

Over-all observations with respect to Captain Safford's testimony:

13. As previously indicated Captain Safford has rather consistently
testified that a true winds execute message was received prior to
December 7. However, there are certain discrepancies in his testimony
tending to show particularly that his recollection of the incident
attending receipt of such an execute has not been definite and has been
developed through a process of elimination.

A. The following testimony, in relation to a winds execute, of Captain
Safford before Admiral Hewitt reflects rather clearly his indefinite
recollection of the winds matter and his efforts to reconstruct a "vague
memory": [92]

"Captain SAFFORD. In the fall of l943 it appeared that there was going
to be a trial or court martial of Admiral Kimmel. It was hinted in the
newspapers and various people in the Navy Department were getting
testimony ready for it. I realized I would be one of the important
witnesses that, my memory was very vague, and I began looking around to
get everything that I could to prepare a written statement which I could
follow as testimony. That was the time when I studied the Robert's
Report carefully for the first

[75] Hewitt Inquiry record pp. 140-142.
[76] Id., at p. 148.
[77] Id., at pp. 433-441.
[78] Id., at pp. 149, 150.
[79] Id., at pp. 73, 78.
[80] Id., at pp. 73, 78.
[81] Id., at pp. 269-271.
[82] Id., at pp. 46, 48.
[83] See Army Pearl Harbor Board (Top secret) record pp. 35-39.
Confirmed in testimony before the committee.
[84] Committee record. p. 4302.
[85] See Clausen investigation record pp. 214, 215. Confirmed In
testimony before the committee.
[86] Committee record pp. 12068-12077.
[87] Id., at pp. 13093-13096.
[88] Clausen Investigation record p. 217.
[89] Id., at p. 194.
[90] Id., at pp. 225, 226.
[91] Hewitt Inquiry record pp. 515-520.
[92] Id., at pp. 112, 113.

484               PEARL HARBOR ATTACK

time and noted no reference to the winds message or to the message which
McCollum had written and which I had seen and I thought had been sent.
And then I began talking to everybody who had been around at the time
and who I knew had been mixed up in it to see what they could remember
to straighten me out on the thing and give me leads to follow down to
where I could put my hands on official messages and things so that it
would be a matter of fact and not a matter of memory. I also talked the
thing over with whatever Army people were still around at the time and
had anything in this thing, and bit by bit these facts appeared to come
together. The investigation was conducted, if you call it that, for the
purpose of preparing myself to take the stand as a witness in a
prospective court martial of Admiral Kimmel."

E. The letters directed to Captain Kramer by Safford and incorporated in
the committee transcript also indicate an indefinite recollection of
events prior to the attack on Pearl Harbor. [93]

C. In testifying before Admiral Hart, Safford stated: [94]

"The "Winds Message" was actually broadcast during the evening of
December 3, 1941 (Washington time), which was December 4 by Greenwich
time and Tokyo time. The combination of frequency, time of day, and
radio propagation was such that the "Winds Message" was heard only on
the East Coast of the United States, and even then by only one or two of
the Navy stations that were listening for it. The other nations and
other Navy C. I. Units, not hearing the "Winds Message" themselves and
not receiving any word from the Navy Department, naturally presumed that
the "Winds Message" had not yet been sent, and that the Japanese
Government was still deferring the initiation of hostilities. When the
Japanese attacked Pearl Harbor, the British at Singapore, the Dutch at
Java and the Americans at Manila were just as surprised and astonished
as the Pacific Fleet and Army posts in Hawaii. It is apparent that the
War Department, like the Navy Department failed to send out information
that the "Winds Message" had been sent by Tokyo. The "Winds Message" was
received in the Navy Department during the evening of December 3, 1941,
while Lieutenant (j. g.) Francis M. Brotherhood, U. S. N. R., was on
watch. There was some question in Brotherhood's mind as to what this
message really meant because it came in a different form from what had
been anticipated. Brotherhood called in Lieutenant Commander Kramer who
came down that evening and identified that message as the 'Winds
Message" we had been looking for."

Yet in his statement and in testifying before the committee Safford has
the message coming in on the morning of December 4, 1941, it being
brought to him by Lt. A. A. Murray.

D. In testifying before the Navy Court of Inquiry Safford said: [95]

"22. Q. Captain, in a previous answer you stated that the copy of the
intercept using the winds code which you saw on the morning of 4
December 1941 indicated a break in diplomatic relations between the
United States and Japan and Japan and Great Britain, and war between
these nations. Was there anything in the establishment of the code
originally which would indicate that a use of that code would indicate
war as contrasted with a mere break in diplomatic relations?

"A. The Dutch translation said "war." The Japanese language is very
vague and you can put a number of constructions or interpretations or
translations on the same message. In very important documents it was
customary for the Army and Navy to make independent translations and the
differences were sometimes surprising; that

[93] See testimony of Captains Kramer and Safford before the committee.
[94] Hart inquiry record, p. 361.
[95] Navy Court of Inquiry record, p. 748.

PEARL HARBOR ATTACK            485

is, a difference in degree. The general facts would be alike. However,
the people in Communication Intelligence and the people in Signal
Intelligence Service and the people in the Far Eastern Section of Naval
Intelligence, as well as the Director of Naval Intelligence, considered
that meant war and it was a signal of execute for the Japanese war
plans.

"23. Q. Captain, I call your attention again to Document 3 in Exhibit 64
which is an English-language translation of the Dutch intercept. Was
this your only source of information that the use of this code would
indicate "a war decision" which is the wording used by the attach in
Batavia?

"A. Mr. Foote's message to the State Department was even more specific.
It said, "When crises leading to worst arises following will be
broadcast at end of weather reports. 1. *East wind rain* war with United
States. 2. *North wind cloudy* war with Russia. 3. *West wind clear*-war
with Britain, including an attack on Thailand or Malaya and Dutch East
Indies." This was apparently a verbatim quotation from the Dutch
translation."

Significantly, in testifying before the committee Safford relies on
Cincaf 281430 as the dispatch serving as basis for interpreting a winds
execute message to mean war. It has now been conclusively shown that
neither the Foote nor Thorpe dispatches were available in the Navy
Department at the time Safford alleges an execute was received and
interpreted to mean war; i. e. The morning of December 4, 1941. [96]

E. The testimony of Captain Safford taken in its entirety reflects
substantial discrepancies as to where the alleged execute message was
received. It was only at the time of submitting his statement to the
committee that Safford stated definitely the message came in at the
Navy's Cheltenham station.

14. Because of substantial discrepancies in testimony given in prior
proceedings with respect to the question of whether a winds execute
message was received in the War or Navy Department, the inquiry
conducted by Admiral Hewitt went fully into the matter, among others, of
determining if such a message was intercepted prior to December 7, 1941.
Admiral Hewitt found: [97]

"The interception of a "winds" message relating to the United States
during the first week of December 1941, would not have conveyed any
information of significance which the Chief of Naval Operations and the
commander in chief, Pacific Fleet, did not already have.

"No message in the "winds" code relating to the United States was
received by any of the watch officers in the Navy Department to whom
such a message would have come had it been received in the Navy
Department. No such message was intercepted by the radio intelligence
units at Pearl Harbor or in the Philippines although intensive efforts
were made by those organizations to intercept such a message. The
evidence indicates further that no such message was intercepted by the
British or the Dutch, despite their efforts to intercept such a message.
Neither the Fleet Intelligence Officer of the Asiatic Fleet nor the
Fleet Intelligence Officer of the Pacific Fleet nor the Intelligence
Officer of the Far Eastern Section of the Office of Naval Intelligence,
recalled any such message. The Chief of Naval Operations, the Director
of Naval Communications, and the Director of Naval Intelligence recalled
no such message. Testimony to the effect that a "winds" code message was
received prior to the attack was given by Captain Safford, in charge of
Op-20-G, a communications security section at the Navy Department, who
stated that such a message was received on December 3rd or 4th, that it
related to the United States, and that no copy could be found in the
Navy or Army files. In his testimony before Admiral Hart, Captain
Safford named, in addition

[96] See in this connection, committee record, pp. 9667, 9668.
[97] For Hewitt Inquiry report, see committee exhibit No. 157.

486             PEARL HARBOR ATTACK

to himself, three other officers who, he stated, recalled having seen
and read the "winds" message. Each of those officers testified that he
had never seen such a message. The only other testimony to the effect
that a "winds" message was received was by Captain Kramer, an
intelligence officer assigned to Op 2O-G, who said that he recalled that
there was a message but that he could not recall whether or not it
related to the United States or England or Russia. It may be noted that
until he testified in this investigation, Captain Kramer erroneously
thought that a "hidden word" message intercepted on the morning of
December 7th had been a "winds" message."

CONCLUSION: From consideration of all evidence relating to the winds
code, 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
Department prior to December 7, 1941. It appears, however, that messages
were received which were initially thought possibly to be in execution
of the code but were determined not to be execute messages. In view of
the preponderate weight of evidence to the contrary, 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.

Go to Next Page