h022113406251456, służby specjalne w sm
[ Pobierz całość w formacie PDF ]
Cryptologia, 30:98–150, 2006
Copyright
Taylor & Francis Group, LLC
ISSN: 0161-1194 print
DOI: 10.1080/01611190500428634
The Zimmermann Telegram Revisited: A
Reconciliation of the Primary Sources
PETER FREEMAN
Abstract A critical examination of the primary sources (some published here
for the first time) on the transmission, interception and decryption of the
Zimmermann Telegram dispels some long-standing myths and misapprehensions,
which are to be traced to inaccuracies in the accounts by the British protagonists
in the affair.
Keywords Count Johann Heinrich von Bernstorff, Theobald von Bethmann
Hollweg, Heinrich von Eckardt, Captain Guy Gaunt, Nigel de Grey, Otto
G¨ppert, Admiral Sir Reginald ‘‘Blinker’’ Hall, Sir Thomas Hohler, Col. Edward
House, Hans Arthur von Kemnitz, Dilly Knox, Robert Lansing, Helmut
Listemann, Count Karl Luxburg, the Swedish Roundabout, Edward Thurstan,
Wilhelm Wassmuss, Woodrow Wilson, Arthur Zimmermann, 13040, 18470,
26040, 3512, 5950, 7500, 89734, 9972
Introduction
‘‘If you knew what I know at this present moment, and what you will see reported in
tomorrow morning’s newspapers, you would not ask me to attempt further peaceful
dealings with the Germans.’’
President Woodrow Wilson, 28 February 1917, the day before the publi-
cation of the Zimmermann Telegram.
1
This article comprises four main sections: a popular outline of the Zimmermann
Telegram story; a list of the main sources (including descriptions of some important
ones which have only recently come to light); a detailed reconstruction of what really
happened; and a discussion and explanation of certain significant cruxes. Texts of
the Telegram in its various guises are given in the Appendix.
I. Outline of the Story
President Woodrow Wilson, whose views accorded with those of most of the
American population, had long kept the United States out of the World War
that was consuming the European powers with neither side in prospect of success.
In January 1917, the German Government decided that unrestricted submarine
Address correspondence to Peter Freeman, c
=
o Cryptologia, Department of Mathematical
Sciences, United States Military Academy, West Point, NY 10996, USA.
1
Quoted by Nickles [28, p. 151].
98
The Zimmerman Telegram Revisited
99
warfare (the sinking of merchant ships without regard to nationality and without
warning) was its best chance of winning: and to begin such warfare on 1 February.
The danger was that the sinking of American ships might overcome the
American official and popular mood in favour of neutrality and tip the United States
into joining the Allies. To strengthen his country’s strategic position if opposed by
the United States, the German Foreign Secretary Arthur Zimmermann planned to
persuade Mexico to declare war on the United States, thus distracting the latter from
the European war. On 16 January 1917, he sent a message to his ambassador in
Mexico City instructing him to propose to the president of Mexico that he should
form an alliance with Germany and attack the U.S. southern border. Germany
would provide financial assistance, and acquiesce in Mexico’s recovery of the states
of Texas, New Mexico, and Arizona (which had been lost to the U.S. almost seventy
years earlier). Zimmermann also told his ambassador to suggest to the Mexican
president that the latter should try to persuade Japan (a member of the Anglo-
French alliance) to agree to the plan, and should mediate between Germany and
Japan (which had taken over several German Far Eastern possessions).
Unknown to the German Foreign Office, its messages were being intercepted
and solved by the British. The work was done in the Admiralty, by a cryptanalytic
bureau best known by its address of ‘‘Room 40 Old Buildings (Room 40 OB),’’
under the control of Captain, later Admiral Sir, William Reginald ‘‘Blinker’’ Hall,
the director of naval intelligence. Hall realised that if Americans knew the German
plan they would be outraged and readily abandon their neutrality. From the British point
of view this would be far preferable to waiting until mounting shipping losses eventually
pushed a reluctant American public and Government into a declaration of war.
And so it fell out. Zimmermann’s message was secretly provided to Washington,
it being understood that the Administration would say nothing of its provenance. Its
explosive effect upon the President and (when it was published on 1 March 1917)
Congress and the American people was everything the British could have desired,
and the U.S. declared war on Germany on 6 April.
There is still much debate, not explored in this article, about the true strategic
effect of the Telegram,
2
but the general outline above is well-known and uncon-
tested. Less clear or agreed, however, are the technical details of how the Telegram
was encrypted and transmitted, how the British came to intercept and solve it, and
the sequence of events between its original transmission from Berlin on 16 January
and its text reaching first Washington officials (on 24 February) and then the
American public (1 March). The evidence has trickled out over the years and contains
a number of contradictions, largely the results of deliberate British misinformation.
II. Sources
This section describes the main primary and secondary sources on the Zimmermann
Telegram incident in broadly chronological order of creation. A number of primary
sources are difficult to come by, being unpublished, out of print, or available
only in national archive centres. Some, of particular significance, are discussed
and illustrated in more detail here.
2
Hereinafter the capitalised word ‘‘Telegram’’ will be used to refer specifically to the
particular message of 16 January on which the story centres.
100
P. Freeman
. A Room 40 copy
3
of one of the codebooks (13040) used for the Telegram
shows that the British recovered the meanings of the codegroups cryptanalytically.
They are written in pages on which the numbers 00 to 99 have been pre-printed,
with the first two or three digits of each group hand-written at the top as page
numbers (the pages with higher numbers, used for proper names, had a slightly
different format). There are extensive blanks where cryptanalysts could not sug-
gest a meaning for rarely-used code groups, and changes and corrections have
been made as understanding developed. Figure 1 shows the upper part of a sample
page, number 149. Other contemporary British material is described separately
below.
. The original plaintext of the Telegram is in German Foreign Office archives (see
Appendix item A). No copy of the encoded text of the Telegram as transmitted (or
intercepted) between Berlin and Washington has been found in German, British,
or American archives.
. There are no contemporary descriptions of the technical details of British
cryptanalysis, but a few items have survived which shed important light on British
handling of the Telegram. Figures 2 and 3 show what German messages
looked like, as intercepted by British censorship and presented to Room 40
cryptanalysts.
a. Figure 2 shows a message sent to Mexico City via the Swedish Foreign Ministry
on 8 February 1917 (a message which has its own role in the Telegram story
and is discussed later).
4
Throughout the war, a basic constraint on all German
communications with the Western hemisphere, including its diplomatic repre-
sentatives there, was that on 5 August 1914, the British cut all cables linking
Germany directly to the Americas. All remaining links to which Germany
could readily connect (e.g., through neutral Denmark) passed through the
U.K. British censorship prevented German messages being relayed. The
Swedish Foreign Ministry came to Germany’s aid by passing off the latter’s
messages as if they were Swedish. Messages sent this way went to the German
embassy in Stockholm, which gave them to the Swedish Foreign Ministry
to send to Swedish representatives abroad who gave them to their German
counterparts (and reverse).
5
b. Figure 3 shows a message sent to Washington on 27 September 1916, illustrat-
ing the use of the American State Department to carry German telegrams.
6
In a
3
The National Archives (TNA), Kew, Survey, HW 3
=
176. At least three copies are known
to have existed. In his account, Nigel de Grey [14] refers to two: his own personal one, and a
central Room 40 one. Another copy was provided to the U.S. after its entry into the war
(Friedman and Mendelsohn [10, p. 17]), which showed the same evidence of cryptanalytic
recovery as the copy now in HW 3
=
176.
4
The original of Figure 2 is in TNA HW 7
=
8. This is a file of Room 40 raw decrypts
of (largely) Berlin to Mexico City messages covering January to September 1917. For each
message, it usually comprises the original intercepted codetext (with interleaved decode) as
in Figure 2, and a separate manuscript translation; occasionally a circulating (typed) version
of the latter is included.
5
McKay and Beckman [25, pp. 45–46]; Kahn [19, p. 1025, fn. to p. 284], citing Swedish
information of 1966, which also said that no copies of such German correspondence were
retained in Sweden. See also Section IV below.
6
This is the start of a message originally dated (internally) 25, September 1916: decode in
TNA HW 7
=
19.
The Zimmerman Telegram Revisited
101
manner similar to the Swedes (though only between Berlin and Washington),
America sometimes passed off German telegrams as American to evade British
censorship. Such messages were given to the American ambassador in Berlin,
who sent them to the State Department where they were given to the German
embassy (and reverse). This practice is usually considered to have started on
2 June 1915 when, according to the memoirs of Count Johann Heinrich von
Bernstorff (the German ambassador in Washington), he asked Wilson to
facilitate his communications while negotiating during the Lusitania crisis;
7
but Friedman and Mendelsohn found a relayed German message dated
12 November 1914 among State Department papers.
8
The British did not
become aware of this practice, although American policy was an important
intelligence priority for them, and by at latest January 1916 they were regularly
decoding American telegrams.
9
. Contemporary manuscript and typed copies of Room 40’s decryption of the
Telegram have survived. The manuscripts were written by Nigel de Grey, the
member of Room 40 responsible for the decryption.
10
These copies, and in parti-
cular the annotations on them, are crucial, and some introductory description of
them is appropriate. One of these manuscript
=
typed pairs is of an early incomplete
decryption, the other is of the complete text as provided to the American Embassy
on 23 February 1917.
11
a. Figure 4 shows the first portion of a partial decode of the Telegram. The
heading of the message has been amended several times by de Grey (and
possibly others), reflecting the fact that though it was originally sent from
Berlin (hence the first heading, of ‘‘B[erlin] No 158’’), it was relayed through
Washington (hence the second thoughts which led to the insertion of
7
Quoted by Tuchman [32, p. 132]. The Lusitania was sunk on 7 May 1915.
8
Friedman and Mendelsohn [10, p. 13, fn. 33]. They also found (p. 14, fn. 36) some
Austrian telegrams (the British embargo applied to all enemy communications), but this
appears to have been an isolated case; the new Austrian ambassador, Count Tarnowski, on
his first call on the State Department on 3 February 1917, had no instructions from his
government in the aftermath of the German declaration of unrestricted submarine warfare:
Washington (Lansing) to Vienna (Penfield) 4 February 1917, TNA HW 7
=
20 (see next
footnote).
9
Churchill College, Cambridge, Archive Centre, Hankey Papers, HNKY 1
=
1, diary
entry for 1 February 1916 (in a section marked by Maurice Hankey [Secretary to the War
Cabinet] as unsuitable for publication): ‘‘I saw Captain Hall again first thing in morning.
He showed me more of [special Presidential adviser] Col. House’s telegrams sent from Berlin.
...I found that Hall had not shown these telegrams even to the First Lord [of the Admiralty
(the Cabinet Minister in charge), then Arthur Balfour]. This information is of course price-
less.’’ Room 40 worked on German messages, but other diplomatic communications were
decrypted by Room 40’s military counterpart, MI1b in the War Office. TNA HW 7
=
20 and
7
=
34 contain Room 40’s copies of MI1b’s distribution (none of MI1b’s own archives appear
to have survived) of decrypts of American telegrams between 25 January 1916 and 4 April
1917; messages about the Telegram have been extracted to HW 3
=
179—parallel citations of
them have been omitted in this article, which relies on the copies in the U.S. National
Archives.
10
The manuscripts are in TNA HW 7
=
8. The typescripts are in TNA HW 3
=
187.
11
Burton G. Hendrick was the first to publish these [17, pp. 336–337 and 333, respect-
ively]. The version published in the U.S.A. on 1 March 1917, differed from that provided
by the British (see text at fn. 84, and Appendix item C).
102
P. Freeman
Figure 1. The top half of page 149 of a Room 40 copy of the 13040 codebook. Note that on
each page groups for numbers (e.g. 14927 and 14979) and punctuation (e.g. 14901) are inter-
leaved; decades are shuffled, but within decades the order is alphabetic.
‘‘W[ashington]?’’). De Grey headed this partially-decoded text ‘‘7500’’
(underlined in the original). At a later stage he wrote ‘‘via U.S.A.’’ at the
top; this, like some of his changes to the heading, is in red ink. The large
‘‘B’’ at the left of the heading is also by him and in red; such markings,
apparently an attempt to categorise messages by subject-matter, appear on
several other decodes in Room 40 files.
b. Figure 5 shows the typed version of this partial decode. In this the number
is given (wrongly) as ‘‘W. 158’’, the date is ‘‘16th Jan. 1917,’’ and there is an
annotation in Hall’s distinctive, near illegible, handwriting: ‘‘main line—not
exposed’’ (i.e., not published). Words marked as doubtful in the preceding
manuscript were typed in red, and appear faint. A few editorial changes
are evident.
c. Figure 6 shows the first portion of a complete decode of the Telegram. ‘‘Feb. 19.
17’’ is written at the top right, where it was Room 40 practice to put the date of
decoding. The routing is given as: ‘‘From: Washington To: Mexico.’’ De Grey
added the parenthetical comment ‘‘[repetition of W No: 158 in 7500],’’ repeat-
ing the incorrect prefix ‘‘W’’ for that number. He wrote ‘‘No. 3’’, the message
number in the preamble added by Washington for transmission to Mexico City;
another hand has crossed this out (producing an effect like the Greek letter
beta) and written in ‘‘1’’ (the internal number assigned by Berlin, which also
appears in the next line of text).
12
12
This writeout is accompanied in TNA HW 7
=
8 by a small note: ‘‘Note: The 13040
[double-underlined] original of this is one of a number of Hohler intercepts filed together
(? At present extracted for D.I.D. 30.11.17)’’ [the question mark is in the original]. Berlin’s
number 7 to Mexico City of 31 January (a long message warning neutrals about the blockade
area to be enforced on 1 February) was also marked as ‘‘(at present extracted 30.11.17).’’
[ Pobierz całość w formacie PDF ]