THE CODEBREAKERS

by DAVID KAHN

  274 staff increased: James, 129.

  274 staff: James, Sky Was Always Blue, 104, 106; Toye, 188; James, letters, February 1 and 4, 1964; Who’s Who, 1962, for Adcock, Toye, Beazley, Savory, Waterhouse, Fraser, Willoughby; Who Was Who, 1929-1940, for Clarke, Monkbretton; 1941-1950, for Dilwyn Knox; 1951-1960, for Ronald Knox, McCarthy, Tiarks, Young; also for Young, James, 90, and George Young and Joseph M. Kenworthy, Freedom of the Seas (New York: Liveright, 1929), at 80-81. Great Britain, Admiralty, The Navy List, gives, in its quarterly issues, under Director of Naval Education (until transfer of the cryptanalysts to Hall), a list of many names of persons presumably assisting in the work; but James states in his letters that the lists—or at least that for July, 1916, which is typical—include several who did not make much of a contribution or were only occasional helpers, and omitted several who did much more and much better work than those so listed. I have therefore followed his recommendations. On Father Ronald Knox, though Evelyn Waugh states in Monsignor Ronald Knox (Boston: Little, Brown & Co., 1959), at 156, that Knox (who did not convert until September, 1917) worked in the War Office and does not mention his Room 40 service, James says that “It is understandable that Evelyn Waugh did not know that he came to us for the last few months of the war.”

  275 50 cryptanalysts: “War Work at the Admiralty.”

  275 social types, typists: Toye, 188; James, xviii, 32-33.

  275 Lady Hambro: W. Lionel Fraser, All to the Good (London: Heinemann, 1963), 62. His ch. 5, 52-62, gives some interesting sidelights on Room 40.

  275 Ewing retirement: Ewing, 195-208; “War Work at the Admiralty” for “pedestrian wits.”

  276 Hall: James, xxiv, 2, 6, 13; Tuchman, 8; Hendricks, 361; Toye, 188-189.

  277 French naval solutions: Marcel Givierge, “Questions de chiffre,” Revue Militaire Française, LXXXXIV (new series) (June 1, 1924), 398-417, (July 1, 1924), 59-78, translated as “Problems of code,” Articles, 4-31, at 19, 23, 27; [Georges J. Painvin], “Conférence de M. Georges Jean Painvin,” Bulletin de l’A.R.C. [Amicale des Réservistes du Chiffre], VIII (new series) (May, 1961), 5-47, at 8. There is a discrepancy with James’s statements: Givierge says that the three-letter code was for U-boats and was monalphabetically enciphered. Painvin’s article—essential for the study of World War I cryptology—is based upon his original working papers.

  277 Nauen solutions: Cartier, “Souvenirs,” II, 49-50.

  277 minimal reciprocation, cruiser, Berlin-Madrid: Cartier, “Souvenirs,” II, 34-36, 32, 33; Painvin, 10. Room 40 also read Berlin-Madrid messages through a captured code (James, 69; Ewing, 199-200; Hoy, 155-157).

  278 Mata Hari: Painvin, 10; Sam Waagenaar, Mata Hari, adaptation by Jacques Haubert (Paris: Fayard, 1965), at 198-203, 215, for texts of intercepts, 206-207, 223 for discussion.

  278 Austro-Hungarian codes: Cartier, “Souvenirs,” II, 33; Painvin, 9-10; James, xx, 159. The cryptanalyst was Painvin, whom the Italians awarded the Cross of the Chevalier of the Crown of Italy for his solutions.

  278 Neumünster: Churchill, III, 113; Corbett, III, 395; James, 120.

  279 Cypher SA: Great Britain, Admiralty, C.B. 0565A: Memorandum No. 7 regarding Ship Cypher ([London], 1918). “The Navy’s War Code,” The New York Times Magazine (September 11, 1932), 19:3, for Davidson.

  280 “cryptographers’ department”: Churchill, III, 112. He adds that Room 40 spared Britain the ordeal of German naval bombardment of her coastal towns because the fleet could not have stayed continuously at sea to prevent it.

  281 Persia: Young and Ken worthy, 80; James, 92; Ewing, 199.

  281 Trebitsch Lincoln: Hoy, 108-114; James, 36-37; Trebitsch Lincoln’s own memoirs.

  281 Casement: Strother, 145, for reproductions of Devoy’s messages; Hoy, 116-123 for OATS and code; James, 112-114 for diary; Ewing, 192-194.

  281 bird, code expert: Hoy, 92, 89-90.

  282 half-past ten: James, 136.

  282 1,000 groups: Robert Lansing, War Memoirs (Indianapolis: Bobbs-Merrill, 1935), 226.

  282 code 0075 used for message: GFO. Friedman and Mendelsohn guessed this (15, 17) and remarked that 0075 was “apparently reserved for messages of the highest importance” (19).

  282 cryptanalysts working for six months: Hayden Church, “A Sherlock Holmes of Secret War Codes,” The New York Times Magazine (November 8, 1931), 17. This is an interview with Ewing, in which Church states that “The deciphering of the code used by Bernstorff, the German Ambassador to the United States, took nearly six months and involved the piecing together of thousands of scraps of paper resembling an immense jigsaw puzzle.” He adds that “In this code the name given to the German Foreign Office was ‘Arthur Foxwell.’” This must be a Room 40 play on Zimmermann’s first name and his attempts at deception. German diplomats were not using jargon codes, and if they were Room 40 would not have taken six months to solve one.

  282 other codes in 0075 series: Friedman and Mendelsohn, 15. Another German diplomatic code is given in skeleton form in NA, RG 76, Mixed Claims Commission, Exhibit 86.

  282 distribution places and date of 0075: GFO. This verifies the “six months” statement of Church.

  282 0075 to Bernstorff by Deutschland: GFO; guessed by Friedman and Mendelsohn, 18, on the basis of Bernstorff’s statement that he received new codes and ciphers on each trip of the U-boat (in Official German Documents Relating to the World War [New York: Carnegie Endowment for International Peace, 1923], I, 313-315).

  283 partial text: The two sources, James, 136, and Hendrick, 336-337, appear to be different copies from the original partial solution. I have collated them with each other and with the final German text. Unsolved portions are indicated by ellipses between parentheses, questionable readings by parentheses, extremely questionable readings by question marks within the parentheses. The single question mark is given, with ellipses, by both James and Hendrick, but what it indicates is not clear, since in the German text only the words “to keep Stop” appear, and these are given in the English partial solution. The bracketed “of Mexico” is my insertion.

  283 copies burned, no word to Foreign Office: James, 136, 138.

  283 political situation: Tuchman, 4, 114, 107-108, 142-144.

  284 Eckardt: German Foreign Office, letter, January 10, 1964.

  284 routes of message: Hendrick, 335-342; Tuchman, 101-104, 128-136; James, 132-133; Friedman and Mendelsohn, 6-14. Britain’s 1915 positive knowledge of German superencipherment: Friedman and Mendelsohn, 9-10. Hendrick, Tuchman, and James list a third route, by radio from Nauen to Sayville, but Friedman and Mendelsohn, 7-8, give strong evidence against the likelihood of this channel’s being used for the Zimmermann telegram, and I have therefore omitted it. Finally, GFO gives only the American route for the message, which may mean that the Zimmermann telegram did not go by the Swedish roundabout. Opposing this is the fact that important messages between Washington and Berlin were frequently transmitted by several routes to ensure their reception. The archivist, Royal Ministry of Foreign Affairs, Sweden, says, in letter of June 1, 1966, that German telegrams “were in fact brought to the Ministry from the German Embassy in Stockholm for further delivery to the Royal Telegraphic Service, where they later were burnt.” Nor were copies kept either by the Ministry or the Swedish mission in Argentina. No copies of the codetext in 0075 thus seem extant, except those in the Room 40 archives.

  285 “highly entertained”: Ewing, 205.

  285 de Grey: Who Was Who, 1951-1960; Burke’s Peerage, 1963, at 2499-2500; John de Grey, letters, June 3, 1963, and undated, several weeks later.

  285 Montgomery: R. D. Whitehorn, letter, February 4, 1958, to Mrs. Tuchman (who kindly lent it to me) and enclosed record from St. John’s College, Cambridge, and undated newspaper clippings; Church, 17, for postcard—I assume the clergyman he mentions is Montgomery, the only one listed in The Navy List.

  286 methods for solution of code: Yves Gyldén, interviews, May 28-31, 1962; Valério; Sacco, §§104-108; Givierge, ch. 15.

&n
bsp; 287 additional traffic: James, 139, 140. These implications that the code was only partially solved at this time are reinforced by a statement of Page’s in a telegram of September 10, 1917, that the Zimmermann telegram “went in a code which the British had at that time only partly succeeded in deciphering” (DSDF, 862.20235/537). This incompleteness of the solution explains why the section concerning dismemberment of the U.S. was not present in the first partial solution that Montgomery showed Hall.

  287 reasons against disclosing telegram: my assumptions, based in part upon those in Friedman and Mendelsohn, 26-27.

  289 February 5: James, 140.

  289 agent T: James, 140-141, 133-135.

  289 Code 13040: GFO for distribution; David Kahn [pseud. Ishcabibel], “A Partial Reconstruction of a German Diplomatic Code,” The Cryptogram, XXVIII (September-October, 1960), 1, 4-7, for construction of code; also Friedman and Mendelsohn, 15-16. Hall’s affidavit states that “The German cipher book covering this system of ciphering [Code 13040] is in our possession, it having been captured by the British authorities in the luggage of a German consul named Wasmuss who was stationed at Shiraz while Wasmuss was engaged in an endeavor to cut a British oil pipe line.” This story is elaborated by James, 69, and Tuchman, 19-21. But it must be some lapse of memory on Hall’s part. As Friedman and Mendelsohn remark, at 17, “It seems unlikely that a German consul engaged in an expedition to cut a pipe line should carry a diplomatic code book in his baggage.” They suggest that the British found parallel plain and code messages in Wassmuss’ baggage. According to GFO, 13040 was never distributed to Wassmuss nor to any German missions in Persia. However, other codes were, and the British discovered “two ‘dictionary’ cyphers” wrapped in several pairs of long woolen underwear in the German consulate at Bushire, Persia, on March 9, 1915 (C. J. Edmonds, “The Persian Gulf Prelude to the Zimmermann Telegram,” Journal of the Royal Central Asian Society, XLVII, January, 1960, 58-67 at 65), and perhaps Hall confused that code with 13040. It is not known what code Wassmuss did have. Hall did not make his statement to cover up British cryptanalysis because earlier in his affidavit he stated that sometimes “our cipher experts were able to decipher the German ciphers.” Whatever the reason for Hall’s statement about a capture, it seems certain that Room 40 solved and did not capture 13040. Friedman and Mendelsohn, 17, point out that the copy of 13040 given by Britain to the United States after America entered the war gives every evidence of being a reconstruction, comprising about half the vocabulary with words and phrases from all sections and with some identifications marked as doubtful, and GFO reports that German archives show no reports of loss or compromise of that code. My text reflects my belief that 13040 was cryptanalyzed. Consequently, I have excluded the romantic stories of Wilhelm Wassmuss and of Alexander Czek, who is supposed to have stolen a German diplomatic code, which in any case may not have been 13040 or 0075 but any of the others; Hall in any event denied any knowledge of Czek (Robert Boucard, Les Dessous des Archives Secrètes [Paris: Les Éditions de France, 1929], 69-83).

  291 variations between the two texts: comparison of them. Printed versions at Official German Documents, II, 1337, for Berlin to Washington; Hendrick, 345-346, and Tuchman, 201-202, for Washington to Mexico.

  291 Britain deliberately holds the message for release: No evidence exists for this crucial statement. That many historians agree in it (views summarized in Samuel R. Spencer, Decision for War, 1917 [Rindge, New Hampshire: Richard R. Smith, 1953], at 62-64) means little, since they were unaware of the underlying cryptologic and espionage problems. Friedman and Mendelsohn, 26-28, believe that the desire to protect the espionage secrets motivated the delay, and both they and Mrs. Tuchman, who has made the closest study of the political circumstances, think that the timing of giving the note to the Americans was based on political factors. I concur in these conclusions. No evidence exists for any different reason for the release, either.

  291 “The danger is”: Hendrick, 324-325.

  291 “much that of a soda-water bottle”: the military attaché, Spring-Rice, quoted in Walter Millis, Road to War: America 1914-1917 (New York, 1935), 403.

  291 Bell: James, 142.

  292 text of message: James, 141; Hendrick, 333, and 345-346 and Tuchman, 201-202 for German.

  292 conference and Balfour: Tuchman, 163-164, 166.

  292 Page telegrams: Hendrick, 332-334. Original is DSDF, 862.20212/69.

  293 Polk, “Much indignation”: Tuchman, 168. 293 Lansing and Wilson: Lansing, 226-228.

  293 Hood, news stories: Tuchman, 175. For some reason the text was given out in a different and weaker translation than that furnished by Page to the State Department.

  293 “Please endeavor”: DSDF, 862.20212/69, Lansing to Page, February 28, 1917.

  293 “never used straight”: Hendrick, 344.

  294 Carlton: Tuchman, 171.

  294 “Some members”: DSDF, 862.20212/82A. The original Bernstorff-to-Eckardt Western Union telegram follows this and bears the same file number.

  294 “Bell took”: Hendricks, 345. Bell’s original decode is DSDF, 862.20212/81½. But only the first page and a few lines of the second are in his handwriting; most of page 2 and pages 3-8 are in handwriting identified by John de Grey in an undated letter of 1963 as that of his father, Nigel. James erroneously states, at 143, that the message was decoded by Bell in the American embassy so that Wilson could say it was decoded by Americans on American soil.

  294 statement to Senate: Tuchman, 180.

  294 pet theories: Hendrick, 356.

  294 Hall instigates: “Admiral Hall on the Zimmermann Telegram,” The World’s Work, LI (April, 1926), 578-579.

  294 Berlin-Eckardt messages: Hendrick, 357-360. Though at first sublimely confident that their code had not been broken (Tuchman, 189, 194), the Germans later resorted to a dictionary code to try to bribe Mexico to remain neutral (Yardley, ch. 6).

  296 hilarity: Hendrick, 356; “Admiral Hall on the Zimmermann Telegram.”

  297 Zimmermann admits: Tuchman, 183.

  297 American reaction: Tuchman, 184-187.

  Chapter 10 A WAR OF INTERCEPTS: I

  298 radio: see also Major R. B. Moran, “Powers and Limitations of Radio Communication Within a Modern Field Army,” Articles, 89-113 and 114-134, at 95-96 (July-August and September-October, 1936); William F. Friedman, American Army Field Codes in the American Expeditionary Forces During the First World War, War Department (Washington, D.C.: U.S. Government Printing Office, 1942), 3; “Signal Communications,” Encyclopaedia Britannica.

  299 first days: Givierge, “Problems,” 5.

  299 French intercept service: Cartier, “Service d’écoute.” This includes numerous photographs of intercept posts and one of Cartier.

  299 direct wire: Cartier, “Souvenirs,” I, 33.

  300 traffic analysis: Cartier, “Service d’écoute,” summarized at Gyldén, 31.

  300 language separation: Givierge, “Problems,” 25.

  300 Circourt: Gyldén, 35-36.

  301 ÜBCHI: Painvin, 11-12.

  302 solution of single-columnar transposition: Gaines, ch. 4; Sacco, §§78-79; Wolfe, ch. 9; Friedman, IV, §§i-iv; Wayne G. Barker, Cryptanalysis of the Single Columnar Transposition Cipher (Rutland, Vt.: Charles E. Tuttle Co., 1961). Frequencies from Wolfe. Adding up the logarithms of these frequencies instead of the frequencies themselves will give a more accurate result.

  302 solution of double-columnar transposition: Sacco, §84; Friedman, IV, §28g.

  303 key in force: Painvin, 12.

  303 cryptanalysts’ difficulties: Carter, “Souvenirs,” I, 25, II, 17; Givierge, “Problems,” 7.

  303 key reconstruction: Givierge, 200-204; Friedman, IV, §27.

  304 October 1: Givierge, “Problems,” 7. All first names of officers come from 1914 and 1919 editions of France, Ministère de la Guerre, Annuaire officiel de l’armée française (Paris: Berger-Levrault).

  304 gossip and subsequent solutions: Givierge, “Problems,”
8-9.

  304 Thielt: Painvin, 12.

  304 new system: Givierge, “Problems,” 8, for November 18; 12, for illusory complication; 9, for December 10; Painvin, 12, for operation; Givierge, 218-223, for methods of solution. Painvin gives December as the date for the introduction of the system, but he was not officially in cryptanalysis then and Givierge was.

  304 Painvin note, first solutions, goes to Bureau: Painvin, 11, 13.

  304 Painvin: Who’s Who in France, 1961-62; J. Rives Childs, letter, August 27, 1962; Yardley, 224; General Desfemmes, “Réflexions sur la guerre éléctronique,” L’Armée, No. 24 (December, 1962), 21-33 at 28.

  305 echelons: Givierge, “Problems,” 16; Marcel Guitard, “Conférence de Marcel Guitard,” Bulletin de l’A.R.C, VIII (new series) (May, 1961), 47-52 at 48; Gyldén, 30, 40.

  306 1915: Painvin, 14.

  306 other problems: Cartier, “Souvenirs,” I, 26; Gyldén, 42.

  306 retrospective solutions: Givierge, vi, and “Problems,” 9-10; Gyldén, 38.

  306 French solutions: Givierge, “Problems,” 14, 7, 18-24, 25, for “terrible regularity”; Gyldén, 38 for faked attacks, 36 for guessed keys; Guitard, 49, for proverbs.