THE CODEBREAKERS

by DAVID KAHN

  374 Riverbank Publications: No. 15 published in 1917; Nos. 16 to 21 in 1918; No. 22 in 1922.

  376 Lohr: letter, January 15, 1948.

  376 Cartier false-dating: Friedman.

  376 treating frequency distribution as a statistical curve: In the Introductory to No. 22, Friedman farsightedly remarked that “when such a treatment is possible, it is one of the most useful and trustworthy methods in cryptography.”

  376 theory (of monographic coincidence): Friedman, II, Appendix 2; 114 for 1925 solution.

  377 Kr, Kp: Sacco has computed a constant called “presences” that differs from language to language and that figures in his formula for “mean quadratic frequency,” or “mfq,” to distinguish monofrom polyalphabetic texts. See his Appendix II and Table 28; also his “Dérivation de la formule des présences,” Revue Internationale de Criminologie et de Police Technique, No. 4 (October-December, 1957), 300-302.

  378 kappa, phi, and chi tests: Friedman, III, §§11, 12, 14.

  384 greatest single creation: Friedman, interview, April 16, 1963.

  384 bewildering variety: see, for instance, Figl.

  384 “cryptanalysis”: letter quoted in David Kahn, Plaintext in the New Unabridged (New York: Crypto Press, 1963), 33-34.

  384 Elements of Cryptanalysis: (Washington, D.C.: U.S. Government Printing Office, 1924). Kerckhoffs and, to a lesser extent, Delastelle had classified ciphers, but their efforts, in addition to being outdated in 1923, did not strike through to fundamentals as Friedman’s did.

  385 ex-prizefighter: “William F. Friedman’s Remarks at His Retirement Ceremony,” The NSA Newsletter, No. 25 (November, 1955), 7-8, 10, at 8.

  385 first two patents: 1,522,775 and 1,516,180.

  385 Teapot Dome: Senate, Committee on Public Lands and Surveys, Leases upon Oil Reserves, Hearings, 68:1 (GPO, 1924), 2483-7, 2515-21, 2548-51.

  385 Mars: “Radio Hears Things as Mars Nears Us,” The New York Times (August 23, 1924).

  386 Signal Corps: The Origin and Development of the Army Security Agency for all administrative details of Army cryptology; also Harris, 330-331.

  386 Navy: Smith, letter of October 30, 1963; Safford’s official Navy biography; PHA, 8:3556. Farago, 36-53, 76-77 has good material on this period. Smith’s “Solution of the Playfair Cipher” in André Langie, Cryptography, trans. J. C. H. Macbeth (London: Constable & Co., 1922), 170-188. Smith felt compelled to emend Macbeth’s sample text by changing fornication to intoxication, cutting off its i to make it fit.

  387 Zacharias: Secret Missions: The Story of an Intelligence Officer (New York: G. P. Putnam’s Sons, 1946), 83-91 at 89 and 101.

  388 solution of Hitt machine: DSDF 119.25/782 1⁄2.

  388 Friedman interests: see listing in Galland.

  389 Walter Reed Hospital: Friedman’s official Army biography.

  390 picture: cover of The NSA Newsletter, No. 25 (November, 1955).

  390 “avalanche”: David Kahn, “Decoding the Bard,” The New York Times Book Review (October 6, 1957), 3, 41.

  390 $100,000: 84th Congress, 2nd session, H.R. 2068 (Private Law 625, May 10, 1956), and accompanying House Report 260 and Senate Report 1815; transcript of hearing (not printed) held by the Committee on the Judiciary. Previous bills were H.R. 5278, 82nd Congress, and H.R. 1152, 83rd Congress.

  391 three Friedman patents: 2,395,863, 2,552,548, 2,877,565.

  392 Safford and Rowlett bills: 85th Congress, 2nd Session, S. 1524 (Private Law 494, July 22, 1958); 88th Congress, 2nd Session, H.R. 7348 (Private Law 358, October 13, 1964).

  Chapter 13 SECRECY FOR SALE

  I am grateful to Boris Hagelin for reading portions of the manuscript and offering suggestions, to R. D. Parker for his several letters and interview, and to the American Telephone and Telegraph Company for making patent files available.

  394 telegraph section, 17th floor: Ralzemond D. Parker, interview, December 4, 1962; Ralph E. Pierce, interview, May 9, 1963.

  394 Vernam: Encyclopedia of American Biography, XXXI (new series) (New York: American Historical Company, 1961), 217-218; Vernam’s daughter, Mrs. Per W. Nielssen, and his brother, Harold, interview, October 22, 1962; The Aftermath of the Class of 1914 of the Worcester Polytechnic Institute (Worcester, 1914), 132; Parker interview; R. B. Shanck, interview, August 9, 1963, for “What can I invent now?”

  394 summer: G. S. Vernam, “Secret Telegraph Systems—Western Electric Patent Applications,” Memorandum of August 1, 1918, in A. T. & T. File of Correspondence, Memoranda, and Notes Removed from Issued Application Folders for U.S. Patent 1,310,719, at 6.

  394 multiplexing: Report of the Chief Signal Officer, 1919, at 139.

  395 altering connections: Vernam memorandum of August 1.

  395 enciphering technique and arrangement: Vernam’s U.S. Patent 1,310,719; G. S. Vernam, “Cipher Printing Telegraph Systems for Secret Wire and Radio Telegraphic Communications,” Journal of the American Institute of Electrical Engineers, XLV (February, 1926), 109-115.

  397 sketch of December 17: Vernam memorandum of August 1, in which he states that “this sketch is the first record which we have of this suggestion.”

  397 meetings of February 18 and March 27 and tests: Vernam memorandum of August 1.

  397 Morehouse keytape modification: U.S. Patent 1,356,546.

  398 8,000 feet: G[ilbert] S V[ernam], “Secret Telegraphy—Double Key System,” Memorandum of June 10, 1918, in A. T. & T. File of Correspondence, Memoranda, and Notes Removed from Issued Application Folders for U.S. Patent 1,356,546.

  398 invention of the one-time system: Though Vernam invented the pulse adding device, no clear-cut contemporary record states who invented the cryptographic method that accompanied it—the one-time system. Following is the evidence on which I base the conclusions that my text embodies.

  Vernam seems to have evolved the randomness of the key. Parker and Vernam’s co-worker, Ralph E. Pierce, both credit him with it, and though no other evidence supports them, none contests them either. Because Vernam had had no prior interest in cryptology, I think it likely that he created the random key in an unthinking way—“let’s punch out some characters for the key”—rather than first making up a coherent key and then realizing its cryptographic weaknesses and deciding to nullify them with a random key. This requirement of randomness, which is not stated either in Vernam’s or Morehouse’s patents or in the A. T. & T. files on either, is first mentioned in Vernam’s Journal of the A.I.E.E. article, p. 113.

  The first mention of nonrepetition of polyalphabetic keys that I have found anywhere in cryptology occurs in a Parker Hitt “Memorandum for Chief of Staff” dated May 19, 1914: “No message is safe in the Larrabee cipher unless the key phrase is comparable in length with the message itself” (Hitt Papers). This was written about six months after Mauborgne had left the Army Signal School for the Philippines, but the statement, which stands without argumentative support, seems so positive and unquestioned as to imply a fairly long period of acceptance. Consequently, I think it probable that it was formulated with Mauborgne present, quite probably with his assistance, and that he adapted the principle to the Vernam machine keys.

  Before I discovered this mention in the Hitt Papers, I had asked both Mauborgne and Parker who invented the nonrepetition feature. Mauborgne replied (letter, March 5, 1963) that “yes, I did it.” (While he was referring to the work with Vernam, this may support the probability of his participation in the formulation of the principle at the Army Signal School.) Parker (letter, March 4, 1963) and Pierce (interview) both denied Mauborgne’s claim, asserting that Vernam invented the feature. However, while the A. T. & T. files not only contain no reference supporting Vernam, they do include documents corroborating several statements that Mauborgne made in his letter. Moreover, without going into details, I believe that the line of development implied by the Mauborgne claim is more likely than that urged by Parker and Pierce, and some evidence that the implied Mauborgne line of development actually happened comes from eng
ineer Donald B. Perry (letter, July 1, 1963). Admittedly, Perry did not join A. T. & T. until June, 1920, and so his testimony is at second hand, but he was a colleague of Vernam’s on the cipher machine from those early days. Pierce’s memory, incidentally, seems influenced by a Parker memoir on the subject; his is not an entirely independent recollection. But in any case I think that the Hitt statement fairly well settles the matter in favor of Mauborgne, though the proof is not absolute.

  All this seems to call for the conclusion that Mauborgne welded the two elements together, and I have consequently used it, even though no documentary evidence exists for it. Other Vernam co-workers threw no light on the question of invention (David E. Branson, letter, June 26, 1963; Roy B. Shanck, interview, August 9, 1963). The National Archives could not find any documents filed by Mauborgne at the time of his work with the Vernam machine, nor any report filed by Lieutenant Griffiths.

  398 unbreakable: William F. Friedman states in his article, “Cryptology,” in the Encyclopaedia Britannica that “a letter-for-letter cipher system which employs, once and only once, a keying sequence composed of characters or elements in a random and entirely unpredictable sequence may be considered holocryptic, that is, messages in such a system cannot be read by indirect processes involving cryptanalysis, but only by direct processes involving possession of the key or keys, obtained either legitimately, by virtue of being among the intended communicators, or by stealth.”

  400 reasons for only partial use: Friedman, III, 71-72; Hans Rohrbach, “Mathematische und maschinelle Methoden beim chiffrieren und dechiffrieren,” FIAT Review of German Science, 1939-1946: Applied Mathematics (Wiesbaden: Office of Military Government for Germany: Field Information Agencies, Technical, 1948), I, 233-257, at 242.

  401 Mauborgne and tri-city circuit: Report of the Chief Signal Officer, 1919, at 140-141.

  401 Gherardi-Fabyan correspondence: photocopies in Piercers possession; originals in A. T. & T. Confidential File 3710.

  401 demonstrations: Demonstrations for the Delegates to the Preliminary International Communications Conference ([New York:] American Telephone and Telegraph Company [1920]), 7-8; Vernam, “Cipher Printing Telegraph Systems,” 115; VernanVs Diary, entry for February 9, 1926.

  402 German activities: interviews with Adolf Paschke, May 3; Werner Kunze, May 4; Rudolf Schauflfler, May 6, all 1962.

  403 publicity on Vernam’s system: “A Secret-Code Message Machine,” The Literary Digest, LXXXIX (April 17, 1926), 22; “Automatic Code Messages,” Science, LXIII (new series) (February 19, 1926), unnumbered page in “Science News” section; Yardley, 363-365; Yardley, “Are We Giving Away Our State Secrets?” Liberty, VIII (December 19, 1931), 8-13.

  403 Vernam cryptographic inventions: U.S. Patents 1,416,765 for stunts, 1,584,749 for handwriting, 1,613,686 for pictures.

  403 to I.T.T.: Vernam’s Diary, July 1, 1929.

  403 SIGTOT: Harris, 586-588.

  403 Vernanm’s death: The New York Times (February 10, 1960).

  404 Gioppi: La crittografia diplomatica, militare e commercial (Milano: Ulrico Hoepli, 1897), at 45-46. Friedman’s Riverbank Publication No. 18 contains a section on “Digraphic and Trigraphic Substitution” at 5-9.

  404 Levine trigraphic substitution: Levine, letter, October 10, 1964. Another of his is in The Cryptogram, XXVIII (January-February, 1961), 54-56.

  404 “Cryptography in an Algebraic Alphabet”: American Mathematical Monthly, XXXVI (June-July, 1929), 306-312.

  404 Hill: obituaries in The New York Times and New York Herald Tribune (January 10, 1961); Mrs. Hill.

  404 Telegraph and Telephone Age: “The Role of Prime Numbers in the Checking of Telegraphic Communications,” No. 7 (April 1, 1927), 151-154, and No. 14 (July 16, 1927), 323-324; “A Novel Checking Method for Telegraphic Sequences,” No. 19 (October 1, 1926), 456-460.

  404 patent: Hill Papers, in my possession.

  405 “Linear Transformation Apparatus”: American Mathematical Monthly, XXXVIII (March, 1931), 135-154.

  405 previous proposals: Buck cited in bibliography by Maurits de Vries; Levine mentioned in M. E. Ohaver, “Solving Cipher Secrets,” Flynn’s Weekly (November 13, 1926), 794-800 at 799-800.

  405 methods of encipherment: A good elementary introduction to the Hill system is given in Lyman C. Peck, Secret Codes, Remainder Arithmetic, and Matrices (Washington, D.C.: National Council of Teachers of Mathematics, 1961), and a clear explanation in Wolfe, III, Lesson 14.

  407 comparison of matrix and linear systems: J. M. Wolfe with Salvatore Bonafide, On Algebraic or Polygraphic Ciphers, annotated translation of Sacco’s Appendix I-D, “Sulla Cifratura Algebraica e Poligrammica” (mimeographed, no date or publisher), notes 13 and 18.

  408 cryptanalysis: I am grateful to Dr. Jack Levine for his comments on the cryptanalytic defenses of the Hill system, and especially for his checking out of my suggestion concerning the possible weakness of two encipherments of a single message. For details, see his “Some Elementary Cryptanalysis of Algebraic Cryptography,” The American Mathematical Monthly, LXVIII (May, 1961), 411-418; “Some Applications of High-Speed Computers to the Case n = 2 of Algebraic Cryptography,” Mathematics of Computation, XV (July, 1961), 254-260; “Analysis of the Case n = 3 in Algebraic Cryptography with Involutory KeyMatrix and Known Alphabet,” Journal für die reine und angewandte Mathematik, CCXIII (1963), 1-30; and, with Joel V. Brawley, Jr., “Involutory Commutants with some Applications to Algebraic Cryptography” Journal für die reine und angewandte Mathematik, CCXXIV (1966), 20-43.

  408 Hill patent: 1,845,947 (with Louis Weisner).

  408 Hill’s later papers: Hill Papers.

  408 Bruton: New York Herald Tribune obituary.

  410 Albert: Who’s Who in America, 1962-63. His paper, “Some Mathematical Aspects of Cryptography,” seems never to have been published. It circulates in manuscript. Wolfe discusses pedagogical value of mathematics in cryptology at I, ii.

  411 rotors: Givierge, 281-285; Sacco, §83c; Eyraud, §§108-109; patents; study of a Hebern machine.

  413 rotor solution: Rohrbach, 253; Sacco, appendix I-B; Ottico Meccanica Italiana, Cryptograph C.R.: A Modern Patented Coding Machine (Roma: Ottico Meccanica Italiana, n.d.), at 13-17; Arne Beurling, interview, November 9, 1963; my own investigations. NA, Microcopy T-311, Roll 83, Frame 7108489, cites a rotor solution.

  415 Hebern: Mrs. Ellie Hebern, his widow, telephone interview, January 16, 1963, and letter, January 21, 1963. The name is pronounced HEE-burn.

  415 early patents: 1,084,010 and 1,096,168 for checks; 1,086,823 and 1,123,738 for keyboards; 1,136,875 for blocks; 1,141,055 for typewriter. The two-typewriter arrangement is mentioned in U.S. Patent Office, Interference 77,716, Edward H. Hebern vs. Austin R. Noll, November 13, 1939, at 14-15 of hearings testimony.

  415 1917: Interference 77,716, Hebern brief. This date is crucial because it awards the priority of the rotor invention to Hebern even though his application for a rotor patent was not filed until after two other inventors had filed theirs. The 1918 date of construction is corroborated by the statement printed in 1922 that the first machine was “completed about four years ago.” (H. H. Dunn, “Electrical Machine Can Make Eleven Million Codes,” Popular Mechanics Magazine, XXXVIII [December, 1922], 849-850; it has photographs of Hebern, a rotor, and the machine.)

  415 Navy in 1921: Admiral Milo F. Draemel, letter, November 23, 1963.

  415 “something radically better”: United States Court of Claims, Case 213-53, Ellie L. Hebern, executrix of the estate of Edward H. Hebern, deceased, and Hebern Code Inc., a corporation, vs. United States of America, May 19, 1953, “Memorandum of Conference on 8 September [1956] in the office of the Judge Advocate General re Hebern Code Inc. infringement, etc.,” at 5-6, quotation of Rear Admiral S. C. Hooper, U.S.N., Retired.

  417 Hebern corporate history: This is assembled from stockholders’ reports of March 1 and October 1, 1922; August 1 and November 1, 1923; August 28, 1924, and November 20, 1925; from news stories in t
he Oakland Tribune on January 28 and December 9, 1923; April 28 and 30, May 2, June 18 and 27, August 6, September 23 and 30, and December 12, 1924; March 15 and 18, April 9 and 30, May 2 and 14, June 18, July 7 and 8, August 27 and 28, and September 12, 1925; January 19, February 1, March 1, 2, 3, 4, and 8, May 14, June 16, 23, and 28, 1926; July 29, 1927; September 10, 1947; and from the proceedings of Interference 77,716 and the Court of Claims case. Hebern’s rotor patents include 1,510,441, 1,683,072, 1,861,857 (the biggest U.S. patent on cryptography), 2,269,341, and 2,373,890.

  419 Friedman solution: Friedman, II, 114.

  419 top naval cryptosystem: Safford in Court of Claims case testimony, at 8-9.

  419 IBM: This is Interference 77,716; Noll was an employee of IBM. Interferences 77,445 and 77,446 were combined with 77,716. Interferences 78,370 and 79,267, also brought on by Hebern against IBM employees, were dissolved.

  420 Koch: information from his son, H. E. Koch of The Hague, obtained by Maurits de Vries, March, 1963. Octrooiraad, letter, March 7, 1963, for patent assignment. The corresponding U.S. patent is 1,533,252.

  420 Scherbius: biographical information from German patents 318,911, 331,419, and 331,683.

  420 Scherbius patents: in the United States, 1,556,964 for codewords; 1,584,660 for numerical rotors; 1,657,411 for standard rotors.

  421= Enigma models: Dr. Siegfried Türkel, Chiffrieren mit Geräten und Maschinen (Graz: Verlag von Ulr. Mosers Buchhandlung, 1927), 71-94, plates M-P.

  421 Gewerkschaft Securitas: assignee of first Scherbius patent.

  421 Chiffriermaschinen Aktiengesellschaft: Handbuch der Deutschen Aktien-Gesellschaften (Berlin: Verlag für Börsen- und Finanzliteratur), 1925, II, 2888, for founding; 1930, III, 3988; 1935, V, 6610, for no dividends and dissolution. Hardie translations.

  421 Postal Union: Türkel, 77-78.

  421 Radio News: Dr. Alfred Gradenwitz, ‘‘Secrecy in Radio,” V (January, 1924), 878, 997-998.

  421 flyers and pamphlets: Die Schreibende Enigma-Chiffriermaschine, undated one-page broadside; Ciphering Typewriter Enigma, undated 16-page pamphlet, with “natural inquisitiveness” at 5.