Communication Theory and Secrecy Systems [mimeographed typescript copy #3 - three of four plates in facsimile]
Murray Hill, N. J. Bell Telephone Laboratories, Inc. May 2, 1949. -40 leaves + 4 plates. 10 7/8 x 8 3/8 inches (plates slightly larger). Mimeographed typescript printed on recto only. Disbound lacking the original staple. Text with four holes punched at spine. Ink stamped "May 2-1949" on upper right corner. Graphic on page 35 is hand drawn in using pencil. Figures (plates) 1, 2, and 3 are provided in facsimile. Figure 4 is the original cyanotype format. Significant creasing throughout. Good. Wraps. 
During World War II, Bell Lab scientists worked on classified government-funded cryptography research. Shannon's research resulted in the classified internal technical memorandum "A Mathematical Theory of Cryptography" dated Sept. 1, 1945. (Sloane and Wyner #24) Never published, the Bell System Technical Journal in October 1949 first published a declassified abstract of this paper titled "Communication Theory of Secrecy Systems." (Sloan and Wyner #25)
The mimeographed typescript offered here, “Communication Theory and Secrecy Systems,” precedes the final Bell System Technical Journal publication by five months. Besides the title change, a quick side-by-side comparison notes some minor textual differences between this paper and the formal publication.
For the content and importance of the final paper, we will quote the experts. It would be worthwhile to examine the different papers and how they relate to the original classified memorandum.
"Shannon’s insight, his great contribution to cryptology, lay in pointing out that [language] redundancy furnishes the ground for cryptanalysis. 'In ... the majority of ciphers,' he wrote, 'it is only the existence of redundancy in the original messages that makes a solution possible.' This is the very basis of codebreaking." (Kahn 1996, p 748)
"The discoverer of information theory examines cryptology in terms of that theory: redundancy, entropy, equivocation. Shannon develops formulas that enable the cryptanalyst to tell when a solution in a particular cipher system is valid. He discusses perfect secrecy (the absolutely unbreakable cipher) and ideal secrecy (practically unbreakable ciphers) and work characteristics needed to solve different kinds of ciphers. (Kahn 1969, pp 324-5)
"[Communication theory of secrecy systems] has provided, for the first time, a well-organized theory of cryptography and cryptanalysis, and it is highly regarded by the expert cryptanalysts." (Pierce, p 271)
Sloane and Wyner do not record this typescript. OCLC/Worldcat shows one copy at the National Security Agency and a second at North Carolina State University (with differing pagination).
PROVENANCE: The personal files of Claude E. Shannon (unmarked). There were three examples in Shannon's files: this example (#3, three of four plates are facsimiles), example 1 (complete), and example 2 (two plates in facsimile).
REFERENCES: (all referring to “Communication Theory of Secrecy Systems”)
Sloane and Wyner, "Claude Elwood Shannon Collected Papers," #24 and #25
Shulman, David, "Theoretical and mathematical aspects of ciphers" in "An Annotated Bibliography of Cryptography" (Garland: 1976), p 122.
Kahn, David 1969, "Secret Writings: Selected Works on Modern Cryptography" in Bulletin of the New York Public Library, May 1969.
Kahn, David 1996, "The Code Breakers: A Comprehensive History of the Secret Communication from Ancient Times to the Internet" (Scribners, 1996, revised edition)
Pierce, John R., "An Introduction to Information Theory" (Dover: 1980)