New York, N.Y. Bell Telephone Laboratories, Incorporated March 1951. First Separate Edition. 15, [1-blank] pages. 10 7/8 x 8 3/8 inches (275 x 213 mm). Publisher's printed grey, blue and black wrappers. Stapled, with five holes punched at the spine as issued. One of two variants known, this example printed on coated paper. Good. Wraps. 
Presented at the Northern New Jersey Sub-Section of the Institute of Radio Engineers, Murray Hill, New Jersey, March 15, 1950. The Bell System Technical Journal, Vol. 30, pp. 50-64, January 1951 first published this paper, here offered as Bell Telephone Systems Monograph (#1819, March 1951), presumed to be the first separate edition.
"A new method of estimating the entropy and redundancy of a language is described. This method exploits the knowledge of the language statistics possessed by those who speak the language and depends on experimental results in the prediction of the next letter when the preceding text is known. Results of experiments in prediction are given, and some properties of an ideal predictor are developed." (abstract)
An oft-cited paper with applications in data compression and cryptography.
Sloane and Wyner, "Claude Elwood Shannon Collected Papers," #69
Reprinted in D. Slepian, editor, "Key Papers in the Development of Information Theory," IEEE Press, NY, 1974, pp 42-46.
COLLECTORS NOTES: 1) There are two known variants of the March 1951 printing. One, as here, is printed on coated paper with the "1819:C.E. SHANNON" on the rear cover adjacent to the middle hole aligned with the "E." in Shannon's name. The other variant is printed on non-coated paper, and the hole on the rear cover is adjacent to the "SH" in Shannons last name. The coated paper copy is somewhat thinner. There is no known precedence.
2) The Bell Telephone System Monograph series offered a way to obtain individual articles by Bell scientists regardless of where their work was first published. Many Monographs significantly postdate the original article publication. Because of this, they rarely constitute the coveted (and traditional) article offprint. If the journal of record issued no offprint, the Monograph might be the first separate publication - the closest the collector can come to a traditional offprint. We have done our best to place each Monograph properly in the article’s publishing history and welcome any corrections or additional information, especially regarding issues unknown to us.