It was sent by a secret cipher furnished by the War Department, which I used myself, for throughout the war I was my own cipher clerk. The ordinary method at the various headquarters was for the sender to write out the despatch in full, after which it was translated from plain English into the agreed cipher by a telegraph operator or clerk, retained for that exclusive purpose, who understood it, and by another retranslated back again at the other end of the line. So whatever military secret was transmitted was at the mercy always of at least two outside persons, besides running the gauntlet of other prying eyes. Despatches written in complex cipher codes were often difficult to unravel, unless transmitted by the operator with the greatest precision. A wrong word sometimes destroyed the sense of an entire despatch, and important movements were delayed thereby. This explains the oft-repeated "I do not understand your telegram" found in the official correspondence of the war period.
I have, since the war, become familiar with a great many ciphers, but I never found one which was more satisfactory than that I used in my messages to Mr. Stanton. In preparing my message I first wrote it out in lines of a given number of words, spaced regularly so as to form five, six, seven, eight, nine, and ten columns. My key contained various "routes" to be followed in writing out the messages for transmission. Thus a five-column message had one route, a six-column another, and so on. The route was indicated by a "commencement word." If I had put my message into five columns, I would write the word "army," or any one in a list of nine words, at the beginning. The receiver, on looking for that word in his key, would see that he was to write out what he had received in lines of five words, thus forming five columns, and then he was to read it down the fifth column, up the third, down the fourth, up the second, down the first. At the end of each column an "extra" or "check" word was added as a blind; a list of "blind" words was also printed in the key, with each route, which could be inserted if wished at the end of each line so as still further to deceive curious people who did not have the key. The key contained a large number of cipher words—thus, P. H. Sheridan was "soap" or "Somerset;" President was "Pembroke" or "Penfield;" instead of writing "there has been," I wrote "maroon;" instead of secession, "mint;" instead of Vicksburg, "Cupid." My own cipher was "spunky" or "squad." The months, days, hours, numerals, and alphabet all had ciphers.
The only message sent by this cipher to be translated by an outsider on the route, so far as I know, was that one of 4 p.m., September 20, 1863, in which I reported the Union defeat at Chickamauga. General R. S. Granger, who was then at Nashville, was at the telegraph office waiting for news when my despatch passed through. The operator guessed out the despatch, as he afterward confessed, and it was passed around Nashville. The agent of the Associated Press at Louisville sent out a private printed circular quoting me as an authority for reporting the battle as a total defeat, and in Cincinnati Horace Maynard repeated, the same day of the battle, the entire second sentence of the despatch, "Chickamauga is as fatal a name in our history as Bull Run."
This premature disclosure to the public of what was only the truth, well known at the front, caused a great deal of trouble. I immediately set on foot an investigation to discover who had penetrated our cipher code, and soon arrived at a satisfactory understanding of the matter, of which Mr. Stanton was duly informed. No blame could attach to me, as was manifest upon the inquiry; nevertheless, the sensation resulted in considerable annoyance all along the line from Chattanooga to Washington. I suggested to Mr. Stanton the advisability of concocting a new and more difficult cipher; but it was never changed, so far as I now remember.
