Page:Red and Purple - A Story Retold.pdf/8

 CRYPTOLOGIC QUARTERLY

The translation of this much of the message is "Number 15 (part 1 of 2) SECRET, to be kept within the department. Paragraph: On March 16, the American Ambassador Grew .... "11

Once a few words were known, it was sometimes possible to guess the subject of the message, and on occasion, to find government documentation on the subject. Some messages were exact quotes of particular documents. Another help in deciphering messages occurred when an operator transmitted a message in a known cipher, and then retransmitted the same message in Purple. There were several occurrences of isologs of this type. Plain text for parts of 15 fairly lengthy messages was obtained. These were subjected to a most "intensive and exhaustive" cryptanalytic study. In this study, the following phenomena were observed:


 * 1. The ciphering mechanism started from a certain initial setting and progressed methodically without cyclic repetition of any sort to the end of the message. The longest message was over 1500 letters long.


 * 2. Two identical plaintext letters in sequence could never be represented by two identical ciphertext letters. This phenomenon is termed "suppression of duplicate encipherments at the first and 26th intervals."


 * 3. Two messages with identical indicators on the same day appeared to be identically enciphered and on direct superimposition - and when written on a cycle of 26 - were monoalphabetic within columns, but with the alphabet constantly, irregularly, and unpredictably shifting from column to column.


 * 4. Two messages with identical indicators on different days were absolutely different.


 * 5. Two messages with different indicators on the same day (i.e., same plugboard arrangement) were absolutely different - no cryptographic similarities whatever.


 * 6. In each line of 26 letters, two identical letters could be identically enciphered except under certain conditions. Two adjacent letters were never identically enciphered, and two letters at intervals two, three, four, or five from each other were rarely enciphered identically.12

At this point in the study, the World War II analysts felt that with 20-25 messages with the same indicator on the same day, it would be possible to find a


 * 11 Ibid., p. 3.
 * 12 Ibid., pp. 4-5.