Page:Red and Purple - A Story Retold.pdf/2

 CRYPTOLOGIC QUARTERLY

In the first M-3 messages, intercepted in March 1933, there were two classes of letters, vowels and consonants, with different frequency distributions. As there was not much traffic, the messages were put into one of those "to be looked at when there is nothing else to do" bins. However, it was postulated that vowels were substituted for vowels, and consonants for consonants. Two reasons for this supposition were that (a) telegraphic regulations allowed cheaper rates for pronounceable text (pronounceable meaning a minimum of two vowels per five-letter grouping) and (b) the transliteration of Japanese characters is unusual in that vowel-for-vowel and consonant-for-consonant substitution produces a pronounceable cipher. l

In March 1934 the characteristics of the cipher changed. Though the text still divided into two classes, one with six letters and the other with 20 letters, there was a mixing of vowels and consonants in each class. At that same time there was also a change in telegraphic regulations: there was no longer a special rate for pronounceable text. 2

The initial study of the intercept revealed the following:


 * 1. The messages had a five-digit indicator in the first group of the message.


 * 2. The text was definitely cipher.


 * 3. The nature of the substitution was such that repetitions were permitted to occur.


 * 4. There was a radical change on the lst, 11th, and 21st of each month.


 * 5. The keying element produced a long cycle which appeared

indeterminate in length. 3

From Manchuria in 1936 came an influx of cipher messages (10-15 per day) which exhibited the above properties. In particular, extensive traffic was received from 11-20 December, a single cipher period.

When the attack against the December messages came to a standstill, the analysts turned to the 1933 messages. One of these analysts, a Japanese linguist, thought of exploiting the fact that the doublet "OO" occurs frequently in Japanese, as does the aba construct "O-O." If the machine used rotors, then the sequence of letters on a wheel could be discovered when the wheel enciphered the same letter consecutively. The analysts made up the following two tables from the vowels in the longest message in the 1933 collection. The first table shows the


 * 1 RED and PURPLE, undated (pre-1960), S-119-785, U.S. Army A.S.A., p. 1.
 * 2 Ibid.
 * 3 Ibid., p. 2.