CRYPTOLOGIC QUARTERLY
- 4. From table 2 pick out pin settings for interrupter wheel and set pins.
- 5. Decipher the first five letters of text. Reset the machine to its original setting. Decipher the next 495 letters. Pause. Pull handle 10 times. Continue. (This break at 500 was in effect for only a few months.)8
Because they were able to read the Red messages, our analysts learned about Purple even before the Japanese had started to use it.
The Breaking of Purple
In December 1938 a message appeared in Red cipher which authorized a man named Okamoto to put certain cryptographic devices into service. The Japanese diplomatic officer referred to the machines as the Type B cipher machines. They were to replace the currently used Type A (Red) machine for highly secret communications between important Japanese embassies throughout the world and the Foreign Office in Tokyo. The B machine would go into effect on 20 February 1939.9
The first message received after 20 February came from Warsaw, Poland. The message contained an indicator different from the normal A type. Six of the 26 letters had abnormally high frequencies, which was characteristic of traffic enciphered by the Red machine. The A machine continued to be used on a regular basis in Hsinking and Shanghai, and occasionally (apparently when the B machine was out of commission) at places which had been provided with a B machine. The B machine was assumed to be a modification of the basic A machine.
By By April 1939 it was possible to decipher the sixes of the Purple messages in most cases. In the set of six letters each letter had about the same frequency; likewise in the set of twenty letters each letter had about the same frequency, but different from that of the smaller set. If the set of six contained two high frequency letters, the frequency of the six would be significantly higher than the twenty. If the six contained two low frequency letters, the frequency of the six would be significantly lower than the twenty. Either of these cases ensured that the sixes could be isolated. It was discovered that, out of 120 different indicators, there were only 25 unique starting points. A distribution table was built from which an analyst could recover the starting point of the sixes, and then fill in all plaintext values for the sixes. Appendix A is the explanation of this distribution table which was given in the course on Purple (available in the Cryptologic Collection).
- 8 Ibid.
- 9 William Friedman, "Preliminary Historical Report on the Solution of the "B" Machine." 14 October 1940, p. 1.
