The History of Hut Eight


When a dayís traffic had to be decoded there was normally a residue of Offizier messages of which only the preambles were readable. If the messages were Barnacle, Cowrie or Conch the problem was not particularly difficult, wheel order and Stecker were known and the message breaking equipment used when the bigram tables changed could be effectively employed. Quite short cribs would break messages by hand methods and the American Hypo would get out what remained. Hypo was a most successful photographic apparatus which encyphered a message at all positions of the machine and examined the results not to see which were language but to see which contained most good and fewest bad letters. As the month progressed one accumulated a number of settings so that many messages turned up on known settings and could be decoded straight away. Also progress was more rapid towards the end of the month owing to a curious non-crashing property of the settings which was common to most keys most of the time. According to this property any single letter occurred once and once only in each of the 3 or 4 positions of the settings; thus if we discover a setting P N B we know that no other setting will have P on the left, N in the middle, or B on the right. As a result of this it was possible to cut down a great deal the work to be done by a machine such as Hypo while, on a 3 wheeled key, if there were only as few as 8 settings unknown, the total number of possible settings was only 512 and it was possible to break messages by trial and error. When discussing wheel order rules we shall meet similar equally stupid habits detectable in those responsible for making up the keys.

On the keys where the Offizier has a different Stecker the problem is much more difficult and an initial break has to be made on the bombe with the help of a crib. Cribbing was difficult chiefly because the addresses were all in the preamble and the Offizier part went straight into the text of the message; also there was little likelihood of routine messages in Offizier. The original break was actually obtained on a 'FORT' story; one message was spotted as a probable continuation of another and the crib FORT followed by the Time of Origin of part 1 was run anywhere early in part II where it did not crash. This was the classical method of Offizier breaking and was used frequently in later years though with rather less success in the long run than we expected; it was an extravagant and laborious process, as normally some 100 menus covering the most likely positions had to be run. Later details of recognition signals and Stichworts were the most fruitful sources of cribs, and also reencodements and various types of cypher error such as inexperienced officers not infrequently made. Opportunism played a very large part in Offizier cribbing, the supreme example being the breaking of the month's Oyster on a 16 letter message sent just after the final attack on the Tirpitz. The result of the attack was not then certain but the message was optimistically cribbed as 'TIRPITZ GEKENTERT' and the crib came out.

-52-


< previous

next >