'pinched' from a German boat, enabling us get the keys for April 22-27 1940. At the same time we pinched a book of instructions telling us the precise form of the indicating system.
To encipher a message the operator chooses two trigrammes out of a book. The first of these trigrammes is called the 'Schluesselkenngruppe'. The choice of this is partly determined by the nature of the message: e.g. all 'dummy' messages have the Schluessel kenngruppe taken from one part of the book and genuine messages have them taken from elsewhere*: we do not know very much about these. The second trigramme is called the Verfahrenkenngruppe. Suppose the Schluesselkenngruppe is CIV and the Verfahren kenngruppe is TOD then the operator chooses two dummy letters, Q and X say, and writes this down
Q C I V
T O D X
From the Verfahren kenngruppe is obtained the window position for the start of the message, by enciphering at the Grundstellung. From the eight letters above one also obtains the indicator for the message, by substitution from a table which gives bigramme for bigramme. The substitution is done by replacing the vertical pairs above with bigrammes, e.g. in this case, if the substitute for QT were DA, and TH for CO, PO for ID, and CN for VX then the indicator for the message is DATH POCN. Apart from the Schluessel kenngruppe feature this is the method we had inferred was being used. This extra feature accounts for the bigrammes in the indicators being almost perfectly hatted. Also the fact that it is never the message setting itself which is chosen at random by the operator eliminates any remaining hope that one might use 'operator's psychology' to help in finding out the alphabets. From our point of view of course the Schluessel kenngruppen might as well not exist, and the (bigramme lists to us remain Foss sheets with one letter entered in each square, and not two. There is however the restriction that there must be exactly 26 occurrences of each letter.
*This has gone out of date