found the double T.O. We then continue to get messages out with the EINS catalogue; each message gives us some values of bigrammes, which are entered on a Foss sheet. From time to time we go through the messages substituting for the bigrammes the values that have been recently found from the messages. With messages for which we know the values of two of the bigrammes we apply the method known as 'twiddling' or 'bonking'. We have to decipher the first few letters of the message at all of the 26 places consistent with our knowledge of the bigrammes. This is usually done in column, one column at a time, each column corresponding to a letter of the message. The twiddling is best done on the Letchworth enigmas, as they have no automatic T.O. Some more messages can be solved when one bigramme is known, preferably that corresponding to the L.H.W. by deciphering a few letters at every one of the 676 places on the test-plate. But this method is rather difficult to work in practice. It seems much more difficult to spot the right answer when one has to look through so many possibilities. The right answer is hardly ever noticed unless it is one of the obvious ones such as BIENE, WESPE, MUECKE, MOSKITO, HORNISSE, KRKR, ANAN, ADMX, GRUPPE, [unreadable]. The case where the R.H.W. bigramme is known cannot be done on the test-plate at all. One can of course use the X-machines in much the same way as was done with the original form of EINS catalogue. This has never been a success. One can also use 'hand methods'. One can go through the message looking for places where two consecutive letters occur on the same rod. The deciphered values also occur on the same rod, and we can examine the rods for possible bigrammes. Combining this with the Turing sheets, Kendrick has solved quite a number of messages. This method is known as 'clicks on the rods'.
[The following sentence apparently added later in Turing's hand]
We now have the EINS catalogue collated with the messages of the day so that
the only necessary work in EINSing is the testing of each possible EINS on
the machine and working out the start.