The cryptographic organisation at Bletchley was highly efficient. Indeed it was the most efficient working organisation I have met, perhaps because there were no trade unions and little or no financial control and because it was run mainly neither by business men nor career civil servants but by mathematicians and chess players, who brought detached and decisive minds to the solution of cryptographic organisational and human problems. Contributory factors were the devotion, high morale and esprit de corps of the picked band of workers. Gifted people were willing to work on boring and repetitive tasks if it was important that these should be done by people capable of spotting the occasional small nugget, which might turn up in the sieve. For after all we were prospecting for something more precious than gold. An interesting symptom of Bletchley’s efficiency was the small size of the central administrative staff which was the servant of the operational chiefs and not their master, as is too often the case in peace-time civil service departments.
Nevertheless in spite of Bletchley’s efficiency, it was clear to me, as it had been to Welchman and Milner-Barry that our party had much to offer the cryptographers. Our communications picture, based on both logs and decodes and thus on an outline of the total working of each ‘star’ was different from, and complementary to, theirs which was based on the sorting of messages into keys and settings. We therefore formed a natural link between cryptographers and could offer them specific services. On Hut 6 blists, the big operational messages relayed on several stars showed up clearly and those of them that served as cribs were often unmistakable. But there were other routine messages, which were not so easy to spot on the blists, but were obvious in the folders of log readers’ summaries, where one could follow the star’s operational procedures repeating day after day. We were thus able to help Hut 6 to spot certain cribs and to call the attention of Hut 6 researchers to routine messages in an unbroken key, which might well prove to be useful cribs, once one day’s settings were broken.
Another service which we could give was to help in getting out the ‘duds’, the messages which did not make sense when decoded on the settings to which they had been classified. The chat about message checking on the logs could help in some of these cases either to get the message out or to suggest that it was in a different setting or to show that the German recipients themselves could not make head or tail of it. At the same time the Hut 6 section dealing with duds and cipher texts generally were in a position to contribute to radio intelligence by tracing retransmissions of messages on different networks.
Then again chat or messages in code or hand cipher on the logs might compromise part of an Enigma setting. This was rare but could be very important as in the case of the experimental signals regiment (see here) or in the extraordinary case of the SS ‘star’ to concentration camps. In addition to Enigma messages in their own key, which we called ‘Orange’, these camps each sent a daily return - number of interns at the beginning of the day, number received, number dead or transferred, number at end of day - a chilling little tally. These returns were sent not in clear but in a simple substitution code with a letter for each digit. This code was changed daily but with the first number of the following day always the same as the last number of the previous day it broke itself. To his amazement the Hut 6 cryptographer found that if the day’s substitution code was written out against the ten digits in order, the five pairs of letters were the stecker instructions for the current day’s Enigma; Orange had only five pairs of letters swapped instead of the usual ten pairs. This prime example of Germans using a pattern to hand meant that Orange could be broken without recourse to the bombe and on a short crib (see here).
However much the most important service, which we could give the cryptographers was to suggest re-encodements and this service became increasingly important.
Re-encodements of a message from a key already broken are clearly likely to provide an excellent crib and were used by Hut 6 from the first. There were for example re-encodements from one day’s setting into the next day’s setting. Obviously once everyone had changed the settings on their machines, it was easier for the Germans to repeat the message in the new setting to anyone who had somehow missed it. Indeed the fact that this was a re-encodement of a message first sent on the previous day was sometimes underlined by giving two times of origin in the preamble - one being the original time of origin and the second the time of the re-encodement.
Then there were re-encodements from one Enigma key into another and from Enigma to a hand cipher. Our log readers with their summaries of all communications on each star were in much the best position to spot such re-encodements. Hut 6 had in fact managed to pick out some re-encodements just by pairing messages with the same time or origin and approximately the same number of letters. But this was very hit and miss and tended to waste bombe time. We were in a position to suggest Enigma to Enigma re-encodements with much greater confidence, while the relatively rare instances of re-encodements between Enigma and a hand cipher could normally only be spotted from the logs.
Accordingly, as soon as I came to Bletchley, I began to visit Hut 6 research sections with suggestions about routine messages chat and other traffic on the logs, duds and re-encodements. And gradually I got into the habit of visiting Hut 6 watch and the blist room as well. As the need for them grew and was appreciated, regular services to Hut 6 were built up. These services became more and more important as Hut 6, by then headed by Stuart Milner-Barry, started to face new problems.
These problems were of two kinds - difficulties of sorting and identification and cryptographic complications. The number of different settings used daily increased, particularly after the Normandy landings; and it became more difficult to find routine messages which provided regular cribs in all of them. The way in had to come more often from a re-encodement. Fortunately as the number of keys increased, so did the number of re-encodements which we could spot on the logs. There were quite a lot of cases, where one station on a star had not got the same key as the others and day after day messages would be re-encoded to it.
With this sort of help, Hut 6 continued to break the bulk of air force and army Enigma messages pretty promptly in spite of new difficulties. Late in 1944 the German army and air force stopped using discriminants (see here) and it was no longer possible for Hut 6 to tell just by looking at the first group of a message what its setting was and into which pile it should be sorted. They had to sort on the call-sign book column of the call sign of the sender or recipient, identify the ‘stars’ from their call sign column and frequency and on this basis hunt for crib messages and re-encodements. Later sorting became still more difficult because the air force stopped using call signs from the predictable call-sign book columns (see here) I made a catalogue of known ‘stars’ their frequencies and characteristics, which was helpful to the log reader and perhaps to the intercept stations and Hut 6 research but it was little use to Hut 6 watch, who only had the Enigma messages and could not wait till we received the logs. They continued to sort by call sign column and to link certain columns together from their use on the same star or on stars linked by message retransmissions. In this way they kept up right until VE Day a volume of messages decoded within a few hours or at worst a day of their transmission, though the number of duds to be dealt with increased. In fact from the Battle of Britain to VE day, Hut 6 never failed for a single day to break Red, usually by 8 o’clock in the morning - and in spite of the proliferation of keys, Red, the Luftwaffe masehinenschuluessel, continued to be widely used and remained the most important source not only about the GAF but generally.
Hut 6 achieved this in spite not only of these problems of sorting but also of complications of the Enigma cipher itself. During the last year of the war, the German armed forces first put a new unkerwaltz (reflector wheel) into the Enigma machine and then later started to change the unkerwaltze connections daily as part of the daily cipher setting. The first problem was tackled by Dr Aitken, the Scottish chess champion, who worked on it for three days, while to deal with the second the Bletchley mathematicians, amongst others Welchman and Turing and Alexander (from the naval section) got together and designed an electronic computer.
The hand cipher cryptographers also had some new problems in the last years of the war. The Germans started to use ‘Stencil’, the kind of cipher in which you write your message along the rows of a pad that looks like a crossword puzzle form and then take out the columns in a pre-arranged order. This is a very difficult cipher to break even with a crib if the pattern of the crossword form is wholly unpredictable. Fortunately the Germans simplified the task with their passion for system and economy of means. They decided to construct their stencil forms out of just 28 different lines. All stencil pads were just arrangements of a selection of these lines. The difficulty presented by Stencil was thus greatly reduced.
I was from time to time able to help the hand cipher cryptographers by suggesting a re-encodement from a broken cipher (usually Enigma), by calling attention to traffic on the same star in a broken cipher or simply by identifying some of the units involved so that signatures and addresses could be guessed.