The Breaking of German Naval Enigma
by Tony Sale | Back to Naval Enigma Index |

Tony Sale's
Codes and Ciphers

Breaking Shark on the Short Weather Signals.

When the four wheel Enigma came into use in February 1942, Bletchley Park could no longer read the vital U-Boat signals, mainly because there were no four wheel Bombes and not enough three wheel Bombes to tackle the extra 104 combinations from the fourth wheel and the fourth wheel and reflector combinations.

Then the capture of the code books from U559 in October 1942 opened up a new possibility for cribs. Among the code books was the "Wetterkurzschluessel" , the Weather Short Code Book of 1941.

At this time only the U-Boats and Naval High Command had four wheel Enigmas, so to "talk to" a shore station or other ships without a four wheel machine, the fourth wheel had a neutral position ("A") at which it and its matching reflector behaved just like a three wheel Enigma reflector and thus messages enciphered at this neutral position could be deciphered by a three wheel machine.

Someone in Bletchley Park realised that Short Weather Signals were sent in the three wheel mode and this elaborate system was devised to exploit this.

The German Meteorology Service needed weather reports from the North Sea and the North Atlantic in order to prepare their weather synoptics. They could get the required information from U-Boats operating in these areas. To achieve this the Central Met. Station sent a request message, eniciphered on a three wheel Enigma, to a U-Boat requesting a weather report at a specific time. At the appointed time the U-Boat would surface, observe the weather conditions and then using its four wheel Enigma set in neutral, three wheel mode, report back to Central Met. But in order to minimise the possibility of the signal from the U-Boat being direction located by the Allies, the weather report was condensed by translating the met observations into a series of short codes by using the Weather Short Code Book. It was this series of codes which the U-Boat Enigma operator enciphered on his machine.

Now comes Bletchley Park's really clever bit. They realised that when the Central Met. Station had collated the U-Boat and other reports, they broadcast a general synoptic in their own Met. Code.

This Met. broadcast used the standard International Met. Code but encoded it further using bigram tables. Mr Archer in the Met. section in Bletchley Park had worked this out, (decoding it was known as Archery) and so from this synoptic the Enigma code breakers knew the weather conditions where the U-Boat was and so could deduce what had been observed by the U-Boat. Finally from the code book captured from U559, they could work out the weather codes keyed in by the Enigma operator to give the U-Boat signal which had been intercepted by the Allies. They had a "Crib" and this was for a three wheel Enigma configuration which the Bombes could break.

Weather Codes in action

The complete Weather Code system is described in A P Mahon's History of Hut 8 and here is a shortened version of that description.

In 1942 the Germans were using Wetterkurzschluessel (2. Auflage) and the following is based on that.

The weather situation was represented by 11 letters, each one representing a different weather condition, but always sent in the same order.
Here are the letters and the corresponding tables in the W Book as shown in Mahon.

and here is the W Book table for Direction and Type of Swell.

The Weather Code letters were preceeded by a single letter which referenced via a table, the three letter message start setting for the Enigma machine.

    This example, taken from the W Book, forms into four groups of four letters: MZYR QSOV KVQM and if "m" gives a wheel start of GWB, for a Grundstellung of: wheel order I VI III Ringstellung DKT Stecker TG DW SX IQ LF UC PK YR BN MJ you can encipher this to get: KQJ MUAU JSZL, with the M unenciphered at the start.

    This was what the Enigma operator gave to the U-Boat's Radio operator to transmit by Morse code back to the Met. Station.

So now assume that this has been intercepted by the "Y" station at Scarborough and sent to Bletchley Park. Also assume that the "Y" station managed to get a Direction Fix (DF) on the transmission so that its position was known.

Further assume that Archery worked and the weather synoptic for that sea area was known. Then the code breakers could reconstruct the Weather Short Code letters the Enigma operator had keyed in. They had a crib.

    KQJ MUAU JSEL . cipher
    ZYR QSOV KVQM . crib
    So now work out a Menu for the Bombe to break the original Enigma settings.

The first thing to spot is a loop:
S<- d ->U (S goes to U at d), with U<- f ->V and V<- h ->S completing a loop.
The remaining letters all form a chain.
L<- j ->M<- c ->Q<- i ->Z<- z ->K<- g ->J<- b ->R with Y<- a ->Q as a side arm.

The linear form of this Menu to be sent to the Bombe room is:


The code breakers would have had to ask the Bombe operators to run through a very large number of wheel orders. They might have reduced the number to be tried down to under 100 by knowledge of the German usage patterns but even so it was a very large task.

Because the Bombe simulation, written in Javascript, runs a lot slower than the real Bombes, I have cheated and set the wheel order and start close to the actual drop position. You can experiment with a more general attack if you have the time and interest.

So now try running this on a Bombe.

Use Menu 6 which is set to the above Menu. The instructions are under the Bombe picture but you need to click on "test all registers" and "show registers".

What comes out as the Bombe stop is:


This would be considered quite a strong stop. One confirmation, JM MJ, and two self Steckers, VV and ZZ. But why CLI when the enigma machine setting was GWB?

What the Bombe has found is the CORE wheel settings which satisfys the Menu constraints for a Ringstellung of ZZZ. The only way to recover the original Ringstellung is by trying to decrypt other messages for which the relative wheel start positions can be deduced.

However it is a useful exercise to see whether CLI, ZZZ is actually the same core positions as GWB, DKT and a very elegant way of seing this is with Turing's Comic Strips. You can ignore the Steckers, just set the wheels to 163, set the strips to Rings DKT and wheel starts GWC. (Remember that the Enigma machine actually enciphers one letter on from the window letter.) Note the lower case letters on each side of each strip along the datum line, *Z*. Now change the Ringstellung to ZZZ and you find that you have to change the wheel start letters to CLI to bring the same lower case letter in a row across the datum, i.e. to get the CORE positions the same.

You can, of course, just do it by arithmetic operations on the letters but I found the Comic Strips gave a real insight into what was actually happening.

The difficulties of it all.

The above exercise has made it all seem very easy. It was not really like that at all. Firstly the codebreakers had to have the right location for the U-Boat. Secondly Archery had to be working so that they could recover the weather synoptic. Thirdly the weather code letters derived by the Germans had to be correct and in the right order (which they often were not) and last but not least the crib/cipher pair had to yield a strong Menu.

This all shows why hundreds of intercepts were needed just to give the posibility of a strong Menu. In setting up this example I had to try over 50 wheel start positions for enciphering the weather code text before I found the above strong example, and even that gives about 10 drops per wheel order on the Bombe.

It all goes to show IT WAS NOT EASY.

This page was originally created by the late Tony Sale
the original curator of the Bletchley Park Museum