Scharnhorst’s Last Message?

The ciphertext message below is suspected to be the last message from the German battlecruiser Scharnhorst. The message was intercepted by one of the the radio officers, D. Rough, on board the English battle ship Duke of York in the evening of 26 December 1943. The message, which was transmitted in Morse code, was received on a HRO receiver and written out on the customary W/T Red Form. The Red Form has been found among the papers of Edward Thomas who was the intelligence officer on board the Duke of York. Today the Red Form is owned by the collector Mr. Colin Waghorn and I am grateful for his permission to publish this historical document here.

In December 2007 Geoff and I were contacted by the Bombe Rebuild Project to hear if we could help with breaking this message. The rebuilt Bombe at BP did not have the necessary Naval Enigma rotors and they wondered if our E-Breaker program would be able to do the job. Unfortunately, we don’t have a Naval Enigma version of this program and both Geoff and I were too busy with work and other tasks to envisage starting yet another codebreaking project. We kindly asked for permission to give the message to Dan Girard, who since then has been trying to break the message with a Bombe simulator using cribs from the plaintext message. Here is a report on his attempts so far:

I’ve had no success with the Scharnhorst message. I’ve tried the vierneunneunzwo crib in various positions; also steuereauftanafjord, steuereaufjtanafjordj, steuerenachtanafjord, steuerejtanafjordjan, etc.; and, at the beginning of the message, the signatures vvvjscharnhorstj, vonvonjscharnhorstj, and vonvonscharnhorst. I’ve also tried mueckevvvjscharnhorst as a crib.  The other urgency codewords, “biene” and wespe, and the variants bine, “mucke” and muke all result in crashes when placed before the signatures, so I didn’t try those.
Nothing has worked. Of course, until the message is broken there’s no way of knowing whether this was due to garbled ciphertext letters or because the cribs were either incorrectly aligned or wrongly worded.

A good account of what happened during the Scharnhorst battle has been given by Patrick Beesley in his book Very Special Intelligence. Here is an extract from page 216 that describes the final hours of the battlecruiser Scharnhorst and the transmission of its last message:

At 4.17 p.m., Duke of York gained radar contact with the enemy and at 4.50 opened fire. Once again Scharnhorst was taken by surprise. At 4.56 she signalled to Gruppe Nord-Flotte: ‘Most Immediate. 72 degrees 39 North, 26 degrees 10 East. Heavy battleship. Am in action.’ Bey evidently mistook some of Burnett’s cruisers for battleships because at 5.22 he signalled that he was ‘surrounded by heavy units.’ The situation soon became hopeless. At 6.02 a signal went off to ‘Admiral of the Fleet and the Commander-in-Chief. Scharnhorst will ever reign supreme,’ and this was followed at 6.25 by one to Hitler, ‘We shall fight to the last shell.’ Scharnhorst’s last signal was made at 7.25, stating that she was steering for Tanafjord at 20 knots. It seems unlikely that she was still capable of this speed because, shortly before she had reported it as only 15 knots, and within another twenty minutes, after repeated hits from Duke of York’s 14-inch guns, from the lesser armament of the cruisers and from eleven of the many torpedoes fired at her by the British destroyers, she sank. Despite every effort only thirty-six survivors could be rescued.

As explained in Very Special Intelligence most of the messages from Scharnhorst and Admiral Northern Waters were quickly broken and decoded at Bletchley Park (BP). The messages were enciphered on the 3-rotor Naval Enigma machine M3 and most of them were using the Allgemein (general) settings from the Enigma key Heimisch. Heimish, later renamed Hydra and called Dolphin by BP, was the Enigma key used by U-boats and surface ships in German home waters, in the North Sea, and the Atlantic. A few of the signals were enciphered using the Offizier setting of the Heimisch key. These signals were usually decoded but with a delay of two to three weeks. One Offizier signal that was transmitted on 18 December 1943 and that instructed the Battlegroup with Scharnhorst to take ‘preparatory measures so that departure would be possible at any time’ seems to have been broken with a delay of only two days.

Ciphertext and suspected plaintext

W/T Red Form

Ship or Station: | Set: H.R.O. No. 2 | Date: 26.12.1943 | Operator's | Opr.: D. Rough. | Time : See below | Remarks: HMS "Duke of York" | | Ended | QSA: 5 | To: | Frequency & System: | Last transmission | From: | 6475 kc/s | from Scharnhorst -------------------------------------------------------------------------------------- 1) de KR ANA KR.KR AN ALLE LA ( Time ended: 1823Z ) 2) Priority dots. 1925 21. UTKZ RBSB YKAE NZAP MSCH ZBFO CUVM RMDP YCOF HADZ IZME FXTH FLOL PZLF GGBO TGOX GRET DWTJ IQHL MXVJ WKZU ASTR UTKZ RBSB ( Time ended: 1832Z. Erratic morse. Nil further heard. )

BP Decode Teleprint


Facsimile of the Red Form and the BP Decode (PDF format)

The Red Form shows two messages labeled 1) and 2). The first message seems to be an emergency warning to everybody (An Alle) that a KR KR message – Kriegsnotmeldung (war emergency message) was to be expected. The regulations for the use of a war emergency message was very strict: Kriegsnotmeldungen dürfen nur bei unmittelbarer Gefahr und höchster Notlage für Schiff und Besatzung abgegeben werden — War emergency messages must only be transmitted when immediate danger and the highest emergency exist to ship and crew. The transmitting station uses the callsign ANÄ, which is an Umlaut callsign — Umlautfunkname. These callsigns, which could not be enciphered, were usually given to fixed land stations. If it was exceptionally given to Scharnhorst for this engagement is not known. Nor is the meaning of LA understood. It could be a frequency reference as the German naval frequencies were given two letter designators, however the frequency lists I have access to shows la as 109 kHz and the frequency 6475 kHz is not listed.

The second message is the suspected last message from Scharnhorst. It starts of with some priority dots and then gives the time of origin, 1925, and the number of cipher groups, 21. A quick look at the message shows that it has 24 groups. If the error in group count is due to the cipher operator on Scharnhorst or the radio operator on the Duke of York is impossible to say. The message is in the standard Naval Enigma format with the two first 4-letter groups being respectively the Schlüsselkenngruppe (Cipher Indicator Group) and the Verfahrenkenngruppe (Procedure Indicator Group), both being repeated as the two last groups of the message. For further information please see the General Naval Enigma Procedure – Der Schlüssel M Verfahren M Allgemein.

As can been seen from the BP decode the last message from Scharnhorst was decoded on 13 January 1944, almost three weeks after it was transmitted. What does this mean? The most obvious reason is that this message was transmitted in the Offizier key and BP had difficulties in breaking this key. As the battle was over there would not have been a great urgency in breaking the message and this could also account for the delay. Another possibility is that BP’s intercept stations failed to receive this message and that they first got a copy after the Duke of York was back in harbour. In this case the decoding of the message could also easily have been delayed by two, three weeks.

Message Broken

The M4 Message Breaking Project started a break on this message on 27 May 2008 at 22:39:21. Less than 24 hours later, on 28 May 2008 at 18:16:21, the message was broken. The German plaintext does indeed correspond to the English plain text in the BP teleprint. The raw German plaintext is:

	steuere j tanafjord j an standort qu aaa ccc vier neun neun zwo
	fahrt zwo nul sm xx scharnhorst hco

Written out in full it becomes:

	Steuere Tanafjord an. Standort Qu. AC 4992,
	Fahrt 20 sm [Seemeilen].

The Enigma key for Scharnhorst’s Last Message:

	UKW	    : B
	Wheel Order : 368
	Stecker	    : AN, EZ, HK, IJ, LR, MQ, OT, PV, SW, UX
	Rings	    : AHM
	Message key : UZV
The full details are given in the M4 Project Report on the Scharnhorst break. By studying Ralph Erskine’s notes on Breaking Naval Enigma (Dolphin and Shark) and The Kenngruppenbuch Indicator System, and using Geoff Sullivan’s Enigma M3 or Enigma M4 simulators you should be able to verify for yourself that the decode is correct.

Dan Girard was indeed only a hair’s breadth from breaking this message with his Turing Bombe simulator. Here is his story:

Since steuerejtanafjordj is one of the cribs I had tried with my bombe simulator, it’s obvious that I must have made a mistake somewhere; and sure enough, on re-checking the menus I had drawn up for this crib, I find that I had misread one of the links from the diagram I had made. I had drawn up seven different menus from the crib, in order to allow for different turnover possibilities; and it was just my bad luck that the only one of the seven on which I made the mistake was the one which had the right turnover assumption. This is embarrassing!
I’ve checked on the vierneunneunzwo crib, and it turns out that one of the positions I tried it in was the correct one; but unfortunately for me I had assumed that since there was a comma between 4992 and speed in the translated decrypt, there would be one in the German plaintext, as vier neun neun zwo y fahrt...; and I included the y in the crib. I’ve just re-run the relevant menu, omitting that link, and found that it would have worked.

This shows yet again that selecting cribs and constructing menus is indeed more of an art than a science; an art where the codebreakers at Bletchley Park were the real masters. We should all stand in awe of their achievements.

Closing the circle

Yesterday, 30 April 2010, I received the following good news from Paul Kellar at BP:

Hello everybody

This is the last link in the chain from the original Red Form.
We have made the rotors 6,7 and 8 for our Typex (which, as you know, we have previously
converted to behave like an Enigma, as they did at BP).
This would output both the original and the decrypt on paper tape which, like a telegram
was gummed to a message form. It groups letters by fives, as you see.
Using the settings from Frode's group we have decoded the original message to reproduce
the decoded output as seen 66 years ago.

Thanks, everybody!


And this is how the Scharnhorst messages would have looked like when it was decoded at BP in January 1944.

Scharnhorst Typex Decode 


We should first of all like to thank the Bombe Rebuild Team who immediately thought of us when they realized they were not in a position to attack this message. Of the team we should like to particularly thank John Harper, Paul Kellar, Mike Hillyard, and John Borthwick, who sent us the copies of the Red Form and the teleprint with the BP Decode. Special thanks goes to John Gallehawk, who through his thorough and persistent search in the National Archives at Kew discovered the BP Decode teleprint. Futhermore, we are most grateful to Dan Girard who stepped into our shoes to do the work Geoff and I had neither the time nor the energy to do. Finally we should like to thank the owner of the Red Form, Colin Waghorn and his partner Dr. Rebecca Hill. And as always we are most grateful to have Ralph Erskine as our friend and mentor, who is always there to hold our hands when the sea gets rough.

Further Information

For more information about our codebreaking projects see our Web Portal: Breaking German Wehrmacht Ciphers.

All rights to original material described on these pages belong to
Geoff Sullivan & Frode Weierud, © March 2010

 Back to Frode’s CryptoCellar Site visitors:

  Breaking German Wehrmacht Ciphers