From our archives...

November 1943 : Slow Progress in Italy

The rivers and mountains in the approach to the Gustav line, combined with the worst weather the area has experienced for many years, lead to slow and expensive progress as the Allies try to push north to take Rome. On the West of the line General Mark Clark mounts two thrusts attacking Monte Camino. Both fail, and after renewed attempts Clark tells General Alexander that his troops are exhausted and need a pause before continuing. Since crossing the Volturno, a month before, his Fifth Army has suffered 7,000 American and 3,000 British casualties. They are up against the 14th Panzer Corps under General von Senger und Etterlin, brilliantly defending the mountainous territory. On the East coast Montgomery and his Eighth Army make better progress, closing up to the Sangro River on the 8th November. Alexander now takes a more direct hand, preparing a joint push by his two Armies to force the Winter and Gustav lines. He orders Montgomery to strike first across the Sangro to take the pressure off Clark who will attack at the end of the month. Heavy rain causes the Sangro to flood so the attack is postponed until the 27th/28th November. It is led by the 78th Division, and gradually the Eighth Army edges forward, taking heavy casualties. The Sangro is crossed and the Germans slowly fall back, fighting all the way, first to the Moro and then to the Gustav line. Montgomery has logistical shortages because much of the transport to Italy is occupied with material to create the strategic bomber airfields that are being built around Foggia. Some of the best units in the Eighth Army, such as the Seventh Armoured division are now being shipped home to prepare for the invasion, but the Army has been strengthened by the redoubtable 2nd New Zealand Division under General Sir Bernard Freyberg. Often the tactical air force was grounded for days at a time by bad weather. Now Alexander has to accept that the capture of Rome is not going to happen before the spring, though he continues to mount attacks to close with the Gustav line, which passes through the formidable Monte Cassino.

This month, helped by the considerable increase in bombe time available, both in the UK and the USA, Hut 6 breaks three new German army Enigma keys; Woodpecker that covered South East Europe, Magpie used during the fighting in the Dodecanese, and Toucan a supply key used in Italy. Partly due to the difficulty of keeping our garrisons supplied without adequate air cover, the good Intelligence on the German Army offensive to regain control of the Greek islands could not prevent the re-capture of Leros on the 16th and then Samos on the 22nd November, the last island in British hands. This brought to an end the disastrous British attempt to re-occupy the Dodecanese at the time of the fall of Italy. The operation had been mounted in haste, and with too few men to make the operation sustainable in view of the difficulty of re-supply without adequate air cover. The failure did not enhance the British reputation with the US, who had steadfastly refused to get engaged. It was a personal blow for Churchill as it marked the eclipse of his strategy to regain control of the Eastern Mediterranean before the Communist influence could spread in the area.

In Italy itself the German Army Enigma keys were not very productive at this time, due to the use of land-lines. So the Albatross keys used by the staff of the German Tenth army and the Puffin keys used between Italy and the High Command near Berlin were only broken occasionally. But Shrike, used by Army Group B, was broken four or five times a week until the traffic died in the spring of 1944. However the shortage of Enigma decrypts was off-set by the non-morse Fish link, Bream, between Kesselring and Berlin, which was broken regularly by the Testery and Newmanry at BP, though usually with a week’s delay. The Luftwaffe Red, of course, continued to be read regularly and currently, providing some Intelligence about the Army, and other Luftwaffe keys provided some tactical information about the fighting. In the relatively static offensive fighting, good Intelligence was not always of great value, except to the top level planners where the knowledge of the enemy order-of battle was always of considerable value. In particular the Puma key now provided, almost every day, a summary by the Luftwaffe liaison officers with the German Tenth Army of the events on their divisional fronts. From decrypts it was learnt that first priority in transport was being given to the movement of troops from Italy to the Russian front. On 18th November BP decrypted instructions for Montgomery’s old foe, the 16th Panzer division to be transferred to Russia. Another decrypt showed transport aircraft moving from Italy to Southern Russia. Decrypts, probably from Fish messages, the Naval Enigma key, Porpoise, and the new Army Enigma key Toucan, provided excellent information on Kesselring’s tanks, and supplies. And the British Y stations were now performing well, often from plainlanguage broadcasts mixed with simple alphabetical codes. The Official Historian speculates that at this time in Italy ‘the Allies often knew almost as much about the enemy’s formations as he did himself’. It is reported that, in conversation with General Lemelsen, Kesselring commented that ‘the enemy always comes on the boundaries’ to which the Commander of the Tenth Army replied ‘the Devil knows how he always finds where we are’. An Enigma decrypt of 29th November listed all eleven German divisions in the line from right to left. But even if a counter-offensive was now beyond the means of Kesselring, it was all too clear that he remained determined to hold the Allies south of Rome for as long as possible.

The U-boat Warfare in the Atlantic. Following the unacceptable rise in U-boat losses, on 24th May 1943 Admiral Dönitz had withdrawn his boats from the North Atlantic convoys routes. At the end of September the battle resumed, by which time the U-boats had been equipped with various new weapons. The acoustic torpedoes, called Gnats by the Allies, proved successful in their first encounter with a convoy when they sank three escorts and six merchant ships for the loss of three U-boats. But the antidote in the form of a towed raft equipped with a noise generator was ready to be deployed, and prove completely successful. Now the U-boats found it increasingly difficult to detect the convoys as the Germans were decrypting none of the Allied signals, and by contrast their U-boat key, Shark, was being broken regularly and rarely with more than a 24 hour delay. Following continuing heavy losses to the U-boats for only minor losses to the convoys, Dönitz had no choice but to order their withdrawal from the North Atlantic convoy routes for a second time. In a message to his boats on 16th November he spoke of the ‘fiasco of the past month’ and said ‘the enemy has all the trumps in his hand’. He instructed the remaining boats in the North Atlantic to join those operating off the Portuguese coast.

Russia retakes Kiev. The Russians had held a small bridgehead over the Dneiper, north of Kiev at the end of their advance following the German defeat at Kursk. The area was so boggy that the Germans did not expect an attack there. Using the dangerous expedient, one no doubt expensive in both tanks and lives, of charging at full speed across the streams, the Russian Fifth Guards Tank Corps occupied the bridgehead. Secretly the whole of the 3rd Guards Tank Army was moved into the bridgehead. On 3rd November they burst from the bridgehead and overwhelmed the German defenders. By 6th November General Vatutin’s troops had taken Kiev, Russia’s third largest city. They race on to take Fastov, Korosten and Zhitomir, only 75 miles from the old Polish frontier. General von Manstein reacts quickly and attempts to repeat his victory of the previous February 1943 when he had severely mauled the over-confident Russians and had recovered much of the land south of Karkov retaken in the Russian winter offensive. Once again he nips off and destroys the Russian armoured spearheads, but now the Russian back-up troops prove to be much better equipped with anti-tank weapons and inflict a terrible toll on the counter-attacking Germans, who recover Zhitomir but little else of the lost ground. As usual Whitehall had to rely on the decrypts from BP to learn of the true progress of the fighting. At this time BP was having little success against the German Army Enigma keys on the Eastern front, and broke few of the Luftwaffe keys in use there except, of course, for the ubiquitous Red. But since May 1943 the Testery and Newmanry had been achieving a steady rise in decrypts from five Fish links to the Eastern front: Squid to Army Group South, Octopus to Army Group A, Tarpon to the Luftwaffe mission in Rumania, and Trout to the German authorities in Lithuania; and then in August 1943 Perch, the link to Army Group Centre, was broken. The naval Enigma keys in use in the Black sea, Porpoise and, from October 1943, Grampus, were also valuable, especially from the naval liaison officers with the German southern armies.

The Joint Intelligence Committee noted that the Luftwaffe strength on the Eastern front had fallen from 2,000 aircraft in July 1943 to 1,550 in September, with low serviceability, due largely to the withdrawal of aircraft to defend the Reich against the RAF/USAAF raids. And now they noted that eight additional divisions had been transferred to the eastern front, five from Italy and three from France and the Low Countries. [They did not yet know that on 3rd November Hitler had instructed that priority was now to be given to the western front, though a Purple decrypt from the Japanese Ambassador to Berlin would confirm this in January 1944]. It was probably in the second half of 1943 that BP re-established a small team, called ISCOT, to look at the breaking of Russian codes. BP had abandoned reading Russian messages when the Germans invaded Russia in June 1941, but perhaps Churchill felt it was now desirable to get back into the Russian codes.

Progress with Colossus. Tommy Flowers and his team at the Post Office Research Centre, Dollis Hill, have been struggling to build and test the electronic valve ‘computer’ machine, Colossus. It is finally made to operate on about the 25th November, though using dummy traffic rather than the actual Fish intercept material for which it has been designed. Now they will have to disconnect the wiring between the individual ‘P.O. racks’ for the journey to BP. Bill Tutte in the Research Section at BP had proposed, in late 1942, that using a statistical approach it should be possible to find, for an intercepted message, some of the wheel start positions set up by the German cypher clerk on his Lorenz machine. This would be based on the supposition that a particular function of the character stream would be less random when the correct start position is used, because of the number of letter repeats in the German language and the properties of the teletype code. Because of the vast number of possible wheel start combinations that had to be tested, this could only be done using a machine; the Testery was using cryptographic methods, taking advantage of German operator errors and cribs for both wheel breaking, (that is find the positions of the teeth on the code wheels that were changed once a month at this time) and wheel setting (that is finding the start position of each of the 12 code wheels for each message). Max Newman and his small team had started work early in 1943 and the first machine to use a statistical approach to wheel setting had been completed by about May 1943. It was called Heath Robinson and the chief designer was probably Frank Morrell of Dollis Hill, with the high-speed electronic valve counters designed by Charles Wynne-Williams, a circuit expert from TRE, the radar research establishment by then at Malvern. (Charles was seconded to BP and established his own small team for constructing electronic circuits for BP ‘one-off’ equipment). Heath Robinson used two tapes that had to be kept running in synchronism, and was difficult to keep in operation. But it did demonstrate that, as the machine ran through the possible wheel-start positions, the correct start position would be shown up by a higher repeated character count. It was Tommy Flowers who realised that it should be possible to replace the key tape by storing the information it contained within the machine, leaving one tape whose content depended upon the particular intercepted message. And by building in five counters in place of the one in Heath Robinson, the time to find the settings would be considerably reduced. Using the tape sprocket holes to provide clock pulses for the whole machine, and replacing the sprocket drive with a set of friction wheels, reduced the wear on the remaining tape and enabled it to speed up to provide some 5,000 characters per second using an optical tape-hole reading system. (Jack Good, a member of the Newmanry, said that one of the greatest secrets of the war was that normal teletype tape would run at thirty miles an hour without tearing). But the problem was that the Mark I Colossus would use perhaps 1,500 valves. However Tommy Flowers was an inspired choice for designer (said to have been proposed by Alan Turing with whom he had discussed such machine building) because he was probably the one man in the country with considerable experience of obtaining long life from electronic valves in his earlier work at Dollis Hill. Flowers ran the valve heaters up slowly, and never switched them off once they were in service. He designed an extremely flexible machine that proved ready to accommodate the large variety of algorithms that would be tried in the Newmanry. Now Max Newman, who had provided the leadership and specification for the building of Colossus, knew the machine would work in principle. But would it be able to perform usefully on Fish intercepts under real service conditions at BP?

Buildings at BP. The last large building at BP, Block G, was completed at the end of October and has now been occupied by ISOS and ISK, the two teams dealing with the Abwehr messages from Hut 18 and Hut 16 respectively. The two-storey middle section of the Block, adjacent and joined to Block D, is occupied by Sixta, probably to improve their liaison with the adjacent Hut 6 team in Block D. A second storey is being added to Block E, the Communication Centre.

The Bletchley Park Trust welcomes the preparation of these notes, but the authors are responsible for the statements and the views expressed.

How you can help

Do you have any information related to the following question:

'Does anyone know the date in November 1943 when Colossus first operated at Dollis Hill? Why did it take so long to get the machine working again at BP, which apparently was not until near the end of January 1944?'.

Read other example archive documents from Bletchley Park