Thursday, November 20, 2014

The British Typex cipher machine

In 1926, the British Government set up an Inter-Departmental Cypher Committee to investigate the possibility of replacing the book systems then used by the armed forces, the Foreign Office, the Colonial Office and the India Office with a cipher machine. It was understood that a cipher machine would be inherently more secure than the codebook system and much faster to use in encoding and decoding messages. Despite spending a considerable amount of money and evaluating various models by 1933 the committee had failed to find a suitable machine. Yet the need for such a device continued to exist and the Royal Air Force decided to independently fund such a project. The person in charge of their programme was Wing Commander Lywood, a member of their Signals Division. Lywood decided to focus on modifying an existing cipher machine and the one chosen was the commercially successful Enigma. Two more rotor positions were added in the scrambler unit and the machine was modified so that it could automatically print the enciphered text. This was done so these machines could be used in the DTN-Defence Teleprinter Network.

The new machine was called Typex (originally RAF Enigma with TypeX attachments). The first experimental model was delivered to the Air Ministry in 1934 and after a period of testing 30 more Mark I Typex machines were produced in 1937. The new model Typex Mark II, demonstrated in 1938, was equipped with two printers for printing the plaintext and ciphertext version of each message. It was this model that was built in large numbers and the first contract for 350 machines was signed in 1938. Typex production was slow during the war with 500 machines built by June 1940, 2,300 by the end of 1942, 4,078 by December 1943 and 5,016 by May 1944. By the summer of 1945 about 11.000 (8.200 Mk II and 3.000 Mk VI) had been built (1).

The following pics from Crypto museum show the Mark II version on the left and the portable Mark VI version on the right:

The Typex cipher machine was used by British civilian and military authorities during WWII and up to the early/mid 1950’s.

Typex compared to Enigma
The Typex was much more secure than the commercial Enigma device since it had 5 rotors in the scrambler unit compared to 3 in the Enigma D and K versions. The 2 rotors at the right end of the scrambler did not move but could be set by hand. The 3 rotors on the left moved according to the same principle as the other Enigma devices. The ‘fast’ rotor moved with each letter and also caused the other two rotors to move thanks to the notches in their ratchet wheels. Unlike other Enigma devices the Typex had multiple notches per rotor. The commercial Enigma and the improved version used by the German Army and Airforce had one notch per rotor, making wheel movement very predictable. The German Navy had three additional rotors, each with two notches.

The Typex on the other hand had rotors with 3-9 notches. For example (2): 

The use of multiple notches per rotor meant that the ‘fast’ rotor turned the other two movable rotors much more often than in the commercial Enigma or the German military’s version thus defeating certain types of cryptanalytic attacks. The Typex also had an advantage in terms of speed and ease of use since it automatically printed the cipher text. In the standard Enigma models the cipher clerk had to write down the cipher text after depressing each key.
However compared to the German military’s Enigma I  and its plugboard the Typex, despite its multistep rotors, was not superior cryptologically (3). According to ‘The Typex cryptograph’:

The Enigma plugboard offers more security, in spite of its reciprocal nature, than the two stator setup in the Typex. While the Army Enigma and the Typex are roughly comparable in design and cryptographic strength, the German Naval Enigma probably possessed an edge over the Typex due to the introduction of the three "thin" reflector rotors which, in effect, made the machine a four rotor device (the reflector was a stator however).
The same article admits that this comparison is only theoretical since the security of each device depended on the way it was used (indicator procedure, introduction of new rotors etc).

The problems with Typex were:
1). Due to the failure of the Inter-Departmental Cypher Committee to select a cipher machine for mass production and the solitary efforts of the RAF in the mid 1930’s there were only a small number of Typex machines available at the start of WWII. The first contract in 1938 was for 350 machines and it’s doubtful that all would have been delivered by September 1939. Note that at that time the Germans had about 10.000 Enigmas in use (4). 

2). The ability to print the enciphered text came at a heavy (literally!) price. While the German Enigma machine was relatively small and compact, the Typex version built in large numbers Typex Mk II was bulky, weighed 54kg and required electrical supply. Thus it could only be used at prepared sites.
3). For the same reason Typex was too complex to mass produce during the war. According to ‘The Development of Typex’:

Some of the reasons for the low production rate are clear. Any rotor-based machine tends to be very complex mechanically. Figures 2 and 3 illustrate just how many different parts a Typex machine included. Typex Mk. VI contained about 700 parts, few of which were common to other models. Typex must have been a quartermaster's nightmare - much more so than Enigma, because of Typex's printer. Typex's relative complexity proved too much for the British machine tool industry. Overloaded as the industry was with the demands of the war economy generally, it took almost two years to obtain the machine tools required to manufacture Typex, despite the priority that would have been accorded to it. Only 2,300 Typex machines had been made by the end of 1942, 4,078 by December 1943 and 5,016 by May 1944.

4). Because they were complex the machines often malfunctioned (5).

The production problems meant that during the war the Brits did not have a cipher machine in widespread use like the Germans did. For comparison’s purposes at least 40.000 Enigma machines were built by the Germans (6).
Use of Typex by British armed forces

The Typex device was originally developed by the RAF but once full scale production started devices were also requested by the Army. Both the Army and the RAF used the Typex as a high level cryptosystem but due to the production problems they couldn’t replace their codebooks with it. Instead the available devices were used for securing the most important communications.
The Royal Navy was also interested in acquiring a large number of Typex machines and equipping its shore stations and naval units with it (7). In fact the first Navy order in 1939 for 630 devices was three times that of the War Office. Unfortunately for the Royal Navy this plan could not be carried out due to the production difficulties. Instead the limited number of Typex devices were used by shore stations, fleet flagships and landing ships headquarters.

During the war the device was upgraded significantly in terms of security. In that sense it’s possible to differentiate between the ‘simple’ Typex of the period 1939-42 and the ‘improved’ Typex of 1943-45.
In the period 1939-42 it doesn’t seem like Typex was used in a completely secure manner and this was despite Bletchley Park’s centralization and intimate knowledge of Enigma theory.  

Insecure procedures
Up to February 1941 the RAF had two sets of rotors for its Typex machines, the Mk I for higher formations and the Mk II for all units but they both used the same settings. From February onwards different sets of settings were introduced for each set of rotors and during the war many additional ‘keys’ were issued for different geographical areas (Middle East, Med, Home, Empire, India, Australia, Canada etc). Regarding the indicator system (showing the starting positions of the 5 rotors) initially the indicators were not enciphered but then an indicator book was introduced with disguised-true indicators and finally the disguised indicators were further enciphered on Typex. (8)
The Army also used an ‘open’ indicator system till May ’41 (for the UK) and November ’41 (for worldwide users), also at the start of the war there was only one set of 5 rotors available. Another weak spot for the Army was the use of stereotyped beginnings till January ’41, when codress burying was introduced (meaning that the address at the start of the message was moved to the middle before being enciphered) (9).
The Royal Navy used the same settings for both Code and Cypher traffic till August ’40 when different settings were made effective. The settings changed weekly till September 1940 when daily change was introduced. Just like the Army the rotors available at the start of the war were only 5 but in June ’41 two more were introduced. Indicators were sent in the clear till November ’41, when the first edition of the Naval message settings book was introduced (10).

Security over efficiency
Although initially the Typex was not used in the most secure way possible during the war it was significantly upgraded through the use of a rewirable reflector, ‘split’ rotors and several sets of indicator books and ‘key’ settings. At the same time different sets of rotors were introduced for different areas and higher levels of command.  

The rewirable reflector (called plugboard) was introduced in 1941 but it took time before all machines were equipped with it. For example the Navy did not introduce Naval plugboard settings keys till March 1942 and it wasn’t until May 1944 that the three services had enough plugboards to introduce an inter-service plugboard key (11).
The ‘split’ rotors were Typex rotors with detachable rotor cores (called inserts). This way the cores could be switched between different rotors. Also the cores could be inserted in two different ways, effectively doubling their numbers. According to report HW 40/89 the first series of split rotors were introduced in November 1942 (12).

The indicator procedure was further modified, first through the use of disguised-true indicator books and then in 1944 by further enciphering the indicator on the Typex to get the message ‘key’ (similar procedure used by the Germans). The new procedure was introduced in February 1944 and used two disguised indicators taken from the indicator book. The true setting of the first one was used to set the Typex rotors and the true setting of the second one was enciphered on it. The resulting 5 letters were the message ‘key’ (13). 
Many different sets of machine ‘keys’ (meaning the 5 rotors valid for each day and the order they were inserted into the machine) were introduced for each geographical area and also for higher commands.  For example by the end of the war the RAF had ‘no less than 30 machine settings in use, excluding Chief Of Staff and Y settings’ (14).
Finally the introduction of several sets of rotors with different wirings meant that even if one network was compromised in some way this would not affect the security of the other networks. At the start of the war the Army and the Navy had only 5 rotors for their machines but during the war they received several more sets. The Navy got 2 more rotors in June 1941 and in 1943 introduced two sets of 7 rotors each, one for Cypher and the other for Code traffic. The original 7 rotors were then used only for inter-service traffic. The RAF also had different sets for Cypher and Code traffic and it is possible that the Army followed this system too. According to ‘The Development of Typex’ the total number of rotors must have been somewhere between 120-252 (15).

All these measures meant that the ‘improved’ Typex was much more secure than the German military Enigma but this came at a cost, as significant resources had to be allocated for the production of Typex rotors, cores, reflectors etc (including an entire RAF unit with 200 personnel tasked with wiring the rotors) (16).
Also the new indicator procedures and the use of different ‘keys’, rotors and inserts meant that the work of the cipher clerks was negatively impacted. This is clearly admitted in reports ADM 1/27186 (17) and AIR 20/1531 (18), which state:

‘Moreover, Typex operation has been complicated in recent years by the progressive introduction of numerous and tiresome procedures and restrictions which the operator must bear constantly in mind in the interest of security. What the Navy requires, and must press for, is a machine which, whilst providing the highest possible security, is nevertheless reasonably simple to operate and maintain in good running order. The Typex Mark II machine is far from reaching this standard.’

As with book cypher, the method of use of the machine became more complex. Indicators, instead of being self-evident, were first of all chosen from a disguised-true indicator book, and, later, indicators were recyphered in addition. A. plugboard scrambling device was fixed to all machines and the drums were provided with removable wirings, which were changed about 3 or 4 times a month. The effect on output in cypher offices was most marked, not only because encyphering and decyphering became slower, but because more messages became indecypherable, and more time had to be spent on correcting them. The rate at which cipher work could be done, even in large offices, fell to about a quarter of that at the beginning of the war.

Research on Typex by German Army codebreakers
The German codebreakers solved many British cryptosystems during WWII, both high and low level. The most important systems compromised were the Royal Navy’s Code and Cypher, the Army’s Cypher, the RAF Cypher and the Interdepartmental Cypher.

Regarding Typex, several sources claim that the Germans tried to solve it but gave up shortly afterwards because they considered the task hopeless (19). These statements are correct for the codebreakers of the Navy, Airforce, OKW/Chi, Foreign ministry and Forschungsamt. However their Army counterparts did not give up so easily!
In the period 1940-41 the cipher research department of the German Army’s signal intelligence agency Inspectorate 7/VI had several talented mathematicians (Pietsch, Steinberg, Marquart, Schulz, Rinow) tasked with examining difficult foreign cryptosystems. The war diary of Inspectorate 7/VI shows that these individuals investigated the Typex device and by May ’41 had ascertained that it was mainly used by the RAF and was issued with 10 rotors. Their research on its internal cipher operation however was slow and had not led to any breakthrough. Things changed in May when they visited the facilities of the Signal Intelligence Agency of the Supreme Command of the Armed Forces - OKW/Chi and were able to examine a Typex machine captured at Dunkirk. The device worked according to the Enigma principle with the two rotors on the left remaining stationary and the wiring of the entry and reflector wheels could be recovered.

Progress report:

The studies of the English cipher machine Type “X” were continued. A summary report of the present results was written by Dr. Pictsch (see Vol. 45); a more accurate report on the individual results with appendixes is still in process.

The most important result was the information about a model of Type “X” at Chi/OKW. This resulted in the following fundamental characteristics: the machine works after the Enigma principle. The entry wheel and reflector wheel are known. The wiring of the five wheels, which are situated between the entry and reflector wheels and which are selected from a set of 10 wheels, is unfortunately unknown. Two of the five wheels, next to the entry wheel, are always at rest; they are not moved during operation, but can only be adjusted manually from message to message. The middle wheel moves at every step and for every complete rotation it moves several times the second wheel from the left, which in turn at every complete rotation moved several times the first wheel from left. The machine prints plaintext and ciphertext at the same time, it is driven by a motor, which results in very light typing. The ciphertext is written in groups of five, the plain text with word spaces. The wheels are equipped with double rings of contact pins and contact pads, which guarantees better electrical contacts than with the Enigma. Due to these facts, partly also due to suspected facts from the previous work that now has become a certainty, could the further studies be directed in a more definite direction. It was now pointless to search over a larger material of several months for parallel passages. These studies should now rather be limited to a single day’s material.
Investigations continued, with visits to the cryptologic departments of the Airforce and the Foreign Ministry in order to share the available information but at the same time these organisations admitted that they were interested in receiving information but could not contribute much because their resources were already stretched too thin. In the period May ’41-November ’41 the Army codebreakers investigated the stepping motion of the Typex rotors, the indicator system, the first and last codegroups from past traffic and developed theories of solution based however on knowing the wirings of the Typex rotors.

Report of June ’41:

The studies of Type X made progress. Mainly it was the completion of a very extensive and detailed report on the present results. Besides the study into the explanation of the movement rules for the wheels, the transport notches and the wheel order, studies were started on the possibility of a deciphering under the condition that the wheel wirings are known.

Report of July ’41:

The studies of Type X made progress. The detailed Report 1 was completed, furthermore 2 memoranda of which one investigated the question what period lengths Type X will have considering three to four transport notches on the wheels or respectively how long stretches without stepping one can expect for special studies. The second memorandum dealt with the possible reconstruction of the machine settings with given cleartext and ciphertext (for further information see Volume 45).

Furthermore, studies are underway to clarify the question if the last five-letter group of Type X messages always are filled with a certain number of letters and if so by which letters. The previous studies in this direction seem to show that the last group mostly is filled with the cleartext letter X.
Report of August ’41:

The studies of Type “X” were continued. In particular, extensive statistical studies of the last 5 letters and the first 5 letters of the cipher messages were carried out because, as expected from the nature of this cipher machine, particularly frequent cleartext letters will appear very seldom in the ciphertext in those positions. It showed indeed that the plaintext letter X has a clearly visible minima in the last four ciphertext positions, from which it can be inferred that the last five-letter group is filled with the cleartext letter X. Furthermore, the statistics of the first 4 letters revealed that the letter A is most common in the 1st position, the letter I in the 2nd, the letter R in the 3rd and the letter X in 4th place, which seems quite understandable, as the message material mainly comes from the Royal Air Force networks. Further studies are on-going, that as a precaution deal with the question of when and under what conditions material can be solved when knowing the wheel wirings or whether the wheel wirings can be determined under certain conditions.
From the available reports it seems that without the wheel wirings they could not solve any traffic, as the report of November ’41 says:

The studies of Type X were temporarily ended. Some further statistical studies were in addition put down in writing. So far the studies of this machine and the statements of the prisoners show it is pointless to process material enciphered with this machine when the wheel wirings are not known.

The Typex compromise investigation
During the fighting in N.Africa both the German and Italian codebreakers were able to exploit many important Allied cryptosystems. Eventually the Allies secured these systems and from mid 1942 they were the ones solving enemy high level codes. At that time Typex was only available at Corps level and there were no indications that it had been compromised in any way. Yet in May 1943, when the fighting ended with the defeat of Axis troops in Tunisia, two German prisoners claimed that a mysterious officer named Wagner had a Typex machine and was using it to solve British army traffic. The two men were Lieutenant Hanswolf Haunhorst, the intelligence officer of the 334th ID, who befriended the personnel of the signal intelligence unit NFAK 621 (supplying radio intelligence to Rommel) and First lieutenant Werner Possel, head of the senior fixed army wireless station in Africa HeFu 7 (20).

Under normal circumstances their statements would have been dismissed as lies but the fact that both had assignments dealing with radio intelligence and their accurate description of a Typex device (with 5 rotors and two sets of printers) meant that a thorough investigation had to be undertaken to check whether Typex had been compromised.

The investigation focused on locating the intelligence officer ‘Wagner’ and on ascertaining whether a Typex device had been captured by the Germans at Tobruk in 1942. These investigations during WWII were inconclusive however at the end of the war they resumed, since many German cryptanalysts had fallen into Allied hands and could clear things up.

The relevant reports show that Typex was investigated but could not be solved. The most important individuals making these claims were Erich Huettenhain (chief cryptanalyst of OKW/Chi), Walter Fricke (Army cryptanalyst responsible for evaluating cipher security, later transferred to OKW/Chi), Otto Buggisch (Army cryptanalyst) and Ferdinand Voegele (chief cryptanalyst of the Luftwaffe’s Chi Stelle). On the other hand colonel Mettig, head of the Army’s signal intelligence agency in the period 1941-43 stated that Typex had been solved in spring 1942 but later claimed that he had made a mistake and the machine had never been solved (21).

The investigations seem to have concluded in 1947, when US intelligence located the archives of the German army’s signal intelligence agency in a camp in Glasenbach, Austria, where they had been hidden at the end of the war (22). As has been mentioned earlier the war diary of Inspectorate 7/VI shows that Typex was investigated in the period 1940-41 but research stopped in November 1941 due to the problem posed by the unknown rotor wirings.
Bletchley Park and Typex (in)security

One thing made clear by the Typex investigation is that Bletchley Park’s knowledge of Enigma theory had not been used to fortify the Typex machine, at least that was the case till late 1943. The famous codebreaker Gordon Welchman wrote in a report (23): ‘after all we ourselves have made no serious attempt to use the experience of the experts on breaking the German enigma to improve the security of our Type X’.

Another report on the Typex case, dated 16 September 1943 (23), says: ‘Colonel Tiltman said that it had been decided some weeks ago that the time was ripe for a large scale investigation of the security of typex as no comprehensive examination had ever been made, but only ad hoc investigations regarding various individual points as they arose’.

The fact that Bletchley Park with its vast resources had neglected the security of Typex for so long shows that cryptologic centralization does not automatically lead to better outcomes. Moreover the emphasis on codebreaking at the expense of cipher security affected not only Typex but also other important cryptosystems such as the Navy’s Cypher No3. According to the article ‘Tunny Reveals B-Dienst Successes Against the ‘Convoy Code’: ‘GC&CS excelled at breaking the codes and ciphers of the Axis powers, and devoted huge resources to doing so. In March 1942, GC&CS employed about 1.600 people on codebreaking operations, but only Travis (in theory) and Dudley-Smith were then assigned to investigating cipher security, even though Comsec was one of GC&CS’s two main functions. It was clearly too few, especially since Travis had no time to devote to Comsec, and Dudley-Smith was not a cryptanalyst. Even in October 1943, when GC&CS’s staff had more than trebled to over 4.800, only Dudley-Smith (in a ‘part-time’ capacity!) and ‘two or three girls’ worked in the ‘Security of Allied Communications’ section, which investigated the security of the Army’s and Royal Air Force’s signals (and even those of some allies), in addition to the Royal Navy’s signals. Comsec is not as glamorous as codebreaking, but is probably more important.

In the interwar period the British military and civilian authorities knew that they would eventually have to replace their dated codebook system with a cipher machine, as it could transmit more information than the codebooks, in less time and do so securely. However their Inter-Departmental Cypher Committee failed to find a suitable device and in the 1930’s the experimental RAF Enigma cipher machine with Type X attachments was the only available model for mass production. In this case it is ironic that the device they chose to produce was already available in the 1920’s in its commercial version. Thus the period 1926 (when the Committee started evaluating cipher machines) to 1938 (when the first contract for mass production was signed) was time lost for no gain.

The Typex machine was similar to the commercial Enigma but with slight modifications that gave it adequate but not impenetrable security for that era. Had it been produced in large numbers and used by naval warships and military units then several WWII campaigns would have ended sooner and with fewer casualties for the Allies.
Unfortunately the decision to add a printer (two for the Mk II version) meant that it was needlessly complex and thus difficult to produce in mass. The numbers built were enough to equip the higher echelons of command but field units and warships were forced to rely on enciphered codebooks that were read by the Germans. Especially in the Battle of the Atlantic the compromise of the enciphered codebooks gave the Germans valuable intelligence on the routes of the Allied convoys that they wouldn’t have been able to get from other sources.

As for its security, in the period 1943-45 the Typex was used with a rewirable reflector and several sets of ‘split’ rotors, indicator books and machine settings. In order to seriously compromise such a device the enemy would have needed to invest huge resources and build their own version of Bletchley Park, so in that sense the device was secure enough.
Yet Typex security in the period 1939-42 was surprisingly weak and this was despite Bletchley Park’s knowledge of Enigma theory and their work on the German version. TICOM report D-83 (24) admits this failure and points out that one of the reasons that the device was not compromised was ‘sheer good fortune’:

‘However, by about May, 1941, OKH were in full possession of all the theory necessary for solving the problem and there can be little doubt that, had they been lucky enough to capture a set of drums at Dunkirk as well as the three machines the bulk of the Typex traffic up to July 1940 would have been read…..The immunity which Typex enjoyed in the first two years of the war was due partly to the care with which the drums were safeguarded, partly to German inability to grasp the potentialities of the problem, but mostly to sheer good fortune’.

The fact that a global power like Britain took more than a decade in order to select a suitable cipher machine, then made it too complex for mass production and finally had to rely on ‘sheer good fortune’ to keep it secure from enemy codebreakers must be rated as a significant failure in the field of communications security.

Sources: Cryptologia article: ‘The Typex cryptograph’, Enigma Bulletin article: ‘The Development of Typex’, ‘Intelligence and Strategy: Selected Essays’, Journal of Intelligence History article: ‘The Admiralty And Cipher Machines During The Second World War: Not So Stupid After All’, AIR 20/1531 ’R.A.F. signal communications: security’, ADM 1/27186 ‘Review of security of naval codes and cyphers 1939-1945’, FO 850/132 ‘Security of cyphers at posts abroad’, HW 40/88 and HW 40/89 ‘Investigation into POW reports that German Sigint authorities (NFAK 621) were exploiting TYPEX (British cypher machine) in North Africa’, War diary of Inspectorate 7/VI, Crypto museum‘Alan Turing: The Enigma’

(1). Enigma Bulletin article: ‘The Development of Typex’ and Journal of Intelligence History article: ‘The Admiralty And Cipher Machines During The Second World War: Not So Stupid After All’

(2). Enigma rotor pic from Wikipedia user Matt Crypto, Typex rotor pic from Ralph Erskine
(3). Cryptologia article: ‘The Typex cryptograph’:

(5). ADM 1/27186 ‘Review of security of naval codes and cyphers 1939-1945’, p 105-6

(6). Journal of Intelligence History article: ‘The Admiralty And Cipher Machines During The Second World War: Not So Stupid After All’, p3
(7). Journal of Intelligence History article: ‘The Admiralty And Cipher Machines During The Second World War: Not So Stupid After All’

(8). British national archives AIR 20/1531
(9). British national archives HW 40/89: ‘Typex questionnaire’

(10). British national archives ADM 1/27186, p30-36
(11). British national archives ADM 1/27186, p35

(12). British national archives HW 40/89: ‘Typex questionnaire’
(13). British national archives FO 850/132 ‘Security of cyphers at posts abroad

(14). British national archives AIR 20/1531
(15). Enigma Bulletin article: ‘The Development of Typex’

(16). Journal of Intelligence History article: ‘The Admiralty And Cipher Machines During The Second World War: Not So Stupid After All’, p3
(17). British national archives ADM 1/27186, p106

(18). British national archives AIR 20/1531
(19). ‘British intelligence in the second world war - vol2’, p639: ‘Like the Typex, the CCM proved to be totally secure; indeed the Germans made no serious attempt to solve either system’.

‘Delusions of Intelligence: Enigma, Ultra, and the End of Secure Ciphers’, p202: ‘No section made an all-out attempt to solve Typex’.
Intelligence and Strategy: Selected Essays’, p165: ‘In early 1940 German army cryptanalysts did basic research on Typex, which led nowhere’.

(20). HW 40/88: ‘First interrogation report on two German Army officers captured in Tunisia’, ‘Answers to GCCS questionnaire’, The Typex Investigation – A WWII mystery
(21). Various TICOM reports including I-66, I-112, I-16

(23). British national archives HW 40/88

(24). British national archives HW 40/169

(25). British national archives FO 850/171
Additional information:

The files of Inspectorate 7/VI, recovered in 1947 by the US authorities from a camp in Glasenbach, Austria include the following reports on Typex:

Countermeasures against cribbing

As an ENIGMA type device (with a reflector) Typex was also vulnerable to the plaintext-ciphertext attacks used by the Allied codebreakers against the German plugboard Enigma. In order to hinder such attacks several measures were employed, such as burying the address in the middle of the text, cyclic encipherment for short messages and insertion of random letters in the text.

For example report FO 850/171 ‘Preparation of telegrams: use of code words: cypher machines and traffic: teleprinter services: en clair messages. Code 651 file 1 (to paper 4968)’ (25) says:

‘When encyphering on the Typex machine, the encyphered version of a letter can never be the letter itself. This sometimes makes it possible to assign with absolute accuracy even a small number of words known or estimated to be in a message to the actual letters of the cypher version by which they are represented. To obviate this danger operators must from time to time press a key not demanded by the text of the message; the additional letters resulting will make the accurate fitting to the cypher version of a piece of clear text quite impossible. Such an insertion should be made on average once in every 10 words while the body of the message is being encyphered; it should be made on average once in every three words during the encypherment of the codress, the prefatory details and the beginnings and endings, whichever of the methods of encypherment in paragraph 25 is being followed; it should also be made on average once in every three words throughout very short messages when they have to be encyphered separately in Typex (see paragraph 27). The insertion should be made within words and not between them.’

Acknowledgments: I have to thank Ralph Erskine for sharing lots of information on the Typex machine, Frode Weierud for translating the relevant passages from the War Diary of Inspectorate 7/VI and Randy Rezabek for TICOM report IF-272, listing the Typex reports of In 7/VI.


  1. You write: "when US intelligence located the archives of the German army’s signal intelligence agency in a camp in Glasenbach, Austria, where they had been hidden at the end of the war".

    Do you have a reference for where these archives ended up?

    Glasenbach was also home to an Allied internment camp in the years immediately after WWII, is that connected somehow?

    1. It could be. Several reports state that the Inspectorate 7/VI archives were found in Glasenbach. For example check:

  2. Did the fact that the Type-X had a reflector rotor mean that it shared with Enigma the very serious flaw that no letter could be enciphered to itself?

    1. That’s correct. The book ‘Alan Turing: The Enigma’ references report FO 850/171 ‘Preparation of telegrams: use of code words: cypher machines and traffic: teleprinter services: en clair messages’:
      ‘PRO file FO 850/171 contains a memorandum of May 1945, from the Cypher Policy Board to the Foreign Office, with instructions for use of the Typex. It explains that 'When encyphering on the Typex machine, the encyphered version of a letter can never be the letter itself. This sometimes makes it possible to assign with absolute accuracy even a small number of words known or estimated to be in a message to the actual letters of the cypher version ...', and gives procedures for burying addresses’.