Security of RAF Cipher Machine (Type-X)

The following is a memorandum to "C", the head of the Secret Intelligence Service, from the Government Code and Cypher School recommending that SIS and the Foreign Office do not use the RAF cipher machine instead of one time pads for encrypted communications. Presumably the RAF machine cipher referred to in this memo is the Type-X.  GC&CS are concerned that as they are reading German Enigma machine ciphers, the RAF cipher machine could also be breakable in a similar way. Indeed a few years later, a German prisoner of war claimed that the Wehrmacht had been reading Type-X messages sent during the North African campaign for a short while.

THE DIRECTOR.

I am rather surprised that this matter has not been raised through the Communications Department.

Whether the Civil Departments should adopt the RAF Machine has been discussed by the Treasury Committee on Cypher Machines and at other times and has been turned down because

  1. It would need a maintenance mechanic at each post.
  2. There might be considerable delay in distribution of spare parts.
  3. It very considerably increases the length of the telegram (3 times) and therefore adds to congestion and very greatly to cost.

At the present moment the supply of the machines is very much behind requirements of the War Office and Admiralty and the supply of cipher drums for existing machines even more so. The position in regard to the latter is becoming serious.

Our experience proves that the machine is vulnerable to a verbatim crib and these are far commoner and much easier to place in diplomatic than in Service traffic. On the other hand the pad subtractor tables ("One time" or "Individual" Tables) are entirely proof against cribs and any form of cryptography.

The machine also is subject to compromise through espionage, an agent, instructed in what to do, can obtain data for breaking fairly simply if he can get access to the machine, the pad subtractor with its very limited distribution and hundreds of pages to be photographed presents a much harder problem.

For these reasons therefore I would hesitate to introduce the machine for your work or the Foreign Office at present.

As regards the latter I know very well that their difficulty is staff both in quantity and quality. This should not be an insoluble problem and given adequate numbers of men or women of the right type I see no reason why an efficient and rapid service should not be provided by their existing cipher methods.

20th January, 1941.

[Source: TNA FO 1093/314, transcribed by www.arcre.com]