How was the Zimmermann Telegram encrypted? Why was it so easily decoded by British Intelligence?

by throwittomebro
thefourthmaninaboat

The Zimmermann Telegram was encrypted in two separate, but similar codes, known as 0075/7500 and 13040. Both of these codes were, essentially, numerical substitution ciphers. In both, each individual word in the plaintext was replaced with a number, also known as a code group. To encipher and decipher it, code books were used. These contained a vocabulary, giving a list of commonly used words and their equivalent code group. The main difference between the two codes was in how words were assigned code groups. In the older, simpler, 13040, a code group was 4-5 digits long. The first two or three gave the page number on which the corresponding word was to be found, the remaining two the position on the page. The words were originally written in alphabetical order, and numbered accordingly, before the pages, with 100 words on each, were shuffled. Each group of ten words on a page would then be shuffled randomly. In 0075, the code groups were four or five digits long. They were assigned randomly to to their corresponding words. Both codes could also have a superencipherment (an additional step in the encipherment). In this, a number was hidden in the ciphered message. The digits of this number would be added to each code group by some scheme - in 0075, the superencipherment method used a seven-digit number, with each digit being added consecutively to each code group, going back to the first digit after the seventh group.

The Zimmermann Telegram was first sent, in code 0075, to the German ambassador to the United States, Count von Bernstorff, as an addendum to a longer message dictating what actions Bernstorff was to take to avoid American entry to the war following the German resumption of unrestricted submarine warfare. This message was sent, like many other messages sent to German embassies in the Americas, over an American telegraph cable - the British had destroyed almost all other telegraph cables leaving Germany. This cable ran through the UK, and so Room 40, the Admiralty's signals intelligence organisation, were easily able to intercept it. Once it arrived in the United States, Bernstorff's staff re-enciphered the telegram in 13040, and sent it on to Mexico City for the German resident there, von Eckardt (who did not have the 0075 codebook). Here, the British were also able to acquire a copy of the ciphertext, likely using contacts within the Mexican telegraph infrastructure.

Neither of the codes used to encipher the Zimmermann Telegram were especially new. 13040 had been in use since 1907, while 0075 had been in use since 1915. As such, the cipher experts of Room 40 had been able to build up a considerable amount of knowledge on the two codes. Code 13040 was almost completely broken, while Room 40 had partially reconstructed the code book for 0075. They had used a number of methods to do so, including comparisons to captured (naval) code books, and by exploiting the semi-alphabetic nature of 13040. When the Zimmermann Telegram was intercepted, the first decipherment was attempted by two of Room 40's codebreakers, Nigel De Gray, and Dilly Knox. They first filled in any previously known groups, before attempting to make educated deductions about the remaining groups, or to spot repeated groups that might betray a new word. They could not fully decode the telegram, but were able to discern the general gist of it; significantly, they could not decode the group 0979, the code group for Mexico. What they could read was that Germany was proposing an alliance with some other country to attack the USA together. As Mexico was the only country in any real position to do so, it was a reasonable conclusion that it was the intended recipient. To maintain the secrecy of Room 40, and to produce a more convincing decrypt for release to the USA, it was decided to obtain the copy, enciphered in 13040, that had been sent to Mexico City. This was done with relative ease, and the message arrived at Room 40 on the 19th February 1917. As Room 40 could read most of 13040, it was straightforward to decode this copy.