100 | 11 BREAkING mACHINES wITH A PENCIl
Dilly encouraged Turing to think about mechanizing the breaking of Enigma messages, and
this would be Turing’s great achievement. They both knew, however, that Turing’s ‘bombes’
were able to break the daily Enigma settings only because the wiring of the Enigma machine
had already been discovered by paper-and-pencil methods—thus Dilly’s greeting to me and the
title of this chapter.
The race to break Enigma
Enigma was first taken seriously as a potential German war weapon by the Poles, and by
Section D of the Service de Renseignements (French military intelligence) whose function
was to acquire foreign cryptographic material—or, more precisely, taken seriously by a lone
Frenchman, Gustave Bertrand, the head of Section D. Bertrand, who was concerned about
the belligerent noises coming from France’s powerful neighbour, was the first to make use of
cryptographic espionage, in 1932. By devious means, and for a considerable sum of money, he
managed to acquire photographs of Enigma manuals and other documents, including invalu-
able key lists for September and October 1932. Hans-Thilo Schmidt, a traitor in the German
war ministry, took the photographs and sold them to Bertrand. Neither Bertrand’s own gov-
ernment nor the British government showed any interest in this ultra-secret material, but the
Polish government certainly did.
At this time Britain was more concerned about communist subversion than with the rise
of fascism, and all our codebreakers were working flat out on Bolshevik codes. The Poles took
Hitler’s threats more seriously, and moreover could intercept the Germans’ low-power radio
transmissions, which we were unable to do at that time. The Polish cipher bureau recruited
three mathematicians from Poznán University and one of them, Marian Rejewski, succeeded
in solving the Enigma machine (Fig. 11.3). He had observed a pattern in the doubly enciphered
indicator groups at the beginning of the messages, and he used mathematical methods, together
with Schmidt’s photographs of the key lists, to solve the machine’s wiring.
Although Dilly had had such success during the Spanish Civil War, he lagged well behind
Rejewski when it came to German military Enigma. When breaking codes, enemy errors are
manna from heaven, and Dilly worked on operators’ errors to break Enigma messages during
the Spanish war. In the case of German Enigma, he received help from no less a source than
the German cipher office itself. In 1930, when the Germans had introduced their plugboard
machine, they issued a manual showing operators how to set up their machines, using the
daily key.^3 This manual gave an authentic ninety-letter example of plaintext with the result-
ing ciphertext, along with the machine settings that produced it. It was an almighty blunder
on the Germans’ part. This authentic example was instantly withdrawn once the blunder was
discovered, and a fictitious example was used in place of this monumental crib. Usually cribs
were the result of patient guesswork by the codebreakers, and were often speculative and quite
short—but this long crib was certain, since the German cipher office had supplied it.
The GC&CS received this manual from the French, probably in 1938, and Dilly attacked
German military Enigma using the monster crib. He had no success, however, because there
was still a fundamental problem to be overcome. When Dilly had first examined the commercial
Enigma machine, he saw that the way the letters of the keyboard were wired to the entry plate
of the machine (see Fig. 12.4) was simply in the order of the letters on the German QWERTZU
keyboard. All his methods were based on this discovery, and fortunately the same wiring held