COPElAND | 121
the letter that lights up at its lampboard—is fed into the second Enigma as input, just as though
this letter had been typed at the second Enigma’s keyboard. Similarly, the output from the sec-
ond Enigma becomes the input into the third, and so on for the fourth and fifth Enigmas.
In practice, the keyboards and lampboards of the second, third, and fourth Enigmas in
the chain can be dispensed with, being replaced by simple wiring. The lampboard of the first
Enigma and the keyboard of the fifth can also be replaced by wiring. We are left with a single
keyboard, attached to the first Enigma, which serves the purpose of inputting letters into the
chain, and a single lampboard, on the fifth Enigma, which records the letter produced at the
end of the chain. (In the bombe itself, even these two items were dispensed with, but imagining
them to be present will help in visualizing the process for exploiting the loops.) We also strip
out all five plugboards: the replica Enigmas used in the bombe had no plugboards and the input
letter went straight into the wheels.
By turning the wheels of each machine in the chain we hope to drive these five Enigmas
into a loop, so that the same letter that goes in comes out again. Because we want this loop to
match our target loop in the crib, the wheels of the first machine should be one step further
on than the wheels of the second machine (since the loop begins at step 13 and then jumps
to step 12). Similarly, the wheels of the second machine should be six steps further on than
the wheels of the third machine, the wheels of the fourth machine should be eighteen steps
ahead of those of the third machine, and the wheels of the fifth machine should be fourteen
steps behind those of the fourth machine. Programming the bombe involved setting up the
machinery so that, as the wheels (drums) of the first replica Enigma rotated, they remained
one step ahead of the wheels (drums) of the second—and so on for the other replica Enigmas
in the chain.
Since the loop begins with E and ends with E, a simple way of looking at the search process
is this: we continually input E into the chain of Enigmas (by repeatedly pressing E at the key-
board), and we keep turning the wheels of the five machines—while maintaining the relative
positions of the wheels, as just explained—until we manage to put the Enigmas into a loop so
that E lights up at the lampboard at the end of the chain. Once we find the loop, then the posi-
tion of the wheels of the first machine in the chain must be the position that the wheels of the
sender’s machine were in at the thirteenth step of the encryption, and from there it is easy to
backtrack and discover the position of the sender’s wheels at the start of the message.
Although this simple way of looking at things is a good starting point, it is rather too simple—
not least because it ignores the fact that at the thirteenth step of the encryption (the beginning
of the loop) it is not the letter E that goes into the wheels of the sender’s Enigma, but rather
whichever letter E happens to be connected up with at the plugboard of the sender’s machine
(see Fig. 12.4).
IE^510 A23figure 12.11 A loop involving
three letters.