I suggest you compare it with Edgar Allen Poe’s cipher – there’s a good image here -
http://www.bokler.com/pix/eapoe_crypto_ltr_small.jpgand Zodiacs 3 part cipher
When using multiple substitutes you need to use them equal amounts – the easiest way is to use them in alternation –
if you compare Poe and the Zodiac 3 part cipher both attempted to do this –
For example Poe had 8 substitutes for the plaintext ‘a’ if you track it through the cipher the following are its substitutes
in order of occurrence – (underlined letters = upside down; italicised = reversed)
I G B T e d y e
G B T c d z e
I G A T c d y e
I B c T d y b
I e G B T c d y b e
T d
I G a y e
B T c d b y e
I G
In the 3 part cipher Zodiac had 4 substitutes for ‘a’ – (^ = triangle; ^. = triangle containing dot)
G S L ^
G S L ^
G S L ^
G L S L ^.
G S ^
S
G
G
^Note how both sequences become chaotic towards the end caused presumably by errors creeping in and possibly the
thought by the composer that it no longer mattered.
Enciphering the same message with B & again with C (see above diag.)- I find C fits the bill to a T! – it explains not only the
quantity, the fact there are nearly as many counting backwards as well as forwards, the length of them and their
distribution – it looks exactly like the patterns in the 340.
A simply shows how one might tick them off to keep a record of which ones are already used so as to use equal
quantities to level out the frequencies.
B shows the system used for a ‘sequential or progressive homophonic’ cipher (the one Zodiac used for most of the
3 part cipher)
C shows one method for a ‘non sequential’ cipher
I had mistakenly thought C would be almost equivalent to a randomly generated one – not so – highlighted in yellow
are the ABAB type patterns that occur in the 408.
Note that in C each line is filled up randomly before moving on to the next line – this allows us to reconstruct the
possible alternate substitutes for the same letter, by looking for all the pairs (in any order) that run throughout the cipher
The following are the only ones that run the full length of the cipher and fit the above profile –
You might like to check the above sequences by computer – lacking the necessary programming skills I found them all by
hand (it took ages) so I might have missed a few etc.
The above is from an attempt I made on it some years ago – it may or may not be of some use.
