Section 4 is a:
keyed Vigenère cipher with a passphrase and an alphabet key.
The passphrase is thehuntingofthesnark
The alphabet key is Boojum (or bojum to avoid duplicate letters)
Here are the steps that I used:
- subtract the plaintext from the ciphertext (mod-26) to try to identify the type of cipher. A majority of the characters are repeating at intervals of 20 or multiples of 20. It looks like a messed-up Vigenère cipher with a likely passphrase length of 20 characters. Therefore, assume that the same 20 character passphrase is being used (thehuntingofthesnark).
- split the plaintext and the ciphertext into 20 character segments and map the ciphertext to 20 rows of alphabets as shown in this example (this specific cipher is also called Quagmire III.)
- and yes, fragments of the passphrase now show up under the plaintext letter B. But, because the passphrase is long, only 25% of the map is populated (130 characters in a 20x26 grid) and the alphabet key is not obvious.
- so back to the puzzle to find words that start with B which are Beaver, Bellman, Baker, Butcher, and Boojum. Boojum gives a match for the ciphertext so the analysis is complete.
Section 6 looks likecould be a Bifid cipher because:
- there is no apparent keylength
- the distribution of letters (IC) is low and close to random
- there are no other cipher methods on Rumkin.com that could provide a similar result, except for Playfair which is not likely because there is no pattern of bigram mapping.
However, all 26 letters are present which is not possible with a 5x5 Bifid. In addition, the plaintext and the ciphertext cannot be mapped to any consistent grid of letters.
This suggests that the Bifid ciphertext has been encrypted again, most likely with a some type of Vigenère cipher. If this is true, I think that it is not possible to replicate the encryption steps unless more information is provided.
Vigenère cipher with a null alphabet or a mixed alphabet.
With an alphabet key of BOJUM, the passphrase would be:oojpzvqdglrszadgpuiyfjxmcvddciqutivnmaljxkazuzeitaizlcxurwenkyiipqhnqsraxaplcvlrhlvqaqsypkaumkrtpzgdqaddamjamadjqcgumkvfsosuhzoorpjsfrjjvol
With a null alphabet, the passphrase would be:ccdvztwinkxqzijccibxbdjfivijipwrfjimggtcwrgzeykifgdzshwrevqmxlqwcvhbpkpghgnshtseoyiogvyycjgdfkepbmqkcgjigfcfggjcoiodfkvmfoseupbbencqnpdduvz
However, these passphrases are the full length of the ciphertext which is no different than enciphering with a one-time pad. So unless, there is some meaning to these passphrases, this solution for Section 6 is likely incorrect.