I found that a very interesting read.
I’m sad my grandmother never knew of this, she was absolutely obsessed with decoding encrypted messages and she probably would have loved to have had a crack at those ones.
It would take so many workers to construct such a "vault" so deep in the ground and move over four tons of metal that there is no way this is true or still exists, some worker would go back and steal it eventually or tell someone about it.
It's not like the pharaohs where they'd murder all the slaves that built their pyramids.
You exaggerate and brag to your friends at the bar, not in an encrypted note precisely detailing the contents of a vault.
It's also just barely at the edge of possible you could trade silver for $13,000 in St Louis in 1819. The population was somewhere between 1600 (1910 pop) and 5000 (1930 pop) [by 1880, nearer when the pamphlet was written, St Louis was the 6th largest in the US].
3000 people in a frontier city isn't a lot to find jewels worth ten years of a skilled Craftsman's wages, but two of those people were the sons of the fur trader who founded St Louis, whose estates were worth > $100,000 a few years later. I'm not sure they'd have $13,000 worth of jewels lying around, but hey, not completely impossible.
It would be feasible to construct a ciphertext that encodes two plaintexts, depending on which document is used to decrypt it.
To cover all possible plaintexts, you'd need 576 word indexes, representing each pair of letters AA, AB, ..., ZZ. It might be hard to find a number i such that the i'th word of the Declaration of Independence begins with Z and the i'th word of the document you are using as the other key begins with Q, etc. But with some license to tweak both plaintexts, this seems within the capabilities of a motivated 19th century prankster.
That said, I can't think of any legitimate reason to use this encoding, so it still points to a hoax or prank.
> attempting to use the Declaration of Independence again on the first note yielded several sequences along the lines of ‘abfdefghiijklmmnohpp’.
If the two un-decyphered noted use the same approach as the decrypted one, a couple of idle thoughts:
I wonder if anyone has tried the same text but with positive/negative word index offsets?
It would presumably be hard to extract a plaintext unless they are based on different texts that have a canonical form. Many books differ slightly between printinga due to errors and/or corrections.
