Iterating multiple substitution ciphers is not mathematically different from using a single substitution cipher (3 Vigenere iterations with a 31-, a 32-, and a 33-letter key would essentially be a single Vigenere cipher with a 32,736-letter key), and a long enough message can be decrypted even with this long key the same way that a shorter message can be decrypted with a short key (likewise for iterating multiple columnar transpositions).
But if I wanted to use 3 Vigenere ciphers and 3 columnar transpositions on the same message, would doing all 3 substitutions first and all 3 transpositions second (or vice versa) be less secure than alternating something like "Vigenere-column-Vigenere-Vigenere-column-column"? Or would the same combined-substitution-transposition decryption techniques apply regardless of the order?
