This is a good, well-written book requiring essentially no mathematical background. It's appropriate for teens and older folks interested ciphers. As another writer has pointed out, it's not current (this is a Dover reprint of a 1972 Simon & Schuster work), and the very few places where it says a code is still in use, it's likely not. I don't view this as a problem.
Ciphers are categorized and historical development is given. Invisible inks and the like are discussed, as are microdots, and SETI, though not by that name. Modern ciphers get no mention whatsoever--thus the non-mathematical nature. The approach doesn't consider computers to any real extent. Also, certain historical items that could have been covered, like the Enigma, aren't mentioned.
There are three weaknesses, IMHO, that keep it below 5 stars.
(1) Most of the ciphers presented have a set of possible setups, which can effectively be considered keys. E.g., the Caesar Cipher, using the Roman alphabet, has 25 different possible versions, which can be considered 25 different keys. Gardner makes no attempt to explain the relative complexities of breaking the various ciphers.
(2) The age.
(3) There is no index, but the table of contents is detailed.
I expect to use a few of these ciphers in introductory computing classes (think CS1, CS2) in the near future--the explanations are clear enough for undergrads with no real background and minimal interest.