A null or an insignificant letter/symbol is occasionally used in ciphertext but does not mean anything. It is used to confuse interceptors.
Use of nulls varies. Use at the beginning or the end of text is effective because these are parts where fixed phrases tend to be used, which may provide a crib for codebreaking. (Still, symbols used exclusively at the beginning or the end can be easily detected as nulls.)
Nulls are often used at word breaks but it may not be very effective as a protection. To the contrary, some Italian ciphers which employ variable-length symbols recommend use of nulls to aid deciphering by the recipient. (See another article.)
Use of nulls at random places within the spelling would be a good practice and have been employed by some people. (Docteur Mauclerc mentioned below is one. Colonel Jardet (see another article) from the Napoleonic age is another.)
A null may also be used within a symbol, that is, between digits/elements forming a symbol that represents one letter or word. A null used in this way may be termed an intercalary null.
Matteo Argenti, who was cipher secretary in the curia from 1591 to 1605, suggested inserting a null within a two-figure symbol (e.g., writing "374" or "3784" instead of "34", with "7" and "8" being nulls) (Meister p.62).
In the cipher no.45 of the Nevers collection, instructions advise that "1" and "0" as nulls should be put in the middle of a two-figure code number as 7831452807, which is parsed as 78(u) 314(o) 52(u) 807(s) ("34" is "o" and "1" is null; "87" is "s" and "0" is null).
In letters of docteur Mauclerc, a null symbol is inserted between the digits of a code group such as "22" (legat) or "49" (espagnol). A code group representing "Guise" (something like "13do") is in many patterns with nulls inserted in various patterns.