During the lexical translation of a Java source file, all Unicode escapes are transformed to their Unicode character counterpart, that is ch\u0061r a ='a'; is transformed to char a = 'a'; char \u0061r ='a'; and char a = '\u0061'; would also be valid. See JLS 3.2 Lexical Translations for more details.
yup -- if you really really wanted -- you could write your ENTIRE program in Unary characters (that'd really promote maintainability and readability, don't ya think?) But that's just sillyness. However, I would expect a question having to do with it somehow on the exam. You're not expected to memorize all the characters, but you should know that its legal to use them. Just as a real life example, I've used Unary characters in our Application Server when dealing with Internationalization and multi-byte characters (Japanese, Korean, etc).
In ch\u0061r a ='a'; the a before the = sign is the variable name. If you look carefully and replace the Unicode escape \u0061 with the Unicode character a, you should get char a = 'a'; [ February 06, 2003: Message edited by: Valentin Crettaz ]