Logic of computer science (unification algorithm)?

The question is from textbook:
In some implementations of the unification algorithm (e.g. in interpreters for the programming language PROLOG), by efficiency reasons, the test "does x occur in t" is left out (the occurrence check).
Give an example of a 2-element set L = {L1, L2} which is not unifiable. Let L1 and L2 have no variables in common, and (still!) a unification algorithm without occurrence check gets into an infinite loop (or erroneously outputs that L is unifiable- depending on the implementation).

I am just wondering how to answer it.