Some Remarks on the Cayley-Hamilton Theorem