In this paper a method is introduced, which appears preferable to the classical one, for the embedding of a ring A into a unitary ring. The method is of course in particular efficacious when A is already by itself unitary or when it may be embedded in a field. The idea is due to the following consideration: if a ring is not unitary, it often has elements which behave like integers; they are the "interi virtuali" introduced here in Definition 2 and by means of which the aim is achieved.