On a classical correspondence between K3 surfaces II