On the identification of the 0− glueball