“…The reaction to Gram's stain is variable so that even this basic tool for identification is of little help. Carbohydrate-fermentation tests often pose problems of reproducibility (Edmunds, 1962;Dunkelberg, Skaggs and Kellogg, 1970;Bailey, Voss and Smith, 1979;Greenwood and Pickett, 1979;Piot et al, 1980), and although glucose, maltose and starch are almost invariably fermented (Dunkelberg, 1977), this is of little use for identification purposes as these carbohydrates are also fermented by a wide variety of other bacteria (Piot et al, 1980). It has been suggested that the ability of G. vaginalis to produce diffuse P-haemolysis on human-but not horse-blood agar can be used to give a presumptive identification (Greenwood et al, 1977;Bailey et al, 1979;Wells and Goei, 1981).…”