The reliability of a scale of 0 to 4 (where 0 is normal) in rating the severity of blepharospasm (BS) and oromandibular dystonia (OMD) was evaluated by the concordance among six neurologists from different neurological institutions. As expressed by k index, interobserver agreement was moderate either for BS or for OMD according to the Landis classification. Neurologists showed different rating attitude toward BS and OMD. In fact, the category analysis showed that raters were inclined to overestimate BS and to underestimate OMD. The familiarity with dystonia influenced reliability more than the length of professional experience in neurology. In fact, when examiners were subdivided into subgroups (each of three raters) according to the former criteria, the level of interobserver agreement increased significantly. Almost perfect agreement was obtained in intrarater comparisons. These results may be of value with regard to the conduct of multicenter epidemiologic and clinical studies on focal dystonias.