The authors developed two scales to be adopted for the evaluation of myasthenia gravis (MG) patients. The first scale (MG impairment scale) is based on objective patient evaluation and on patients' responses to standardized questions relating to the functioning of specific muscle groups. It consists of 13 items exploring strength and 10 items exploring fatigability. The second scale (MG disability scale) evaluates disability in those everyday activities that are often impaired in MG patients. Test-retest reliability of each item and of the global score (sum of single item scores) was assessed by the weighted K statistic and by the intraclass correlation coefficient. Reliability was invariably 'substantial', and for single items 'almost perfect' for the MG impairment scale, and invariably 'almost perfect' for the MG disability scale. The internal structure of the MG impairment scale was explored by means of the principal component analysis. This analysis resulted in three main (rotated) factors, which loaded respectively onto 'ocular', 'spinal' and 'bulbar' functions. For these factors, we report factor score coefficients that can be used to compute single patients' scores, which in turn may be used in further analyses, particularly for follow-up studies. We also report the results of an analysis of the correlations between the two scales. The MG impairment and the MG disability scales are proposed for application in both clinical and research settings.