The purpose of this study was to evaluate interrater and intrarater reliability of the Fahn-Tolosa-Marin Tremor Rating Scale (TRS) in essential tremor (ET). Proper treatment of ET is contingent upon correct assessment of the severity, loss of function, and disability related to tremor. Videotape recordings of 17 subjects with ET evaluated with the TRS were produced and sent to 59 raters. Once the raters returned the videotape and completed the score sheet, they were mailed a second tape with the same recordings presented in a different order. In the interrater reliability evaluation, modified Kappa statistics for seven tremor type composites ranged from 0.10 to 0.65 in the first videotape and 0.17 to 0.62 in the second videotape. Interrater reliabilities were greater for Part A items (magnitude of tremor in different body parts) than for Part B items (tremor in writing and drawings) of the TRS. The average Spearman correlation was 0.87, indicating very good consistency between the two videotapes, but correlations for Part A were somewhat better than for Part B. It is best when the same rater performs repeated measures of tremor on a patient, particularly when judging tremor in handwriting and drawings. Training of raters on use of the TRS would help standardize judgement.