Tizanidine (Sirdalud) was compared to baclofen (Lioresal) in a randomized, double-blind, cross-over trial. Each medication was introduced over a three week titration period and then maintained at the highest tolerated dose for five weeks. The two treatment phases were separated by a one week drug withdrawal and a two week washout period. Sixty-six patients entered the trial and forty-eight completed both treatment phases. At the end of the trial, neurologists and physiotherapists thought that baclofen was superior on the basis of perceived efficacy and tolerance (p less than or equal to 0.05). Although the efficacy of tizanidine or baclofen was judged as good to excellent by 24 and 39% of patients respectively, this difference was not statistically significant. Muscle weakness was the most common adverse effect. This was significantly more troublesome in patients treated with baclofen. Somnolence and xerostomia were more common in patients treated with tizanidine. Both baclofen and tizanidine appear to be useful adjuncts in the treatment of spasticity in patients with multiple sclerosis. Preference of either drug is tempered principally by side-effects.