Are the oculomotor disturbances in myotonic dystrophy (MD), i.e. reduced smooth pursuit (SP) gain and reduced saccadic peak velocity (PV), of muscular or central origin? To answer this question the following two approaches were used. (i) The performance of SP was compared with the patient's ability to suppress the vestibulo-ocular reflex (VOR) visually (VOR suppression; VOR-S). In the latter task the SP system is involved, but the eyes hardly move within the orbits. A parallel impairment of SP and VOR-S would indicate a central dysfunction. (ii) Peak saccadic velocity was compared between two saccades performed to and fro in rapid succession. The intention was to measure any myotonic effect which might build up after the first saccade and slow down the second saccade. We studied 15 MD patients and 15 age-matched controls. Stimuli for slow eye responses consisted of sinusoidal horizontal rotations of the SP target and/or the vestibular rotation chair at frequencies between 0.1 and 0.8 Hz. Saccades were analysed in terms of PV. accuracy, duration and latency, comparing centripetal versus centrifugal saccades at short and long intersaccadic intervals (ISI; 400 ms and 900 ms, respectively). The SP gain was reduced in patients compared with the controls, the effect being most pronounced (32% less) at the highest stimulus frequency. Whereas VOR was normal in the patients, VOR-S was clearly impaired (50% worse at 0.8 Hz). Despite normal saccadic accuracy, peak saccadic velocity was significantly lower in the patient group (23% less for saccades of 12 degrees amplitude), similarly for centrifugal and centripetal saccades; all these differences were independent of the ISI. Latency was normal with centrifugal saccades, but was considerably increased with centripetal saccades at short ISI (67% longer compared with controls). The observation of a parallel degradation of SP and VOR-S in the patients is interpreted in terms of a central deficit in the SP pathways. Thus, it appears that slow eye movements were not impaired by muscle dystrophy and myotonia to a considerable degree in our patients. The increase in saccadic latency for centripetal saccades at the short ISI also reflects a central deficit. However, the observed slowing of saccades might have a myopathic or neural origin; a distinction was not possible at present. A myotonic origin of the saccade slowing seems unlikely, because the effect was independent of the presaccadic activation of the relaxing (antagonistic) eye muscle.