he had mild limb and gait ataxia. There was no muscle atrophy nor weakness, and tendon reflexes were normal; the plantar response was flexor. There were no involuntary movements, no sensory nor autonomic disturbances. Electro-oculographic findings were almost identical to that of his brother.The main clinical feature of our patients was adult onset slowly progressive cerebellar ataxia, thought to be hereditary as there were no apparent causes including metabolic diseases. The most conspicuous finding of our cases was an abnormal ocular movement during eye closure. It was a rapid irregular multidirectional oscillation with no intersaccadic intervals, and appeared to fulfil the criteria of opsoclonus. To our knowledge, the appearance of opsoclonus, only during eye closure, has not been previously described. Opsoclonus can be caused by an abnormal activity of burst cells and pause cells in the brainstem saccadic pulse generator.' Hain et a12 described a case with neurodegenerative disease showing bursts of ocular flutter during blinks, and attributed it to a blink-related reduced discharge of pause cells. As ocular flutter and opsoclonus form a continuum, frequent aggravation of opsoclonus at attempted fixation3 seem to be due to a similar neural dysfunction. Opsoclonus only during eyeclosure could be explained by the presence of eye-closure related abnormality of saccadic pulse generator, and increased opsoclonus during eye closure3 seem to support this speculation.Square wave jerks, saccadic dysmetria and saccadic pursuit seen in our cases seem to be common ocular motor abnormalities in hereditary cerebellar ataxia.4 Opsoclonus, however, has not been clearly described in hereditary cerebellar ataxia the absence of which was even thought to be a characteristic finding.