In addition to many other symptoms, Huntington's Disease (HD) also causes an impairment of oculomotor functions. In particular, saccadic eye movements become progressively slower and more difficult to initiate; ultimately, patients are forced to recur to large head thrusts as means to initiate gaze shifts. We wondered whether, as a precursor of this condition, head movements would facilitate gaze shifts already in early stages of the disease. We studied horizontal head movements and eye-head coordination in 29 early stage HD patients (Ps) and 24 age matched controls (Cs). Subjects tracked random horizontal steps of visual or auditory targets while their heads were either stabilised (saccade amplitudes <or=40 degrees) or free to move (amplitudes <or=160 degrees). Subjects were to react either immediately (reactive mode), or wait until a go signal was sounded (delayed mode), or by antisaccades. Ps' head velocity was found to depend on the age of disease onset in a similar way as their saccadic eye velocity does, being clearly reduced in early affected Ps, but increasing to normal levels in lately affected Ps. Yet, saccade and head velocity were only loosely correlated although both exhibited a negative correlation with the severity of Ps' genetic condition (number of Ps' CAG repeats). Eye-head coordination turned out to be identical in Ps and Cs except for quantitative differences caused by the lower saccade and head velocities of Ps. Specifically, the timing between head and eyes and the head contribution to gaze shifts were similar in both groups. Moreover, preventing head movements did not affect the saccade latency or accuracy of Ps. Although Ps made more small involuntary head movements in this condition than Cs, these movements were not instrumental in generating saccades since they occurred only late after saccade onset. Thus, the head manoeuvres of severely affected patients must be considered a late adaptive behaviour. Finally, the ability of both Ps and Cs to suppress immediate reactions in the delayed and antisaccade conditions diminished as target distance decreased, with failure rates in Ps being much larger than in Cs. Unlike eye and head velocity, these failure rates were not correlated with age and, by the same token, neither with the variations in head and eye velocity nor with the number of CAG repeats. Hence, the pattern of brain areas prominently affected by HD is likely to vary significantly among individuals.