Neurons in primate visual cortical area MT show precise visuotopic organisation and are known to be highly sensitive to motion direction. It is generally thought that the large receptive fields of these neurons make them incapable of providing precise signals for spatial position. The spatial precision of neuronal populations depends, however, on both receptive field size and the extent of overlap in receptive fields. Here we directly compare spatial- and motion selectivity of area MT by measuring from populations of MT neurons. Recordings were made in area MT from marmosets (Callithrix jacchus) (n=3). The animals were anaesthetised throughout the duration of the experiment (induction: 12 mg/kg Alfaxan, i.m., maintenance: 6-12 µg/kg/hr sufentanil, i.v.) and artificially ventilated through a tracheostomy. Muscular paralysis was then used to prevent eye movements (0.3 mg/kg/hr pancuronium bromide, i.v.). Level of anaesthesia was established by continuous monitoring of EEG and ECG; anaesthesia was taken as dominance of low frequencies (< 5Hz) in EEG and absence of changes in EEG/ECG during application of noxious stimuli (paw pinch). A craniotomy was made and an electrode array (10×10) implanted in and around area MT in one hemisphere of the animal. The visual stimulus was a white disc moving along each of several possible linear trajectories at 20 °/s. Multi-unit activity at each electrode was counted in 0.05 s bins. Linear support vector machines (SVM) were used to establish whether neural activity in area MT was able to discriminate between pairs of spatial locations up to 8° apart, along trajectories that differed in motion direction by up to 36°. Discriminability depended on a combination of both factors. An elliptical surface fit to the data indicates that for moving objects an average 1.69° (±.04, 95%CI) difference in spatial position was as discriminable as a 30° change in motion direction. Tuning curve width of these neurons was on average 3.2° (±2.0, s.d.) in space and 45.7° (±12.1, s.d.) in motion direction. Receptive fields of neurons in area MT are smaller when close to the fovea, and the performance of the SVM paralleled this. Near the fovea, a separation of 1.48° (±.06, 95%CI) was required, and in the periphery a separation of 2.13° (±.27, 95%CI) was required, to match performance for a 30° change in motion direction. We conclude that populations of neurons in area MT provide signals for spatial position that are as precise as the signals they provide for motion direction. In addition our analyses show that while neurons show broad tuning for spatial position and motion direction, overlap in these dimensions allows precision in the population response. We speculate that the combination of broad tuning and high overlap is an efficient mechanism for simultaneous encoding of multiple dimensions of visual analysis.