1. In these experiments we examined the receptive field mechanisms that support the optic flow field selective responses of neurons in the dorsomedial region of the medial superior temporal area (MSTd). Our experiments tested the predictions of two hypotheses of optic flow field selectivity. The direction mosaic hypothesis states that these receptive fields contain a set of planar direction-selective subfields that match the local directions of motion within optic flow fields. The vector field hypothesis states that these receptive fields are uniquely sensitive to distributed properties of planar, circular, or radial optic flow fields. 2. Experiments using large-field stimuli revealed that some neurons showed changes in optic flow field selectivity depending on the position of the stimulus in the receptive field; these are position-dependent responses. However, other neurons maintained the same optic flow field selectivities in spite of changes in stimulus position; these are position-invariant responses. We have used the position dependence or invariance of optic flow field selectivity as a way of testing the direction mosaic and vector field hypotheses. Position dependence is more consistent with the direction mosaic hypothesis, whereas position invariance is more consistent with the vector field hypothesis. 3. To test for position effects, we examined the optic flow field selectivity of small subfields within the large receptive fields of 160 MSTd neurons. First, we centered small-field optic flow stimuli of various sizes over the same position in the receptive field. Most MSTd neurons showed decreasing response amplitude with decreasing stimulus size but maintained optic flow field selectivity. 4. We then placed small-field stimuli at various positions within the large receptive field of these MSTd neurons. Position-invariant response selectivity was most prominent in single-component neurons, suggesting that they were more consistent with the vector field hypothesis. Position-dependent response selectivity was most prominent in triple-component neurons, suggesting that they were more consistent with the direction mosaic hypothesis. However, the variations in planar direction preference throughout the receptive field of these triple-component neurons were not consistent with a direction mosaic explanation of the large-field circular or radial selectivity observed. 5. Small-field position studies also demonstrated the existence of zones within the receptive field in which either direction-selective inhibitory or direction-selective excitatory responses predominated. The degree of overlap between these zones increased from nonselective to triple- to double- and finally to single-component neurons.(ABSTRACT TRUNCATED AT 400 WORDS)