Spatio-temporal convergence (STC) properties of 19 vestibular only neurons' responses to translational head movements were examined in an alert, behaving squirrel monkey. In addition to standard tests that included 1.2 Hz yaw/pitch/roll rotations and inter-aural/naso-occipital/dorso-ventral translations, we also observed responses to translations along multiple directions in several orthogonal planes. Neural responses were fitted first by a model that permitted STC in all planes, characterized by a non-zero minimum and a gradual shift of phase. We then evaluated statistically whether models with fewer independent dynamic parameters yielded equally satisfactory results. The responses of 13 neurons were adequately fit by simple cosine models (1-D) as well as models that allowed complex STC behavior. Of the six neurons exhibiting STC, five could be modeled with two independent phase parameters (2-D) while the remaining neuron required a model with three independent phase parameters (3-D). The maximum translation sensitivity and phase, Smax and varphimax, and minimum translation sensitivity and phase, Smin and varphimin, were estimated from the reconstructed sensitivity and phase surfaces. The tuning ratio, Smin/Smax, in STC neurons was>0.40 while in 1-D neurons it was <0.25. Furthermore, the maximum response vectors of most 1-D neurons lay within 20 degrees of either the horizontal or sagittal plane while those of STC neurons lay >20 degrees from both planes. No difference in other response properties, such as varphimax or rotational responses, was found between neurons exhibiting STC and the simple cosine tuning. Our results suggest that the STC behavior observed in otolith-related vestibular neurons probably arises from summing inputs from afferents, with diverse response dynamics, innervating different otolith macula.