Stereomotion experiments often feature monocularly visible stimuli. When such stimuli move in depth, different monocular motion signals are presented simultaneously to each eye. Though this 'inter-ocular velocity difference' cue to stereomotion has been investigated in a number of studies, the observer behaviour could theoretically have been based on properties of an individual monocular image. To isolate such potential artifacts, we performed a 2IFC speed- discrimination task using random dot stereograms moving either directly or obliquely (at an angle of ±0.25 deg, interocular speed ratio 2:1) away from the observer. Stimuli (7.3 × 1.3 deg; 600ms; standard disparity speed, 0.62deg/s) were displayed using ferro-electric shutter glasses on a high-speed fast-phosphor monitor (120Hz per eye). A background pattern of static random dots allowed us to avoid visibility issues as the DRDS passed through zero relative disparity, while monocular half-occlusion artifacts were minimised by using horizontally extended stimuli. Multiple interleaved staircases were used to measure speed-discrimination performance for conditions in which the trajectories were direct in both intervals, or in which the trajectory in one interval was direct and in the other oblique. The points of subjective equality in the 'direct' condition were, as expected, unbiased, while those in the 'mixed-trajectory' conditions werr biased on average by 9% and 15% (oblique faster) for the 2 observers (one naive) tested. However, discrimination thresholds were similar across all conditions (∼25% Weber fraction). Though these data are inconsistent with the monocular strategy of responding on the basis of the image motion experienced in a particular eye, they are consistent with two alternative hypotheses: i) oblique stereomotion appears faster than directly receding motion, or ii) the perceived speed of stereomotion is biased by the monocular image motion from the eye experiencing the faster motion.