The experiments reported here address the issue of whether the pathways which extract motion from first-order and second-order spatial patterns remain separate or whether they combine at some higher level in the motion system to form a single pathway. The question is addressed by investigating the interaction of first-order and second-order stimuli in the processing of a global-motion stimulus [a variant of the task introduced by Newsome & Pare (Journal of Neuroscience, 8, 2201-2211, (1988)]. Two experimental procedures were used. The first consisted of determining the effect of the addition of dots of one type (e.g. first order) undergoing purely random motion on the ability to extract the global-motion signal carried by dots of the other type (e.g. second order). The second experimental procedure consisted of determining the effect of maintaining a coherent-motion signal in one type of dot, moving in the opposite direction to the global-motion direction, on the ability to extract the global-motion signal carried by dots of the other type. The dots were matched for their effectiveness in producing a global motion percept and the results for both procedures were the same. First-order dots impaired the ability to extract second-order global-motion, and second-order dots had no effect on first-order global-motion extraction. It is argued that the sensitivity of the second-order global-motion system to the first-order dots is due to the ability of the second-order local-motion detectors to detect these dots. The present results are thus interpreted as indicating that the first-order and second-order motion pathways remain separate up to and including the level in the motion system at which global-motion signals are extracted.