If attention is drawn on either side of a bar prior to its presentation (by showing a spot) the bar will appear to move away from where the attention was drawn presumably due to some form of local facilitation. We call this effect motion induction (MI) and have shown, using this paradigm, that split-attention can be easily demonstrated but only under conditions of attribute similarity (ARVO, 92). On the other hand, the basic MI effect with a single attentional focus is obtained using any combination of attributes between the attention getting spot and the bar. In the condition of split-attention, two spots are presented on the screen followed by the bar which then appears to be drawn from both sides and to collide in the center. This effect can be tracked mentally. Further, we demonstrate that the potentiation of attentional focus is generally multidirectional but can be biased by the geometry of the stimuli used. For instance, if two spots are presented and then two bars are presented, one between the two spots and the other perpendicular to the bar over the left spot, a central collision will not be seen but movement to the left and up will be observed. We beleive that this type of split-attention influence on MI can be represented by a minimal resistance model of motion induction which is both directional and attribute based.