The role of perceptual load in selective attention to multidimensional objects was investigated by independently manipulating the load along the task-relevant and the task-irrelevant dimensions in the central search array, which was flanked by congruent, incongruent, or neutral peripheral distractors. The relative bottom-up perceptual saliency of these dimensions in capturing attention was manipulated between experiments. When the task-relevant dimension was the color of the letter and the task-irrelevant dimension was the visual shape of the letter (Experiment 1), manipulation of the letter shape perceptual load had no impact upon the pattern of congruency effects in responding to the color, i.e., smaller congruency effects under higher color perceptual loads and larger congruency effects under lower color perceptual loads. When the task-relevant dimension was the shape of the letter and the task-irrelevant dimension was the color of the letter (Experiment 2), there were no congruency effects in responding to the letter shape under high color perceptual loads irrespective of the letter shape loads. When only the target and the flanker were colored whereas the distractors in the central array were not (Experiment 3), the task-irrelevant color information reduced or eliminated the impact of letter shape perceptual load on the congruency effects in responding to the letter shape. These findings suggested that selective attention to multidimensional objects follows the general principles suggested by the perceptual load theory, but the bottom-up perceptual saliency plays a primary role in the distribution of attentional resources over objects and dimensions.