The purpose of this dissertation was to describe and test a framework (the coordinate relations hypothesis) for understanding the types of shape representations and recognition tasks that are mediated in dissociable neural subsystems. The results from Experiment 1 demonstrate that there is a right hemisphere advantage for recognizing upright animals, but that inverting animals eliminates the right hemisphere advantage for basic-level animal recognition. The results of Experiment 2 showed that there is a right hemisphere advantage for physically discriminating objects that differ only in their metric properties but that there is not a right hemisphere advantage for objects that can be discriminated using a structural description. Experiment 3 tested whether showing the metric change objects at different orientations eliminates the right hemisphere advantage for the metric change trials. Experiment 4 was designed to investigate the effects of metric vs. geon changes in priming. Unfortunately, the results of Experiments 3 and 4 were inconclusive (possibly due to a lack of statistical power). Although the results of Experiments 3 and 4 were inconclusive, Experiments 1 and 2 provide strong evidence for the coordinate relations hypothesis and are inconsistent with other current hypotheses regarding the neural subsystems for visual recognition. The results thus suggest that one subsystem mediates structural description representations and subserves any recognition tasks that can be performed using a structural description and the other subsystem mediates coordinate relations representations, and mediates recognition tasks that require a coordinate relations representation.