Using the classic mental rotation task paradigm, I compared participants' performance with Shepard and Metzler (1971) cube figure stimuli and tetrahedron molecule-like stimuli. In Eperiment 1, I used the classic mental rotation task paradigm to compare performance with the original cube figure stimuli and molecule-like tetrahedron stimuli as well as axis of rotation – picture- vs. depth-plane. In Experiment 2, I tested the effect of display dimensionality – stereoscopic 3D versus standard mono, and also included additional trials to investigate practice effects. In Experiment 3, I looked at the moderating effect of ability on performance and strategy use. In all experiments, response time and error rate as well as slopes and intercepts were calculated to examine overall performance and individual differences. Specifically for trials with rotations about the horizontal depth-plane axis, the left and right elements did not move laterally across angular disparities. This allowed for a non-rotation strategy that involved comparing the relative locations of the adjacent elements, rather than utilizing a mental rotation strategy on rotations about the horizontal depth-plane axis. Individuals were assigned to strategy use classes based on both self-report and statistical clustering of individual slope coefficients (i.e. rotation rates). The moderating effect of ability on performance and strategy use was assessed three measures of ability: Cube Comparison, Paper Folding, and Abstract Reasoning. Extending Shepard's classic mental rotation task paradigm, I utilized desktop virtual reality software for stimuli presentation and data collection. On depth-plane rotations, the angular disparity effect (ADE) was significantly greater on cube figures compared to tetrahedrons, while similar slopes were observed on both stimuli types for picture plane rotation trials. I found general agreement between both assessments of strategy use–- self-report and clustering on individual slope coefficients. Two divergent strategy classes emerged after repeated practice with molecule-like tetrahedron stimuli: a) a mental rotation strategy evidenced by a positive angular disparity effect, and b) a non-rotation strategy with slopes too small to consider mental rotation. Participants picked up on the non-rotation strategy about 70% of the time and discriminant analysis showed the classifier had a predictive error of 30% against self-reported strategy as labeled data. In general, rotators scored higher on Cube Comparison and the two strategy groups did not differ on Paper Folding or Abstract Reasoning. In terms of performance, higher Cube Comparison scores were predictive of larger slopes for tetrahedron stimuli, indicating use of a mental rotation approach for both stimuli types. High abstract reasoning scores predicted larger slopes for cube figure stimuli and flatter slopes for tetrahedron stimuli - indicating successful strategy switching. This indicated that higher performing participants were able to flexibly switch between an orientation-independent strategy for tetrahedron stimuli and a mental rotation for the cube figure stimuli. These findings improve our understanding of the interplay between ability and strategy for reasoning about spatial tasks with domain-specific stimuli. Importantly, I provide evidence that the molecule-like tetrahedron stimuli were not necessarily processed the same way as stimuli used by Shepard and Metzler (1971). Cube Comparison Test score predicted use of a mental rotation strategy and Abstract Reasoning predicted use of an alternate non-rotation strategy for the tetrahedron stimuli. Individual differences in ability are discussed in the context of spatial judgement tasks for organic chemistry instruction.