In this paper, we quantify the effects of two physical affordances on user interaction in an augmented virtual environment: the grounding of datasets on a workbench and the augmentation of a Magic Lens interface tool. We investigated the effects of these affordances on subjects' performance and behavior in an information gathering task. Our results indicated that grounding had a significant main effect and that there was significant interaction between grounding and interface factors. Specifically, subjects tended to perform better with some level of affordance than none at all, although performance with grounded datasets decreased in the presence of augmented Magic Lens. Further, both affordance factors influenced behavior by reducing head mobility.