We investigated how the visibility of targets influenced the type of point used to provide directions. In Study 1 we asked 605 passersby in three localities for directions to well-known local landmarks. When that landmark was in plain view behind the requester, most respondents pointed with their index fingers, and few respondents pointed more than once. In contrast, when the landmark was not in view, respondents pointed initially with their index fingers, but often elaborated with a whole-hand point. In Study 2, we covertly filmed the responses from 157 passersby we approached for directions, capturing both verbal and gestural responses. As in Study 1, few respondents produced more than one gesture when the target was in plain view and initial points were most likely to be index finger points. Thus, in a Western geographical context in which pointing with the index finger is the dominant form of pointing, a slight change in circumstances elicited a preference for pointing with the whole hand when it was the second or third manual gesture in a sequence.