A sense of direction (SOD) computes the body's facing direction relative to a reference frame grounded in the environment. The authors report on three experiments in which they used a heading-recall task to tap the functioning of a SOD system and then correlated task performance with self-reported SOD as a convergent test of the task's construct validity. On each heading-recall trial, the participant judged the photographer's allocentric heading when photographing a pictured outdoor scene. Participants were tested over the full range of SOD ratings in Experiment 1, and in Experiments 2 and 3 heading-recall at the SOD extremes was tested. In all experiments, there was wide variability in heading-recall accuracy that covaried with self-rated SOD. Parametric manipulation of various task parameters revealed some likely functional properties of the SOD system. The results support the psychological reality of a SOD system and further indicate that there are large individual differences in the efficacy with which the system functions.