The discharge characteristics of postsubicular head-direction cells in a fixed environment were described in the previous paper (Taube et al., 1990). This paper reports changes in the firing properties of head- direction cells following changes in the animal's environment. Head- direction cells were recorded from rats as they moved freely in a 76-cm- diameter gray cylinder. A white card, occupying 100 degrees of arc, was taped to the inside wall of the cylinder and served as the major orienting spatial cue in the animal's environment. Rotation of the cue card produced near-equal rotation in the preferred firing direction of head-direction cells, with minimal changes in peak firing rate, directional firing range, or asymmetry of the firing-rate/head- direction function. Card removal had no effect on peak firing rate or range of firing, but in 8/13 cells the preferred direction rotated by at least 24 degrees. Similarly, changing the shape of the environment to a rectangular or square enclosure caused the preferred firing direction to rotate by at least 48 degrees for 8/10 cells in the rectangle and 3/8 cells in the square, with minimal changes in the peak firing rate or directional firing range. Hand holding the animals and moving them around the cylinder had no effect on the preferred direction or firing range of the cell, but decreased the maximal firing rate in 7/9 cells. On 2 occasions, 2 head-direction cells were recorded simultaneously. The rotation of the preferred firing direction for one cell was the same as the rotation of the preferred direction for the second cell after each environmental manipulation. These results demonstrate that specific visual cues in the environment can exert control over the preferred firing direction and indicate that head- direction cell firing is not a simple sensory response to visual cues, but rather represents more abstract information concerning the animal's spatial relationship with its environment. The constancy of the angle between the preferred firing directions of pairs of simultaneously recorded head-direction cells suggests that there is a fixed mapping of the population onto direction within the environment. Thus, environmental manipulations appear to cause only a change in the reference direction, but leave all other discharge characteristics of directional cells unchanged. In the discussion, comparisons are drawn between the responses of head-direction cells and hippocampal place cells to similar environmental manipulations (Muller and Kubie, 1987), and ways in which these 2 spatial systems interact in navigation are discussed.(ABSTRACT TRUNCATED AT 400 WORDS)