Understanding the spatial ecology of animals is an important part of conserving biodiversity. In this context,manual radio-tracking is a primary means for understanding the movement of freshwater fish. Our aim was to determine the suitability of using linear or area-based estimates to quantify diel use of space by Macquarie perch, Macquaria australasica (Percichthyidae), in a reservoir. Area-based estimates were calculated as minimum convex polygons (MCPs). Four-hourly radio-tracking of 23 adults within diel periods resulted in 136 pairs of linear and area-based estimates. We encountered problems with MCPs that were not associated with outlying radio-fixes. MCPs were more likely to overlap with land as the extent of an individual's range increased. Corrections were applied to account for land intersecting with MCPs (33.1% of cases); however, underestimates of diel range occurred in most of these cases. MCPs were also small despite substantial movement along one dimension if individuals occupied only straight-edged sections of the reservoir and did not traverse bays or peninsulas (8.1% of estimates). Mean (+/-s.e.) diel-range length was 516+/-89m (n=23) and records varied from 6 to 1972 m. In waterways with convoluted shorelines, linear estimates of diel range are more reliable than area-based estimates derived from manual radio-tracking.