In an attempt to examine three main hypotheses on the evolution of leks, data on female home ranges, distance between leks and male site fidelity in the great snipe (Gallinago media) were obtained. In a 30 x 40 km area in central Sweden, the locations of 12 leks were identified. Six were within the study area (7 x 11 km) in which probably all leks were known. In the study area, nearest neighbor distances between leks were longer than the predicted drawing area (diameter of female home ranges plus the detection range of leks). Furthermore, in only I of 13 radio-tracked females did the estimated home range enclose many leks. These data do not fit the prediction from the hotspot hypothesis on the dispersion of leks, namely, that the distance between leks should be less than the drawing area and hence an average female home range should enclose more than one lek. The observed distance between leks was about the same as predicted by the female preference hypothesis. This hypothesis also predicts females visit mainly one lek; however, we found they sometimes visit two and nest close to a third. We propose that males may first settle according to hotspot rules, but females will resettle according to preferences for certain males and/or larger leks. In this way some hotspot leks are abandoned, and the distance between leks is increased as males become more tightly clumped. However, when all males and females have settled, the leks still existing would be on hotspots. In the female preference hypothesis, males are assumed to aggregate on leks because females prefer clustered males as they can be more easily compared than if they are dispersed. AlterOffprint requests to: J. H6glund