How should bird conservation efforts be prioritised? This article reviews two approaches—the Collar et al. (1994) global study of threatened bird species and Garnett's (1992b) analysis for Australia. Each of these studies uses IUCN ‘extinction risk threshold’ criteria to derive conservation ‘hotspots’ where effort should be concentrated. Difficulties with IUCN rankings stem from two implicit assumptions: (i) that the conservation objective is always to maximise the number of surviving taxa and (ii) that conservation in different ‘hotspots’ is equally costly. Garnett attempts to circumvent these limiting assumptions by supposing lexicographic conservation preferences. This is partly successful but still leads to prioritisation anomalies.