Myers et al.1, in their new analysis of global biodiversity hotspots, recommend areas where conservation actions should be focused to minimize losses in the imminent extinction crisis. We strongly support initiatives to produce clear, efficient and practical goals for conservation to guide biodiversity planners and decision-makers in governments, agencies, conventions and non-governmental organizations (NGOs). However, as things stand there is only limited consensus on global conservation priorities at international level. We believe that the time is now right for scientists and practitioners to work together to develop a commonly adopted blueprint for action.

A key first step would be a structured debate to identify common goals; to pool data sets; and to agree on the contributions that non-equivalent measures of priority, such as diversity, endemism, threat, viability and ecological function should make. All this should lead to a powerful and cost-effective product. Most important, however, is that the sound analyses, breadth of expertise, and strength of consensus underpinning such an agenda would greatly increase the probability that decision-makers would adopt the resulting priority system, and take action to stem the loss of wild species and maintain critical ecosystem processes.

Currently there is much duplication of effort across organizations. Prioritization programmes with comparable goals to that of Myers and his collaborators at Conservation International are in progress at WWF-US2, BirdLife International3, IUCN4, World Resources Institute5 and The Nature Conservancy6. All these are enormously costly, involving independent programmes to gather data, undertake analyses, and publish and advertise products. The result is considerable redundancy, and generation of competing rather than complementary priority sets (sets of areas that result from systematic prioritization). We appreciate that organizations have to develop specific products to meet their own objectives. Nevertheless, we believe that leading NGOs must cooperate in data-gathering, analysis and priority- setting if there is to be a widely supported strategy that the entire conservation community can work together to implement.

In our view, the analysis of Myers et al.1 highlights several other issues that must be addressed in reaching consensus. First, a co-ordinated programme must take account of the many recent advances in systematic priority-setting7. The analysis by Myers et al . does not include new techniques allowing the identification of priority sets which, by paying explicit attention to patterns of between-site complementarity, greatly boost the efficient representation of all mapped species and not just endemics targeted in pre-selected sites7,8,9,10. These techniques also allow planners to justify the choice of particular areas, to determine the flexibility of site selection and to balance representation and conservation efficiency against a suite of socio-economic costs (such as land price, degree of land transformation, and human population density)7,9,10,11.

Second, there remain several serious limitations to all current assessments of global priority. For instance, most conservation decisions take place at a finer geographical scale than that discussed by Myers et al.1. Hence new techniques, which allow analyses to move freely between different scales, are needed before global analyses can be translated into effective action on the ground. At these finer scales there are already some encouraging collaborations among academic researchers and NGOs (see the letter by Fonseca et al., below).

Equally pressing, priorities identified for one taxon may fail to reflect diversity in other groups12 — hence analyses should be developed from independent data sets from several taxa, ideally including invertebrates: the taxon that represents most terrestrial biodiversity. Biological priorities must also be integrated with socio-economic considerations, but current developments in this area are still at a rudimentary stage.

Last, conservation practitioners and academic researchers alike are concerned that systematic prioritization focuses on patterns of species and community distribution, yet largely fails to address the conservation of key ecological and evolutionary processes which maintain those patterns13. There is growing evidence that conserving the pattern will not by itself guarantee the conservation of these processes. Yet we currently lack robust measures for quantifying the extent to which different areas contribute to core processes, or for evaluating the overall performance of priority sets in terms of process maintenance. In consequence, although it is claimed that the hotspots approach is cost-effective1, problems concerning selection of areas, scale of analysis, socio-economic concerns and maintenance of key processes mean that areas of high conservation value may have been missed.