Lists of endangered species designed to protect flora and fauna may themselves be condemning a substantial proportion of plants and animals to extinction.
Australian botanist Mark Burgman, of Melbourne University, believes that the biases of the scientists creating the lists have serious consequences for conservation policy.
"Policy-makers, politicians and the general population trust us to be balanced in our provision of advice. Of course we are not," he said.
Professor Burgman began investigating the lists in a review of threatened plants in the Australian Journal of Biology .
He realised that most of the information on the turnover of species was "noise" due to changes in knowledge rather than the impact of human activity.
He calculated that revised lists varied between 10 and 20 per cent and that between 1993 and 2000 the number of endangered vascular plants (those able to transport liquids around) in Australia doubled.
Most of the new ones were species that had not been assessed before. There were no documented extinctions for lower species such as fungi or algae. Similarly, endangered animal lists tend to feature more bird and mammals than invertebrates.
Professor Burg-man believes the lists have value if their limitations and biases are understood.
He is working with the National Centre for Ecological Analysis and Synthesis at the University of California to develop new methods for evaluating species' conservation status.
Scientists and governments need to know the changes in numbers of extinct and threatened species, the causes of past extinctions and the threats in future. "Current lists simply do not provide this information," Professor Burgman said.
Scientific bias towards the cute, unique or spectacular has long been recognised. In 1995, Lord May, of Oxford University's zoology department, described the international lists as popularity polls.