A geologist questions a grand theory.

Atmospheric oxygen concentrations are falling. Breathing is difficult. Those that can't cope are collapsing and dying with symptoms akin to altitude sickness.

This may read like the first page of a Hollywood script, but, according to the oxygen-stress hypothesis, a similar scene occurred 251 million years ago at the end-Permian mass extinction, when up to 95% of all animal species died out. Like all good prevailing hypotheses, this one makes predictions that can be tested, if only the right rocks can be found.

Enter Tyler Beatty of the University of Calgary in Alberta, Canada, and his colleagues. They recently set up camp in the remote reaches of northwestern Canada, where rocks spanning the end-Permian extinction show a shift from Permian sandy carbonates to Triassic sand and mud. They found that fossils of entire creatures are not common at the boundary, preventing taxonomic analyses, but that fossils documenting sediment disturbance by animals are (T. W. Beatty et al. Geology 36, 771–774; 2008). This is fortuitous because such disturbance in marine sediments is linked to oxygen concentration. So these rocks may preserve a 'smoking gun' for an oxygen-stressed world.

However, the shallow marine sediments of the Early Triassic were pervasively burrowed by diverse organisms of the period, including large, oxygen-demanding arthropods. Only deeper-water sediments, deposited below wave-mixed surface waters, had the expected oxygen-stressed fossil traces.

This complicates the oxygen-stress story for the end-Permian mass extinction. Beatty et al. stop short of asking whether the end-Permian mass extinction was really caused by a massive reduction in atmospheric oxygen. But in light of their results, I am not holding my breath.

Discuss this paper at http://blogs.nature.com/nature/journalclub