Mass extinctions are, by definition, harsh, but they each seem to be disastrous in their own unique way. After all, the KT extinction was likely caused by an asteroid, but other mass extinctions may have involved glaciation, global warming, volcanic activity, sea level changes, and changes in oceanic or atmospheric oxygen levels, among other factors. Based on these divergent triggers, it might seem that each mass extinction should affect diversity differently, changing the rules of extinction in different ways. David’s hypothesis — that during mass extinctions having a broad geographic range protects clades from extinction — was inspired and supported by North American mollusks in the KT mass extinction, but how would the idea hold up against data from other organisms in other times and other places? In other words, how general is this “rule” of mass extinction? In the 20 years since David first formalized the idea, many other scientists have checked it in different time periods and against their own study organisms — and, perhaps surprisingly, have generally found it to be a good fit. For all five of the Earth’s major mass extinctions (whether caused by an asteroid impact or not), the rule seems to hold. David suspects that this wide-range rule may actually operate during normal times too, but that during these times, so many other factors enhancing survival are also in play that geographic range only comes to center stage during mass extinctions, when the benefits of these other features melt away. This will be an exciting area of research in the future, he says.

The generality of this rule paints an illuminating picture of how mass extinctions actually worked. Why should the rules of survival change during mass extinctions? Why might traits that are advantageous during normal times be useless during mass extinctions, and why should having a broad geographic distribution be so important? David reasons that mass extinctions must involve an almost total collapse of normal biological cycles, nullifying the advantage of traits adapted for those normal times. Furthermore, he hypothesizes that in all the chaos of a mass extinction, there must be a few, small, randomly scattered refuges offering protection from the destruction. Surviving, then, would be largely dependent on having some of your group members living in one of these refuges — a lucky break more likely to be granted to genera with broad geographic ranges than those with narrow distributions. David describes this situation as something of a global holocaust with tiny oases scattered randomly around the world: “In a mass extinction there must be a fundamentally global disturbance, but there must be little patches around the world that are refuges, where the severe effects are damped in some way. And so it must be that somehow these widely distributed genera are more likely to encounter the refuges than are those that live in narrowly restricted areas . . . The interesting thing is that it’s not that there was just one place on Earth, or a few places, that were reliably safe on a broad scale. It’s not as if Spain was a refuge or Florida was a refuge — because we’ve actually tested for the existence of these large-scale refuges [and haven’t found them] . . . Instead, it must have been on a very fine scale — a bay here or an island there.” David has been poring over geographic patterns of extinction to test the idea of small scale refuges: “Someday it would be wonderful to find an example of one of those refuges, where there’s exceptionally low extinction, but no one has found one yet. And we’ve really looked hard.”