By Richard Black
Environment Correspondent, BBC News website, Darwin
The woolly mammoth and the dodo have gone; the dinosaur kingdom lies withered in its fossil graveyard.
Eidothea hardeniana is very active and genetically diverse
The gorilla and the bonobo dwindle, along with countless fish and rainforest beetles as yet un-numbered.
Golden eagles and rhinos, meanwhile, gambol in their newly found multitudes.
What is it, then, that decides who lives and who dies - which species teeter and fall after the dodo, and which, like the eagle, regain their numbers and soar again?
It is a key question for conservationists, desperate to deploy scarce resources to best effect as the human race's expansion funnels so many plants and animals towards the vortex of extinction.
So you would think there would be a definitive answer, and you might think it would involve the size of a population; the fewer there are, the worse the prospects.
Think so? Then read on...
Intuitively, it might seem there is a level below which numbers must not fall if any chance of recovery is to remain.
This number has been dubbed the "minimal viable population" (MVP), and it crops up regularly in the conservation literature.
One commonly-used definition - there are others - is to ask "what is the population size which gives the species a 90% probability of surviving for 100 years?"
Common sense would tell you that the MVP ought to be different for each species, depending on factors such as how fast it reaches maturity, how frequently it reproduces, and so on.
The MVP ought to follow logical rules; so Barry Brook and Corey Bradshaw set out to find what they might be.
In research submitted to the scientific journal Ecology Letters but not yet published, they calculated MVPs for 1,198 species using standard computer models of population.
These include mammals, insects and plants, and were chosen because there is good data about how populations have changed over time - enabling a good estimate of MVP.
The scientists, both based Charles Darwin University in the city of Darwin, the capital of Australia's Northern Territory, then asked whether MVPs could be predicted from information about the particular species.
"We looked at things like range size, whether the species exists around the world or just locally," Corey Bradshaw told me.
"We looked at things like the level of human impacts, fertility rates, and so on, and we did this massive analysis for all these species.
"At the end of the day, we had an extraordinarily low predictive power of the minimal viable population size; we were unable to really predict what the MVP should be based on all these correlates."
The things which according to human reason should matter, then, cannot tell you how big a population needs to be in order to remain one of nature's rhinos rather than a mammoth.
Not much for conservationists to hang onto there, then, other than the simple conclusion that population size is not important; and that conclusion is ringing bells across the other side of Australia, at the Royal Botanic Gardens in Sydney.
Here, Maurizio Rossetto is seeking the rules of survival using a different approach; whereas Bradshaw and Brook went for a broad look across the smorgasbord of life, he is surveying a few species in intimate detail in rainforest fragments.
"There are a lot of species that have been there for a very, very long time - lineages that have been there for 100 million years," he told me.
"So if we can start to understand the patterns - what has happened in the past as far as contraction and expansion, why some species are more and less widespread, are there ecological traits that are affecting this - once we have this information, we can start to develop much better predictive tools."
Two trees tale
One chapter in Dr Rossetto's unfolding story concerns two rainforest trees which are both now limited to tiny pockets in northern New South Wales.
Eidothea hardeniana was discovered only five years ago; and only one population is known, numbering about 90 individuals.
Yet it appears to be a highly active plant, flowering young and growing vigorously; and the single population is genetically very diverse, a trait which is usually favourable to survival.
Maurizio Rosetto is trying to figure out the rules that determine survival
One mechanism which enables it to maintain this diversity is a block, presumably genetic, on self-pollination; only cross-pollinated seeds survive.
Each tree blooms infrequently, so no given pair of trees is likely to cross often.
And it is able to live for hundreds, perhaps thousands of years by re-sprouting - dying back, and allowing a new stem to form.
The second tree, Elaeocarpus williamsianus, possesses some of the same mechanisms which make Eidothea hardeniana so locally successful.
Nine small populations have been found in New South Wales; so you might presume it is in a better state than hardeniana.
Not a bit of it; here re-sprouting appears to have got out of hand.
"When we went to look at those populations and analyse them using DNA techniques, what we found was that every single population is one individual," recounts Dr Rossetto.
"Because, like Eidothea, it doesn't self-pollinate, all those populations - except for one, where there are two individuals - are sterile. So because of the success of this re-sprouting habit, this species has cornered itself in an evolutionary impasse."
So it is heading for extinction?
"It is not in a very practical position, let's put it this way!"
Effects if size
Two trees, then, each able to re-sprout and to prevent self-pollination; but it is williamsianus with its apparently larger population size which is in trouble, while the tiny colony of hardeniana looks set to cope much better.
"Sometimes there is this rationale - like 'we'll save the big area and get rid of the small area' - it's not as simple as that, and we have lots of examples where the small population has more diversity than the big population," says Maurizio Rossetto.
"You need to have better understanding; and if you have sufficient data from a sufficient number of species, that will help us to make some true generalisations."
Taken far enough, could detailed investigations including fieldwork and genetic analysis produce a kind of rulebook - a definitive list of the traits which make a given species more or less likely to survive a period of sustained stress, as the human race and its six billion members are now creating?
"I think it's not impossible," Rossetto says.
Until that happens, what conclusions would Corey Bradshaw draw from his overview? Are there any traits which place species in greater or less danger?
"If you think about it, nothing really evolves to go extinct," he comments, "it's anti-life, anti-Darwinian, it makes no sense.
"What does make sense is if, suddenly, a novel mortality source comes along, acting on a scale which is much less than the lifetime of an individual; then you're getting into a problem."
And here may be a clue; because in the animal kingdom, the lifetime of an individual tends to parallel body size.
"If your body size is large, chances are you're not going to be able to adapt to novel mortality sources.
"If I were going to predict whether an elephant or a mouse would go extinct, I would say an elephant most likely, because it's not going to be able to adapt to the changes we're making to the planet."
What does all of this prove? Basically, that extinction is a complex business - size, at least in terms of a population, is far from everything.
And allocating valuable conservation resources is best done from a position of knowledge.