Well, I’m impressed! I attended one of Tim Flannery’s promotional talks for his latest book, The Weather Makers, and he’s a great speaker.
I’ll second The Future Eaters as a great book if you’re interested in the development of the Australasian environment (including New Zealand, New Caledonia and I’ve forgotten the other area). It covers many of the weirdosities about this region of the world, how they came about, and so on.
I’m sure any number of Aussie men who live at isolated sheep stations could answer that for you. 
The only problem with this theory is that Australia has only recently been mostly desert. Australia’s been slowly creeping north through the climatic zones. Go back a few million years and most of Australia is forest and grassland.
That’s a very simplistic if not inaccurate assessment.
The dingo does seem to have replaced the Thylacines. However it is still debatable whether it outcompeted the Thylancines. Given that a Thylacine and a Devil survived on Tasmania while both groups finally vanished around the time of the Dingo’s arrival in the mainland this suggests aloink. However since Devils continue to survive in Tasmania despite the presence of dogs for centuries this also suggests that disease or environmental change were far more important. Australia was home to numerous Thylacine species. Most of the m became extinct long before dingoes arrived, suggesting strongly that the remaining species was already under pressure well before the arrival of dogs, a point backed up to some extent by the fossil record.
So while the dog has occupied the niche vacated by the Thylacines it’s probably not the case that they literally replaced them.
Cats are also replacing the Dasyurids. However as with the dog/Thylacine replacement this is a movement of a novel predator into a vacated niche. Dasyurids are preyed upon heavily by foxes, which seems to be the main reason for their decline, along with hunting by humans.
Rabbits, rats and mice have not and are not replacing any native herbivores in Australia. Australia’s suite of native rodents are well established and under no threat from exotics. The only introduced rodents are the two cosmopolitan rat species and mice. And mice are the only species to make much inroads into rural areas. Rats remain primarily a pest of human habitat as they are in most of the world. Mice are fairly widespread, but only reach significant numbers in agricultural land. They pose no threat to native herbivores from competition, and at this late stage it’s doubtful they pose any disease risk.
Rabbits have been responsible for decimating populations of bilbies by competing with them for their nesting burrows. But bilbies aren’t primarily herbivores. There’s some suggestion that rabbits may also have had an impact on other burrowing species. But once again, this isn’t replacement. Those herbivore species that have vanished or declined dramatically in Australia have done so primarily because of ecological change induced by humans or introduced predators. Introduced herbivores have had relatively little impact through competition.
I think Lemur has mostly answered this: small landmasses have fewer species so random chance says that they will have fewer survivors. But there are a few important point is that need to be added.
Firstly, placentals don’t tend to drive marsupials extinct. In South America for example the Marsupial predators drove the placental predators into extinction, while the placental herbivores killed of the marsupial herbivores. So it’s just as possible for marsupials to eliminate placentals as the other way around. It’s only when a marsupial suite has to face all the species from the entire world that they are likely to lose out. But this comes back to relative landmass size. Random chance says that the species from the smaller landmasses will lose out no matter what they are. And we see this time and again. North American indigenous bear, bison, elk, deer and numerous other placental species became extinct when they faced contact with their Eurasian counterparts. Today all those groups of animals in North America are descendents of Eurasian species. The American species are gone. The same goes for the placentals in South America and Madagascar for example, along with numerous Australian placentals.
The truth of the matter is that when the entire world competes against an island the rest of the world usually winds, but that can be accounted for by pure chance.
The other important point to acknowledge is direction of flow. Literally hundreds of species of placental animals were deliberately introduced to Australia. That includes cats, dogs, monkeys, mongoose, antelope, cattle and so forth. And of course there are hundreds more accidental introductions. About 12 of them managed to survive in the face of competition form the indigenous species. That’s not exactly overwhelming evidence that placentals replace marsupials. In fact it suggests the opposite that even with major human intervention placentals just can’t compete. Even rabbits failed to establish in Australia initially, an it was only after much hard work and several decades that viable populations were established. In contrast no marsupial species at all have ever deliberately been introduced anywhere else on the planet. Yet there are at least 3 feral marsupial population in placental territory on different continents. Based on the direction of flow marsupials are actually far more likely to establish in the face of placental competition than vice versa. Of course really the data is too poor to draw either conclusion.
But my point is that the question is not “Why do placental animals tend to displace marsupial species”. The question is “Why are so many placental species given the opportunity to try to displace marsupials, and why are there so few opportunities for marsupials to move the other way”.
The third point, particularly as it pertains to Australia, is that the system was already a wreck from human activities. Massive ecological change, mass extinction of plant and animals major alterations of landscape had been caused by humans in the few thousand years before the latest wave of placentals arrived. These things weren’t entering into anything like a functioning pristine system to compete with marsupials on their home turf. Hey were entering a decimated and impoverished wreck. Needless to say they often managed to establish fairly well, specifically because they niche they were entering had been so recently vacated and the normal predators exterminated.
Placentals and marsupials were initially both present in Australia, The placentals died out at some later date. The marsupials came to dominate all ecological niches and displaced the placentals for whatever reason.
The problem with this idea is that marsupials invest les in pregancy, but far more in post-partum care. As a result if a marsupial loses a young it loses young during pregnancy it is less of a loss, but if it losses young significantly post-partum it is a far more significant loss.
The things is they weren’t more successful everywhere else. In South America marsupials dominated as carnivores. So any explanation is going to have to explain why marsupials co-dominated on one continent as well as why they dominated on a another. The simplest explanation seems to be simply pot luck. Some group had to be dominant everywhere. Random chance meant that more places were dominated by placentals which then gave them the edge, once more through chance, when isolated areas were rejoined.
[Moderator hat on]
I appreciate all the replies that are based on science. I DON’T appreciate all the crap replies that are turning this forum into MPSIMS. Quit it. NOW.
[/Moderator hat OFF]
Another question on this topic, what was the last common ancestor of placental and marsupial mammals? I find the degree of parallel evolution that occured after that absolutley amazing as we have animals of both types that look very similar and reside in similar ecological niches, from flying squirrels to dogs.
According to Wikipedia :
That means that the last common ancestor lived before the Cretaceous-Tertiary extinction, 65 million years ago.
Actually, there were several more groups than the four outlined by Wiki, most of which were small insectivores and fructivores. Only those four survived the K-T transition (implicitly in the case of Monotremes, which are only known back to the Miocene, but which clearly had a long separate evolution).
The issue of convergence is made even more interesting by Tertiary South American mammals, where marsupials, flightless birds, and, briefly, land-adapted crocodylomorphs fought it out for the large- and medium-sized carnivore slots, while edentates filled the rhino and semi-bipedal-leafeater econiches, and “local brand” ungulates unrelated to the ungulates elsewhere filled the camel, horse, hippo, elephant, and even hare niches.
I’ve just been reading Richard Dawkins’ The Ancestor’s Tale, which traces the evolution of the ancestors of humans. According to that the last common ancestor of placentals and marsupials lived about 140 million years ago, shortly* after Godwana broke up. He or she was approximately our 80-million-greats-grandparent.