Images: Monash Science Centre, fossil; DRAGA GELT, reconstruction

of a diminutive shrew-like animal a mere 8 centimeters long (right; jawbone area in white) is shaking the established family tree of mammalian evolution. This possibly placental creature, named Ausktribosphenos nyktos, apparently scurried out of the way of dinosaurs in what is now Australia during the Cretaceous--some 110 million years before the accepted appearance of such mammals there.

In the 23 years that paleontologists Thomas H. Rich of the Museum of Victoria and Patricia Vickers-Rich of Monash University have headed an effort to find the origins of Australia's unique mammals and birds, they have learned to expect the unexpected. But when excited field workers at a dig near Melbourne showed Tom Rich a fossil of a tiny mammalian jawbone on March 8, 1997, it was his jaw that dropped. As the implications of the discovery sank in, Rich looked up from the microscope with a smile breaking across his face. All he could say was, "My God."


CAUTIOUS ELATION. Tom Rich, lead author of the paper, looks up from his microscope after examining the A. nyktos jaw for the first time.

Rich's astonishment confirmed what the digging team had suspected since the fossil was discovered earlier that day: this bone simply should not be in Australia. The jawbone, barely 16 millimeters long, did not bear the characteristics of an egg-laying monotreme, such as the platypus; nor did it resemble the jaws of pouched marsupials, such as kangaroos and wombats. Instead, it seemed to be the jaw of a placental mammal--the group that includes horses, cats, bats, whales and people.

The trouble is that the fossil was wrested from Cretaceous rock that was 115 million years old while well-established theory holds that the first placental mammals in Australia were island-hopping rodents that arrived a mere 5 million years ago. "The case that terrestrial placentals had not reached Australia until 5 million years ago was so strong," Rich recalls, "that the thought of them being in Australia this early never crossed my mind."

The researchers' conclusion that terrestrial placental mammals may have lived down under 110 million years earlier than expected, as reported in the November 21, 1997 issue of Science, could all but uproot the mammalian family tree. Based on the fossil record, it is generally assumed that mammals arose some 200 million years ago, when the continents were still joined together in a single landmass called Pangaea. By the Cretaceous period (146 to 65 million years ago), the continental configuration had changed, and two supercontinents arose: Laurasia in the north and Gondwana in the south.

Current theory holds that the southern hemisphere saw the emergence of the egg-laying monotremes, which then crossed into South America, where they eventually died out. Marsupials, whose young continue their growth in the mother's pouch, seem to have originated in North America. And placentals, whose young form entirely inside the mother, arose in Asia.


But the fossil unearthed at the Flat Rocks site near Inverloch apparently lived in Gondwana--unexpectedly far south of the Asian placentals. Rich suspects that the animal would have been only about 8.5 centimeters long and a "generalized insectivore." He cites the spineless hedgehog as the closest modern analog to this ancient creature. It lived in a temperate valley among the polar dinosaurs." Because these warm-blooded creatures had to endure the darkness of winter in the polar regions, Rich's group gave it the official name of Ausktribosphenos nyktos, the "Australian Cretaceous tribosphenic mammal that lived by night."

Central to Rich's placental interpretation are A. nyktos's tribosphenic molars. All marsupials and placentals (termed therian mammals collectively) are characterized by this tooth type, in which hollows and cusps of corresponding upper and lower molars occlude in a mortar-and-pestle fashion to macerate food. According to Rich, details of the talonid basin, a lower molar hollow, in A. nyktos speak against a monotreme interpretation. And considering the absence of certain other features, Rich excludes the possibility that this creature was a marsupial. Rich also points to A. nyktos's inferred dental formula--the number of premolars and molars--as evidence of its placental nature: Placental mammals and, according to Rich's figuring, A. nyktos have five premolars and three molars, described as the 5/3 pattern.

Though intriguing, this mouthful of evidence has not convinced some scholars. "The talonid basin of A. nyktos doesn't look like that of any self-respecting placental mammal," declares San Diego State University paleontologist David Archibald. He points to a series of small cusps found on the tongue side of A. nyktos's talonid that are not seen in placental talonids. Any special similarities between this Australian beast and placental mammals, says Archibald, are probably the result of convergent evolution. Nor does Archibald think that the dental formula is conclusive. He notes that a 5 /3 pattern has also been observed for some monotremes, and suspects that this pattern may be a primitive trait for all therian mammals.


SCORE! Nicola Barton, the team member who spotted the fossil in the shattered rocks, clutches the fragment containing the jaw moments after the magnitude of the find was obvious.

Both Archibald and Richard L. Cifelli of the University of Oklahoma believe that A. nyktos looks as though it is related to an extinct monotreme called Steropodon (a fossil of which is known from the Early Cretaceous of Australia) or the peramurans, the group that gave rise to marsupials and placentals. Cifelli centers his arguments on A. nyktos's "really bizarre" premolar. "This style of molarization," he asserts, is "strongly suggestive of peramurans." The notion of A. nyktos as an intermediate step between monotremes and peramurans would lend support to the idea that monotremes share a much more recent common ancestor with marsupials and placentals than previously thought.

A more conservative opinion comes from University of Alberta paleontologist Richard C. Fox. Like Rich, he thinks that A. nyktos is a primitive tribosphenic mammal, but disagrees that it can be categorized more specifically than that. "It has some placental-like characters, but not the whole suite of characters that we would use to determine that it was a placental definitely."

So far, Rich's ideas have won tentative support from David W. Krause of the State University of New York at Stony Brook. "In terms of what we know about the evolution of early mammals, Rich's is the most parsimonious interpretation," he says. Because so little is known about Gondwanan mammals, Krause is wary of dismissing Rich's interpretation "just because we don't expect, based on current knowledge of early mammalian evolution on Gondwana, to see a placental mammal in the Early Cretaceous of Australia." He adds that if this fossil had been discovered in the Late Cretaceous of North America, "I don't think anyone would have doubted for an instant that it was anything but the jaw of a slightly weird placental."

Krause's own research has turned up unexpected mammals in Madagascar and India. Previously known only from Argentina, these "gondwanatheres" were announced within weeks of the A. nyktos find. And Krause notes, "Both of these discoveries together underscore to me that we really do need to keep an open mind about the biogeography of early mammals on Gondwana."

If there were placental mammals in the Early Cretaceous of Australia, Krause says, it would "push back the record of placentals farther than we expected on any southern land mass and in many ways revolutionize our concept of early mammalian biogeography."


Still, explaining how they could have gotten there remains a tall order. "Almost undoubtedly you would need some intermediate land masses to show the presence of placental mammals and right now we don't have those records," he admits. But Krause thinks his explanation of gondwanathere dispersal may be of help. "Finding gondwanatheres in Madagascar and India forced us to rethink how that distribution could make sense, and we made sense of it by suggesting that Antarctica, not Africa, served as the biogeographic link...The same kind of scenario could hold for this Early Cretaceous mammal."

Considering A. nyktos along with a contemporaneous placental from Mongolia called Prokennalestes, Rich guesses that placentals may have had a wider distribution around the world during the Early Cretaceous than previously thought. Perhaps placentals arose earlier that expected--say, while the continents were still lumped together into the single supercontinent, Pangaea. Even if terrestrial placentals were present in Australia 115 million years ago, however, Rich is quite certain that they went extinct, perhaps displaced by marsupials, and only reentered the continent 5 million years ago.

It would seem that the only way to resolve these many differing views is to find more fossils. And that's exactly what Rich and his team hope to do. They have extensive plans to continue working at the site where A. nyktos was discovered. In the meantime, the debate will surely continue. "This is the fun part," Archibald says, "discussing what it might be."