The discovery in Ethiopia's Middle Awash region of a handful of nearly six-million-year-old teeth is adding fuel to a long-standing debate among scholars of human evolution. At issue is whether the base of our family tree is as streamlined as a saguaro or as shaggy as a shrub.

When it comes to classifying fossils, paleoanthropologists generally fall into two camps. There are the splitters, who parse the human fossil record into numerous genera and species, and the lumpers, who recognize fewer, more variable taxa. Both factions agree that several hominid species coexisted during the later stages of human evolution, between three million and 1.5 million years ago. The number of forms that shared the landscape shortly after humans diverged from chimpanzees, on the other hand, is vigorously disputed.

In a report detailing the new findings, published today in the journal Science, Yohannes Haile-Selassie of the Cleveland Museum of Natural History and his colleagues assign the ancient teeth to a new hominid species, Ardipithecus kadabba. In light of the discovery, the team argues, remains previously attributed to the subspecies A. ramidus kadabba should now be considered part of the new species, which is older and more primitive than A. ramidus.

Particularly important in their analysis are the upper canine and lower third premolar that turned up. All fossil and modern apes, particularly males, have large, tusklike canines that are continually honed against the lower third premolars, which keeps them sharp for fighting (mostly over access to mates). Humans, in contrast, have smaller, more incisorlike canines, which scientists have interpreted as indicative of increased male cooperation. For their part, the A. kadabba canine and premolar exhibit a mix of apelike and hominidlike traits, prompting Haile-Selassie to speculate that this species might be the first on the human line after the chimp-human split.

A. kadabba is not the lone contender for the title of earliest member of the human lineage. Two other putative hominids dating to the late Miocene epoch--Sahelanthropus tchadensis from Chad and Orrorin tugenensis from Kenya--surfaced in 2002 and 2000, respectively. But Haile-Selassie and his collaborators suggest that the teeth of these specimens indicate that they are very similar to A. kadabba. On the basis of the available evidence, they contend, all three may belong to the same genus or even species.

A contrary view comes from David R. Begun of the University of Toronto, who counters that the A. kadabba, Sahelanthropus and Orrorin dentitions differ in important ways. "Rather than a single lineage, the late Miocene [hominid] fossil record may sample an adaptive radiation, from a source either in Eurasia or yet undiscovered in Africa, the first of several radiations during the course of human evolution," he writes in an accompanying commentary. But the level of uncertainty about the fragmentary fossils known thus far makes it impossible to reconcile these differences of opinion between lumpers and splitters. "The solution is in the mantra of all paleontologists," he concludes. "We need more fossils!"