By Lucas Laursen

When a new species comes to light, its effect on the arrangement of its family tree might be better measured by statistics than by headlines. In a study of primates and flightless dinosaurs, researchers at Bristol University, UK, have found that the likelihood of any given find shaking up the family tree depends on how complete that tree was to begin with.

The catarrhine family tree--one of two primate lineages, and the one that includes humans--regularly attracts requests for revision from palaeontologists bearing fossils such as Ida, according to palaeobiologist James Tarver, lead author of the latest study. Quite a lot of this tree has already been filled in and, because knowing so much about the relatives means that it's easier to know where to put newcomers, the latter usually fit in on existing branches.

Yet the family tree for flightless dinosaurs, which palaeontologists began to construct with the discovery of the first dinosaur fossil in 1840, is still pretty patchy. As a result, it is more likely to need to sprout new branches to fit in later discoveries.

"What we've done is look at the two most intensively studied groups," Tarver says, and highlighted differences between the relatively stable catarrhine family tree, and the less certain family history of the dinosaurs. He says that statistical analysis could help to indicate which areas in a given family tree are already well-sampled and which might yet reveal more influential finds.

Linking the missing

For example, Tarver says, beyond the extant catarrhine species, there is only one known fossil of a chimpanzee species and no gorilla fossils, so more of those might be of greater value for understanding primate evolution than smaller additions to the large body of hominid fossils. "Hopefully researchers will start looking in the black holes, the larger gaps," he says--perhaps by using the statistical approach that he and his colleagues outlined August 30 in the Proceedings of the Royal Society B.

Tarver and his team tested several ways of measuring the completeness of a family tree. First, they looked at Colless's index of phylogenetic tree imbalance, which indicates whether the distribution of new discoveries throughout a family tree is random or biased. Second, the researchers examined the state of the two groups' family trees at ten-year intervals, beginning in 1790 for the catarrhines and in 1840 for the flightless dinosaurs.

They found that knowledge of catarrhine taxonomy progressed randomly, making little difference to the overall structure of the family tree. This suggests that the group has been well understood for a long time. However, because dinosaurs lived across a larger geographical area than catarrhines and most of the initial studies took place in Europe and North America, the later addition of dinosaur fossils from Asia and South America has changed the shape of the dinosaur family tree more.

Jørn Hurum, a palaeontologist at the Natural History Museum at the University of Oslo, whose classification of the fossil he called 'Ida' has come under fire, says that such a system, if adopted by palaeontologists, "would be fantastic". He adds that, in palaeontology, "phylogenies change from week to week". Tarver and colleagues' study of catarrhines did not include Ida, so they did not measure her effect on the family tree--which is something palaeontologists continue to debate (see 'Fossil primate challenges Ida's place').

"I really want to do this kind of study on early primates because then we can see how robust the phylogenies that are published just now are," Hurum adds.

However, the idea of comparing a narrow branch of primates with all flightless dinosaurs troubles other researchers who were not part of Tarver's study. Fred Grine, a palaeontologist at the State University of New York in Stony Brook, says, "basically what they're looking at comes down to nothing more than biogeography" and that "if you started finding catarrhines in South America, they'd also change the phylogeny".

But the statistics go beyond looking at how densely sampled a particular geographical site is, Tarver says, because they provide a measure of certainty that a specific group of species is well-sampled. "Just because you've searched one [geographical] area thoroughly doesn't meant you've reached that threshold where new discoveries stop having an impact [on taxonomy]," he explains.