In 2012, the controversial case over whether or not Archaeopteryx lithographica, perhaps the most iconic dinosaur species of all time, was a bird was settled. Apparently. (free pdf) This was an important analysis for two reasons. Firstly, it countered a previous study showing that Archaeopteryx was more closely related to dinosaurs like Velociraptor and Deinonychus, and secondly used advanced, sort of non-traditional methods in palaeontology, called maximum likelihood and Bayesian analysis, to work out its relationships. None of this is news. It was extensively covered upon release way back, and also revisited recently with the discovery of Aurornis xui, which was found to be the most basal bird (i.e., that which was closest to the origin of the group, typically known as Avialae, or Aves).
Now a new study, from the December issue of a journal called Cladistics, has re-examined the data from the 2012 study, asking the question of whether or not the analyses they used even made sense to do in the first place. This is actually quite important. Maximum Likelihood and Bayesian analysis where designed with models in mind for describing the evolution of nucleotide data, that is genetic sequence data. Whether or not they can actually be used in fossil data, or that which is based purely on anatomy, is questionable, as anatomical features, or characters, evolve in a different manner to genetic data (e.g., there aren’t anytransversions or transitions). Given how often now these methods are used for genetic data, and to great effect, it’s a bit weird that it seems to have been largely overlooked by palaeontologists and scientists who use morphology to reconstruct evolutionary relationships.
Usually, palaeontologists use a method called parsimony analysis, which is a much simpler model for generating evolutionary trees. To use any of the programs below, download one of the nexus files from here (dinosaurs!), and run it with the program – worth a fiddle for a bit of science-ing!