Supertrees are a useful method of constructing large-scale phylogenies by assembling numerous smaller phylogenies that have some, but not necessarily all, taxa in common. Birds are an obvious candidate for supertree construction as they are the most abundant land vertebrates on the planet and no comprehensive phylogeny of both extinct and extant species currently exists. In order to construct supertrees, primary analysis of characters is required. One such study, presented here, describes two new partial specimens belonging to the Primobucconidae from the Green River Formation of Wyoming (USA), which were assigned to the species Primobucco mcgrewi. Although incomplete, these specimens had preserved anatomical features not seen in other material. An attempt to further constrain their phylogenetic position was inconclusive, showing only that the Primobucconidae belong in a clade containing the extant Coraciiformes and related taxa. Over 700 such studies were used to construct a species-level supertree of Aves containing over 5000 taxa. The resulting tree shows the relationships between the main avian groups, with only a few novel clades, some of which can be explained by a lack of information regarding those taxa. The tree was constructed using a strict protocol which ensures robust, accurate and efficient data collection and processing; extending previous work by other authors. Before creating the species-level supertree the protocol was tested on the order Galliformes in order to determine the most efficient method of removing non-independent data. It was found that combining non-independent source trees via a “mini-supertree” analysis produced results more consistent with the input source data and, in addition, significantly reduced computational load. Another method for constructing large-scale trees is via a supermatrix, which is constructed from primary data collated into a single, large matrix. A molecular-only tree was constructed using both supertree and supermatrix methods, from the same data, again of the order Galliformes. Both methods performed equally as well in producing trees that fit the source data. The two methods could be considered complementary rather than conflicting as the supertree took a long time to construct but was very quick to calculate, but the supermatrix took longer to calculate, but was quicker to construct. Dependent upon the data at hand and the other factors involved, the choice of which method to use appears, from this small study, to be of little consequence. Finally an updated species-level supertree of the Dinosauria was also constructed and used to look at diversification rates in order to elucidate the “Cretaceous explosion of terrestrial life”. Results from this study show that this apparent burst in diversity at the end of the Cretaceous is a sampling artefact and in fact, dinosaurs show most of their major diversification shifts in the first third of their history.
PDF
download
878987797
028-85220240
