Abstract

The iconic moa (Aves: Dinornithiformes) from New Zealand continue to attract much scientific scrutiny, as they have done since their discovery in the 1840s. Here, we review moa research since 2001 that advances our knowledge of the biology of these families; in particular, their breeding, diet and phylogenetic relationships. Then we perform a phylogenetic analysis based on morphological characters using a broader range of taxa and many more characters than hitherto used in moa analyses. Finally, we provide revised diagnoses of all moa taxa to reflect current knowledge. In this last decade, molecular analyses have been at the forefront of much of this research. Indeed, moa have become model subjects for advances in ancient DNA technology on account of their preservation and young geological age, and the fact that several of the foremost proponents of ancient DNA research are New Zealanders. Much of this research has been about extending the capacity of ancient DNA technology as much as about answering biological questions, but the resultant insights with regard to the latter have been profound for moa. Complete mitochondrial genomes for three species of moa have been published and extensive datasets of a number of mitochondrial genes are now available for all species over their entire geographic range. Analyses of nuclear DNA is limited to a sex specific gene and some preliminary microsatellite identifications, but it seems likely that improved technology will allow greater use of this resource in the near future. Phylogenetic analyses of mitochondrial molecular data have precipitated several changes to moa taxonomy and nine species are now recognised. The significance of deep phylogenetic structure among populations in some taxa continues to attract debate and likely will require nuclear data and a more profound understanding of natural variation in extant species to resolve. Significantly, molecular data have enabled new insights into diet, with direct identification of species responsible for coprolites, and by its new-found propensity to identify eggshell, foreshadows further advances in understanding their breeding biology and distribution. Our phylogenetic analysis, based on 179 characters scored for 23 ingroup palaeognath taxa and three galloanseres as outgroups, resulted in several strongly supported relationships. Firstly, the Eocene palaeognath Lithornis was either sister to remaining palaeognaths or had a weak affinity towards tinamous. All ratites formed a monophyletic clade exclusive of tinamous. Moa were monophyletic and sister to aepyornithids in the unconstrained analysis. Attempts to constrain moa as sister to tinamous to reflect molecular-based conclusions resulted in moa as sister to all ratites in a clade that was unresolved with respect to tinamous and Lithornis. This relatively basal position of moa was not a significantly worse reflection of the data compared to their more crownward location in the initial analyses. The casuariids were sister to Struthio and the rheas. In our revised diagnoses for Dinornithiformes and all its constituent taxa, we give updated information on the type specimens based on recent research by the authors. We accept three families, six genera and nine species, and make the new combinations of Euryapteryx curtus curtus (Owen) and E. curtus gravis (Owen). Complete or near complete exemplars of the skull of all moa taxa, most not illustrated before, are shown in dorsal, lateral and ventral views.