Abstract
The recent description of an anterior caudal vertebra purportedly belonging to a diplodocid sauropod from the Early Cretaceous of China has the potential to drastically alter our interpretation of the evolution and timing of geographical dispersal of a major dinosaur lineage. However, comparison with a wider taxonomic sample points more strongly towards titanosauriform affinities for this specimen, which is in keeping with the affinities of all other sauropods known from the Cretaceous of Asia. We explain the disparity in phylogenetic interpretation of this isolated vertebra as a by-product of scoring differences and analysis of fragmentary material using repurposed data matrices. Rescoring the isolated vertebra based on our interpretation of the anatomy and rerunning the original analyses removes the specimen from Diplodocoidea but does not place it within Titanosauriformes, because of inadequacy in taxon and character sampling inherited from the repurposed data matrices. We suggest that phylogenetic analysis must begin with an initial hypothesis of affinity, based on comparative anatomy and spatiotemporal distributions, that must be adequately tested by the data matrix employed – i.e. data matrices should be tailored to sample anatomically, geographically and temporally relevant clades, and new characters should be added in tandem with new taxa so that the potential synapomorphy pool is not diluted. This is especially important for analyses of fragmentary specimens, which are likely to return coarse phylogenetic results with general evolutionary and palaeobiogeographical implications.