Were There Miocene Meridiolestidans? Assessing the Phylogenetic Placement of Necrolestes patagonensis and the Presence of a 40 Million Year Meridiolestidan Ghost Lineage

The enigmatic mammal Necrolestes patagonensis from the Miocene of Patagonia possesses a highly apomorphic osteological form that has confounded phylogenetic interpretation for over a century. In this time it has been affiliated with both eutherians and metatherians; however, a recent study by Rougier et al. (Proc Natl Acad Sci USA 109:19871-19872, 2012) raises the intriguing possibility that Necrolestes is a relictual member of a clade of South American non-therian dryolestoids, the Meridiolestida. This group is known chiefly from the Cretaceous of South America and assignment of Necrolestes to Meridiolestida implies a ghost lineage of about 40 million years. Such a lengthy ghost lineage requires strong evidence, which minimizes potentially circular assumptions of anatomical homology. Here, we vary the coding of cusp homologies in Necrolestes, previously assumed to diverge from the metatherian pattern, and add zalambdodont and incipiently zalambdodont metatherian taxa to the analyses, in order to assess the effects of non-independent characters and taxon sampling on the original topology. The results of both maximum parsimony and Bayesian analysis using the Mk model show that these possible sources of bias have little effect on the topology and ultimately increase confidence in the placement of Necrolestes in Meridiolestida and its concomitant 40 million year ghost lineage. Additionally, our Bayesian analysis resolves Australosphenida in a trichotomy with Peramus and Vincelestes + Boreosphenida. This contrasts with the majority of existing topologies, and raises interesting questions regarding both the evolution of tribospheny and the use of the Mk model with paleontological datasets.