These changes arise
from two principal corrections: 1) recounting the molars and 2) identifying vestigial prefrontal and postfrontal bones wherever present. Often enough these have been overlooked by prior workers. That’s easy to do.
Giant Andrewsarchus and smaller Sinonyx have moved from
the Tenrec + Odontoceti clade to the Vulpavus ovatus clade (Figs. 1, 2). This clade of terrestrial tree shrews arose in the Paleocene following the demise of the dinosaurs. With Phenacodus in this lineage known for relatively small brains and giant jaw muscles, these all look to have been bear-like to hippo-like herbivores, despite the fang-like canines (convergent with pigs, hippos and musk deer).
Not only does the LRT grow by adding taxa,
it also gets pruned and corrected. Taxa that were added a decade ago are reexamined in light of better data brought on by more taxa. Reversals and convergence can be tricky. Freshman mistakes, some due to traditional errors not yet recognized by workers, some due to their order of addition, continue to be corrected. The root of the Placentalia continues to be fragile, completely resolved, but now less prone to change due to scoring errors. Corrections are not yet, nor ever completed. That’s the way paleontology should be handled, but rarely is.
Spaulding et al 2009
considered Andrewsarchus an artiodactyl entelodont (= large extinct pig relative). The LRT tests entelodonts. Andrewsarchus is not an entelodont. Taxon inclusion and accurate scoring resolves all such issues.
it only takes a few scoring changes to move Andrewsarchus and kin closer to tenrecs and elephant shrews, where they nested for the last decade. This is convergence at its trickiest.
Spaulding, M O’Leary MA Gatesy JH 2009. “Relationships of Cetacea (Artiodactyla) Among Mammals: Increased Taxon Sampling Alters Interpretations of Key Fossils and Character Evolution”. PLOS ONE. 4 (9): e7062. Bibcode:2009PLoSO…4.7062S. doi:10.1371/journal.pone.0007062