Important new paper by Susanna Kümmell et al. on carpal evolution in synapsids: http://app.pan.pl/archive/published/app65/app007092019.pdf
Most notably overturns ‘accepted wisdom’ on homology of the mammalian lunate bone, arguing it is not derived from the ‘reptilian’ intermedium but instead the lateral centrale.

Homology between the lunate and intermedium was proposed based on extant taxa by Gegenbauer in 1864, codified with reference to fossil synapsids by Broom and Romer in the early 20th century, and has stood largely unopposed since then.

Looking at the excellent wrist specimens of Dimetrodon studied by Romer, this homology makes sense: the intermedium and lunate are in the same place in modern reptiles/mammals, and in pelycosaurs the intermedium is the largest bone occupying this space.

However, reliance on Dimetrodon shows a pitfall of focusing on exemplars distant from the crown rather than covering the whole synapsid lineage. Because in therapsids, the intermedium is quite a small bone, and the dominant element in this region is the lateral centrale.

Though this may not be the final word on the subject, it does make more sense that the already-reduced intermedium was lost in mammals, and that it was instead the lateral centrale that became the mammalian lunate.

It’s a good example of how important it is to re-examine ’textbook’ hypotheses initially based on very limited fossil data. Crucially, the new hypothesis only became clear based on detailed, personal examination of the vast holdings of synapsid fossils around the globe.

The evolution of mammals from early cynodonts is still one of the absolute best-understood macroevolutionary transitions in the fossil record, but it’s a mistake to think of mammal origins as a ‘solved problem’--within the broader framework there’s still lots to learn!