Re: back to Blogging SVP

["David Marjanovic" <david.marjanovic@gmx.at>]

> Could someone add info as to which of the abstracts does this finding refers 
> to?

It's not mentioned in the abstract:

>>
Poster Session I (Wednesday, October 30, 2013, 4:15 - 6:15 PM)
A PHYLOGENETIC ANALYSIS OF PALAEOCENE MAMMALS: IMPLICATIONS FOR THE ORIGIN OF 
PLACENTAL MAMMAL ORDERS
HALLIDAY, Thomas, University College London, London, United Kingdom; UPCHURCH, 
Paul, University College London, London, United Kingdom; GOSWAMI, Anjali, 
University College London, London, United Kingdom

The Palaeocene is arguably the most important interval in mammalian evolution. 
Prior to the Cretaceous-Palaeogene (K-Pg) extinction, the mammal fauna is 
largely, though not exclusively, restricted to the scansorial or terrestrial 
insectivore niche; afterwards, there is a broad diversity of large herbivores, 
large carnivores, and later gliders, flyers and aquatic forms quickly evolved.

The majority of the taxa known from the Palaeocene, however, belong to 
‘wastebasket’ taxa or clades of unknown affinity – these include cimolestids, 
pantodonts, and the “condylarths”. Only Rodentia, Carnivora and possibly 
Primates have well-supported Palaeocene members. Clarifying the relationships 
of Palaeocene mammals is thus essential for any reliable macroevolutionary 
study into the early phases of placental mammal evolution.

Here we present the results of an extensive cladistic study of fossil mammals, 
focusing on Laurasiatheria and possible laurasiatheres, with 130 taxa coded for 
681 dental, cranial and postcranial characters. Preliminary analyses in TNT 
resulted in 2448 trees of length 6474 steps when constraining Afrotheria as a 
monophyletic group. A strict consensus after pruning the seven least stable 
taxa yields highly resolved relationships for the enigmatic Palaeocene mammals, 
including the majority of the “condylarth” groups. Relationships between extant 
taxa are largely upheld, although miacid and viverravid carnivorans are not 
recovered as monophyletic, and Eulipotyphla is not recovered. “Condylarths” are 
found to be polyphyletic, as expected, with apheliscids and pleuraspidotheres 
falling closest to Artiodactyla, and phenacodontids closest to Perissodactyla. 
Arctocyonids are reconstructed as a paraphyletic lineage leading to miacid 
carnivorans, while periptychids lie at the base of a clade containing the 
majority of the non-euungulate Laurasiatheria. Cimolestids lie on the eutherian 
stem and are separate from Pantodonta. One novel result is the placement of 
Leptictida as sister taxon to Afrotheria. With the exception of the leptictid 
Gypsonictops, all Cretaceous taxa are resolved as stem Eutheria, supporting a 
Palaeocene origin for the majority of placental orders.

There is much debate over the role of the K-Pg extinction in the origin of the 
placental mammal orders. This phylogenetic analysis will provide a useful basis 
for many future studies of major evolutionary patterns in early crown placental 
mammals.
<<

As expected for an SVP meeting abstract (the submission deadline was when 
again, in May?), it was a bit outdated by the time of the presentation. :-) In 
particular, there is no Euungulata ( = Artio- + Perissodactyla to the exclusion 
of all other extant taxa), instead the artiodactyl total group is the 
sister-group to all other laurasiatheres; all phenacodontids except 
*Phenacodus* come out as stem-perissodactyls, *Phenacodus* falls on the 
artiodactyl side; the "arctocyonids" aren't all stem-carnivorans, but are 
spread out over the stems of most or all laurasiatherian lineages. I don't 
remember if the placement of Leptictida held up, but I think they fell on the 
eutherian stem.

Also not mentioned in the abstract is that the arctostylopids come out as the 
sister-group of Atlantogenata ( = Afrotheria + Xenarthra to the exclusion of 
all other extant taxa). That's pretty fascinating (they're from Asiamerica, but 
named after the notostylopid notoungulates from South America and have long 
been thought to be possible relatives of some or all South American 
"ungulates"). While Afrotheria was constrained to exist, Atlantogenata was not!

> Additionaly, refering to some of the previous posts, if David Peters actually 
> was the first to propose Chriacus - bats close relationships, and it will 
> prove to be right in the future, he shouldn't be taken credits back ("just a 
> lucky strike") for this idea here on DML just because many of his other ideas 
> are generally treated as unscientific.

I don't remember if he found bats and *Chriacus* as sister-groups, and didn't 
think of looking it up; if he did, I won't withhold credit. However, in that 
case, I'd very much like to see DP's data matrix first, given my *sigh* 
experience with his amniote matrix. -- I do, however, and have always credited 
him for insisting that as many taxa as possible should be included in 
phylogenetic analyses. He's not the first to say that, but he's consistently 
been saying and doing it for years.