Re: No Cretaceous placental mammals?

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

Now I can answer my own questions:

> Isn't there a pregnant leptictid in Messel or something?
> Or am I making this up?

In Messel at least there's none; thanks to HP Trevor Dykes.

> But in any case leptictids have never been reported to
> have epipubes or a narrow pelvis, AFAIK.
> If the Maastrichtian tooth taxon *Gypsonictops* is a
> leptictid, we have placental reproduction in the Mesozoic.

Indeed leptictids lack epipubes, indeed this is autapomorphic of the 
(Leptictidae + Placentalia) clade, and indeed *Gypsonictops* is a leptictid 
(one of two in the matrix of Wible et al.); thanks to HP Tom Holtz. The 
placental mode of reproduction must have evolved very late, apparently in 
the Maastrichtian, just in time for the mass extinction.

More on the phylogenetic analysis:

- 21 characters are parsimony-uninformative; one of them is even constant. 
That makes 387 rather than 408 parsimony-informative characters. Of course 
that's still a lot!
- There seems to be some duplication in the matrix: for example the number 
of premolars, the number of molars, and the number of postcanines are three 
separate characters.
- All characters were unordered. Someone should have a look what happens if 
this is changed. I wonder if the *Purgatorius*-*Protungulatum*-*Oxyprimus* 
clade depends on this.
- I think every single named zhelestid, asioryctithere, and zalambdalestid 
is in the matrix. Surprisingly, these come out as four distinct clades, with 
most of the zhelestids forming one of them.
- Zalambdalestidae and Glires are very far apart, even though their 
enlarged, ever-growing lower incisors were coded as homologous to bias the 
results towards the assumption that Cretaceous placentals are known. As 
further explained in the supplementary information, the teeth are almost 
certainly not homologous: embryology says the lower gnawing tooth in Glires 
is the no-longer-deciduous second incisor of the standard placental three, 
while in Zalambdalestidae it is the first of the standard therian four. 
AFAIK no therian with more than four lower incisors is known, so the 
zalambdalestids have most likely not lost an incisor in front of the 
enlarged one.
- *Deccanolestes* (consisting of two species of Maastrichtian eutherian from 
India, each named for an isolated astragalus, IIRC) is in the matrix! It 
comes out as the sister-group of (Zalambdalestidae + Placentalia), far from 
both Placentalia and Cimolestidae (one or two previously supported 
positions).
- There are two carnivorans (*Miacis* and *Vulpavus*), the sister-group of 
the crown of Cetartiodactyla (*Gujaratia* -- formerly *Diacodexis 
pakistanensis* which would have made *Diacodexis* paraphyletic), and three 
"mainstream condylarths" (*Hyopsodus*, *Meniscotherium*, *Phenacodus*) in 
the matrix; together with *Erinaceus* (western European hedgehog, or maybe 
the eastern one, I forgot), *Blarina* (a shrew), *Solenodon*, and *Eoryctes* 
(an Eocene "insectivore"), this is all of "Laurasiatheria" that is in the 
matrix. Someone should add more.
- Especially, *Apheliscus* and friends, found close to Macroscelidea in 
other analyses, are not in.
- *Potamogale*, the only afrosoricidan, comes out as the sister-group of 
*Eoryctes* (see above), but all other afrotherians form a clade: 
((*Procavia* [hyrax] + *Moeritherium* [proboscidean]) + (*Orycteropus* 
[aardvark] + *Rhynchocyon* [elephant shrew])) -- and that without Eocene 
hyraxes, Eocene elephant shrews, or apheliscines in the matrix!
- Interestingly, the three "mainstream condylarths" form a clade. I wonder 
if adding perissodactyls would change that.
- Except for the "condylarths" (the mentioned ones plus *Protungulatum* and 
*Oxyprimus*), no representatives of Paleogene-only "orders" are included. Is 
*Uintatherium* a Giant Horned Bunny? Are taeniodonts "cimolestans"? I smell 
a couple of surprises waiting to be discovered.
- Over the next few years I expect a series of papers to be built on this 
matrix. Now that people who want to investigate eutherian phylogeny no 
longer need to build their own matrix from scratch, but can simply add taxa 
and maybe a handful of characters to an existing one, the motivation for and 
the ease of finding out where, say, the South American "ungulates" belong or 
which planet the bibymalagasy comes from have without doubt greatly 
increased!

In other words, this paper was just the beginning! :-)