Cretaceous, yes -- Netherlands, yes -- marsupial, unlikely

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

The most spectacular news first. Mammal hair is mentioned in passing to 
occur in amber from the Campanian of Siberia!!!

-----------------------------------

I've now read the paper:

James E. Martin, Judd A. Case, John W. M. Jagt, Anne S. Schulp & Eric W. A. 
Mulder: A New European Marsupial Indicates a Late Cretaceous High-Latitude 
Transatlantic Dispersal Route, Journal of Mammalian Evolution 12(3/4), 
495 -- 511 (December 2005)

The authors don't even try to find out whether *Maastrichtidelphys*, called 
a "pouched mammal" once, is a crown-group marsupial or not. Before the 
paleobiogeography section the word Metatheria is never mentioned; even in 
it, *Didelphodon*, peradectids and pediomyids are "marsupials". Only 
*Marsasia* and *Asiatherium* are not "marsupials". It seems to follow that 
"Marsupialia" is used for something like the "Neometatheria" of 
http://www.fmnh.helsinki.fi/users/haaramo/Metazoa/Deuterostoma/Chordata/Synapsida/Metatheria/Metatheria.htm 
or the "Archimetatheria" of 
http://www.fmnh.helsinki.fi/users/haaramo/Metazoa/Deuterostoma/Chordata/Synapsida/Metatheria/Metatheria_2.htm, 
which may or may not be the crown-group depending on the phylogeny, and is 
much bigger than the crown-group on the currently accepted phylogenies. 
Which phylogenies Martin et al. consider likely (let alone why!) is never 
explained, and neither do they even just mention that some (most??) people 
use a different definition of Marsupialia. IMHO this counts as a significant 
failure of peer-review.

Martin et al. say *Maastrichtidelphys* is a herpetotheriid, say that 
Herpetotheriidae is known from the early Maastrichtian (of North America) 
onwards, and say that this group is "seemingly related to the North American 
opossum". In the "Systematic Paleontology" section, they accordingly put 
Family Herpetotheriidae into Superfamily Didelphoidea and Order 
Didelphimorphia, which they put into Cohort Ameridelphia -- I'm certain they 
know that this latter taxon is paraphyletic. (Marsupialia is a supercohort, 
in case you're wondering...) In other words, they clearly don't think that 
only clades (and species) should be named.

*Maastrichtidelphys* "shares the greatest number of character states with 
species (albeit not all) of" the early Eocene herpetotheriids 
*Amphiperaterium* (Europe) and *Garatherium* (Africa). I wonder what this 
exercise in phenetics is for.

Despite the phenetics, a phylogenetic analysis is done. *Nortedelphys* 
(described a year earlier by the same authors) comes out as a 
herpetotheriid.

First analysis -- only North American taxa and *Maastrichtidelphys*, to find 
out which of the former is closest to the latter:

--+--+--*Maastrichtidelphys*
 |  `--*Nortedelphys*
 `--+--*Didelphodon*
    `--+--*Alphadon*
       `--+--*Albertatherium*
          `--+--*Protalphadon*
             `--+--*Pediomys*
                `--*Turgidodon*

The authors do not explain how this tree was rooted.

Second analysis -- to find out where *Maastrichtidelphys* lies among "other 
latest Cretaceous and Paleocene didelphoid marsupial taxa":

--+--*Didelphodon*
 `--+--*Alphadon*
    `--+--*Jashkadelphys*
       `--+--*Mizquedelphys*
          `--+--*Pucadelphys*
             `--+--*Itaboraidelphys*
                `--+--*Maastrichtidelphys*
                   `--*Nortedelphys*

This tree is rooted with *Didelphodon* and *Alphadon* as outgroups. Note 
that the topology contradicts the first tree. *Pucadelphys* and 
*Itaboraidelphys* come from the Paleocene of South America, making the idea 
that Herpetotheriidae belongs to the crown-group more plausible 
(*Pucadelphys* is close to the crown, but apparently outside). But note that 
the monophyly of "Didelphoidea" is not tested; not one non-"didelphoid" is 
included, if we assume that the authors still include everything in 
"Didelphoidea" that was included _fifty_ years ago.

I could obviously be wrong, but I conclude that the authors have likely a 
priori assumed that their "Didelphimorphia" and "Didelphoidea" are clades 
and built their conclusions (like a South American origin for 
*Nortedelphys*) on this assumption that they have not tested. Both of their 
poorly supported cladograms can probably be safely ignored (in spite of the 
rather high number of characters -- 41 in total).

On the other hand, the biogeography looks quite robust, despite the 
disgusting mixture of Cretaceous and recent coastlines in fig. 4.

BTW, the specific epithet should end in -orum instead of -i. Mr Meuris and 
Mr Smet are two people. <blood pressure rising>