turning turtles

[Bill Adlam <sagitta_elegans@yahoo.com>]

I've now got my hands on the Hedges & Poling turtle  paper in Science.
It's entitled 'A molecular Phylogeny Of Reptiles' (by reptiles they
seem to mean all non-mammalian amniotes).

To summarise, the maximum likelihood phylogeny, based on 11 nuclear
proteins (1943 amino acids) was:
(Rodents + Primates) + (Squamates + (Birds + (Crocs + Turtles))).
Tuataras weren't included due to lack of data.

A consensus phylogeny based on 4 genes (785 amino acids)
(Rodents + Primates) + (Squamates + (Turtles + Crocs + Birds +
Tuataras)).

To me the big surprise is not so much that turtles are grouped with
archosaurs as that they come within archosaurs, closer to crocs than
to birds.  This node is very strongly supported, with various
techniques giving confidence measures of 97 to 100%.  The
turtle-croc-bird node is supported 66 to 90%.

I would like to think that the unusually high rate of evolution in
birds has made them so different that they have lost some archosaur
apomorphies, leaving crocs more like turtles, but I don't think that's 
sustainable.  (That would be a case of short branches attracting.)
Hedges & Poling's maximum likelihood phylogeny actually has birds as
the most similar to the ancestral amniote.

Apart from the nuclear amino acid data, they looked at a few nuclear
nucleic acid genes, which support the same conclusions, and a few
mitochondrial genes, which don't.

These and other genes were used as a molecular clock, calibrated by
the divergence of synapsids at 310Ma.  According to the clock,
squamates diverged at 245 +-12Ma, birds at 228 +-10 and crocs at 207
+-20.  I.e. early, mid and end Triassic, respectively.  Crocodiles are 
known from around the divergence time given, and turtles from a little 
before.  However, there would be a 100 million year gap for squamates. 
As for Triassic birds, I'm not saying anything.

Even if the clock is discounted, the phylogeny still requires a very
early origin for (stem) squamates.  The authors note that there is
virtually no morphological support for the turtle-croc clade.  They
suggest dermal armour as a feature linking the groups, and note that
aetosaurs shared this. Some aetosaurs had one or two other chelonian
features.

Bill
_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com