RE: 11th specimen of Archaeopteryx

[Mickey Mortimer <mickey_mortimer111@msn.com>]

hammeris1@att.net wrote-

> So ... Wikipedia is a sick puppy that probably can never be cured, but can
> someone give me the latest/greatest paper, speech, book stating what the most
> recent view is of where Archie is? Is it still in Aves? Is it the majority
> view that it is a split-off no longer in the direct line of birds?

The majority view, as in the conclusion found by almost every cladistic 
analysis which has tackled the problem, is that Archaeopteryx is a basal 
avialan.  Avialae as most commonly used consists of taxa closer to modern birds 
than to Deinonychus.  Aves has a few possible definitions.  It is often used 
today as the group of crown birds (paleognaths plus neognaths) which 
Archaeopteryx certainly isn't a member of, but has also been commonly defined 
as the group consisting of everything at least as closely related to modern 
birds as Archaeopteryx (which by definition includes Archaeopteryx).

The recent controversy has been over the analysis in Xu et al.'s (2011) 
description of Xiaotingia.  This is a version of the Theropod Working Group 
analysis which goes back to Norell et al. (2001).  Specifically Xu et al. 
(2011) added a few taxa and several characters to...
- Zhang et al.'s (2008) analysis which added Epidexipteryx and a few characters 
to...
- Senter's (2007) analysis which completely recoded and added many characters 
and taxa to...
- Kirkland et al.'s (2005) analysis which added several characters and a few 
therizinosaurs to...
- Hwang et al.'s (2004), which is based on Xu et al.'s (2002), which is based 
on more analyses still all the way back to Norell et al. (2001).

So it's a re-re-re-re-re-analysis of a huge dataset.  Running this dataset with 
Xiaotingia results in Archaeopteryx being a basal deinonychosaur instead of a 
basal avialan.  Note this isn't a big move, since Deinonychosauria and Avialae 
are sister groups.  It just moves from the base of one group to the base of the 
other.  Note also that despite what the hype would indicate, this result isn't 
very well supported.  Forcing Archaeopteryx back to its normal position as a 
basal avialan only takes TWO more evolutionary steps.  That's not significant 
at all, and furthermore the analysis itself is flawed as detailed below.  

When you run an analysis like this, certain characters need to be "ordered".  
So that for instance, taxa with six sacral vertebrae are seen as intermediate 
between taxa with five and taxa with seven sacrals.  If you don't order the 
character, "six sacrals" is counted as a character that has no definite 
relationship to other numbers of sacral vertebrae, so you'd get weird results 
like grouping Rahonavis and Shenzhouraptor together to the exclusion of more 
derived birds because of their shared primitive sacral number.  Also, ordering 
changes the amount of steps a character takes to evolve, since if it's 
unordered, you can go from five to nine sacrals as easily as you can go from 
five to six sacrals.  When we try to find out if Xu et al. ordered their 
characters, we have to follow the lineage of analyses all the way back to 
Kirkland et al.'s (2005) version, which only says one character was ordered but 
doesn't say which.  Xu et al. seem to ignore that anyway, since I get their 
results by running their analysis unordered.  Running Xu et al.'s analysis with 
all characters ordered adds over 100 steps, which of course completely 
overpowers our two step difference we noted above.  While not every character 
should be ordered, many should be.  Importantly, when characters are ordered, 
Archaeopteryx comes out as a bird.

So how much of Xu et al.'s result is due to not having the right characters 
ordered?  We don't know unless someone goes through the tedious steps of 
looking through all the characters and choosing which should be ordered.  You 
can see how this could be important for Archaeopteryx, since any intermediate 
state it has between birds and deinonychosaurs will be counted as equally 
different from both instead of being a bit closer to birds (assuming the 
deinonychosaurian condition is primitive).

Another problem is that Xu et al.'s analysis doesn't include all of the 
relevent taxa and characters that other versions of the analysis do.  Senter's 
newest (2010) analysis (which is a modification of his 2007 one) includes most 
of the same taxa but has many new codings and takes seven more steps to place 
Archaeopteryx in Deinonychosauria.  Zanno et al.'s (2009) analysis (which has a 
rather different lineage going back to Hwang et al. 2004 and so doesn't include 
any of Senter's numerous modifications) contains a different mix of characters, 
adds Mahakala and Shanag, but lacks scansoriopterygids, Sapeornis, Protopteryx, 
NGMC 91 and Bambiraptor.  Forcing deinonychosaurian Archaeopteryx is six steps 
longer in it.  Makovicky et al.'s (2010) analysis (which is more similar to 
Zanno et al.'s) includes yet a different mix of characters and taxa needs eight 
more steps.  Most recently, Turner et al. (2011) have a TWG-based analysis 
centered on deinonychosaurs and birds, including taxa not found in the 
Xiaotingia analysis like Hesperonychus, Graciliraptor, Tianyuraptor, 
Austroraptor, Mahakala, Jinfengopteryx, two undescribed basal troodontids, 
Jixiangornis and a lot of birds.  Based on their taxon sample I bet they also 
included the numerous bird-related characters of Clarke's analyses.  And this 
analysis found Archaeopteryx to be a bird, though I can't say how well 
supported that is since they haven't released their data matrix yet (grrr).

So we can see that most analyses find Archaeopteryx to be 6-8 steps more likely 
to be a bird, while Xu et al. found it to be 2 steps more likely to be a 
deinonychosaur.  Each analysis includes some data others don't, and all have 
miscodings.  Until someone combines the information (which I'm finishing up), 
we won't know if say adding Xiaotingia to Turner et al.'s analysis would make 
Archaeopteryx a deinonychosaur, or if adding Jinfengopteryx to Xu et al.'s 
analysis would make Archaeopteryx a bird.

Until that time, I'd say it could be either, but that both the number of 
analyses and the strength of support in those analyses slightly favor it being 
a bird.  

> And is it
> felt that it does have known close-ancestors in the record?

(from The Theropod Database...)

Several taxa have been referred to Archaeopterygidae or Archaeopteryx itself in 
the past, but most do not belong there. Parkinson (1930) and several other 
authors have alluded to an Archaeopteryx-like specimen from the Tendaguru 
Formation of Tanzania, but this is based on a misreading of Stremme 
(1916-1919), who noted the isolated carpometacarpus is unlike Archaeopteryx. It 
is probably paravian, and may be an ornithurine. Lambrecht (1933) cites the 
species Archaeopteryx "vicensensis", but this nomen nudum was later claimed to 
be pterosaurian (Kleinschmidt pers. comm. to Brodkorb, 1978). Jensen (1981) 
identified a proximal femur (BYU 2023) as Archaeopteryx, but it is also similar 
to basal dromaeosaurids such as Microraptor and Unenlagia. Kessler and Jurcsak 
(1984) described an incomplete humerus from the Early Cretaceous of Romania as 
Archaeopteryx sp., but it could easily belong to another small maniraptoran as 
well. Paul (1988) referred all dromaeosaurids to Archaeopterygidae, which has 
not been recovered in any phylogenetic analysis. "Proornis" was originally 
called "the North Korean Archaeopteryx" (e.g. anonymous, 1995), but seems to be 
a confuciusornithid instead. Weigert (1995) described 103 teeth from the 
Guimarota Formation of Portugal as cf. Archaeopteryx sp., but they are not 
avialan and may belong to a basal deinonychosaur instead. Protarchaeopteryx was 
assigned to the family by Ji and Ji (1997) and Paul (2002), but is a basal 
oviraptorosaur. Forster et al. (1998) found Rahonavis and Unenlagia to clade 
with Archaeopteryx in some most parsimonious trees, but Rahonavis has generally 
been found to be an ornithurine or dromaeosaurid since, while Unenlagia has 
been found in basal Avialae or Dromaeosauridae. Rauhut (2002) referred 
Paronychodon to Archaeopterygidae, but this was based on comparisons to the 
Guimarota teeth noted above. Ji et al. (2005) described Jinfengopteryx as being 
more closely related to Archaeopteryx than to Aves, but is now recognized as a 
basal troodontid. Xu et al. (2011) rece!
ntly used
er's (2007) TWG analysis to refer Anchiornis and Xiaotingia to 
Archaeopterygidae, but this requires further confirmation.

Notably, when Archaeopteryx is constrained to be a bird in Xu et al. (2011), 
Anchiornis and Xiaotingia are still deinonychosaurs.  When the analysis is run 
with all characters ordered they form a clade of birds more basal than 
Archaeopteryx.

Mickey Mortimer