Phylogenetic fusion patterns in pterosaurs

This post has been modified from its original content. It’s important to remember that pterosaurs are lizards. They follow lizard-type growth patterns as reported by Maisano 2002 in which some lizards fuse bones and keep growing while others never fuse certain bones into old age. Pterosaurs also grow isometrically, with long-snouted, small eyed embryos known.

Traditional thinking follows the paradigm
that the unfused scapulocoracoid (s/c) in pterosaurs demonstrates immaturity. I tested this in a phylogenetic analysis. Turns out the patterns are not ontogenetic, but clearly phylogenetic. Scapulocoracoid fusion is on again, off again in patterns that are not the random pattern one would expect if ontogenetic in nature.

 

Figure 1. Click to enlarge. Pterosaur family tree (May 2013) highlighting scapulocoracoid fusion in pterosaurs (bright green) and lack of fusion (bright blue). Other taxa do not preserve the s/c. If ontogenetic we would expect a more scattered, randomized pattern. That’s not the case here as fusion patterns follow phylogeny, not maturity. Some taxa here do not preserve the scapula and coracoid. Not listed here, but related to Cearadactylus, Barbosania does not fuse the s/c. Some taxa have complete fusion. Others retain a line of fusion. Among the higher ornithocheiridae there is the greatest randomness in fusion.

Pterodaustro is known from embryos to fully mature individuals
Codornú et al. (2013) report on 300+ individual specimens from a single bone bed: ”Interestingly, proxies for full skeletal maturation are thus far present only in isolated elements (i.e., all complete or semicomplete specimens belong to osteologically immature individuals). These proxies include the complete fusion (lack of any sutural evidence) between the extensor tendon process and the shaft of the first wing phalanx, the complete fusion between the tibia and the proximal tarsals, and the fused distal secondary ossification centers of the humerus.” Note they did not report fusion of the scapula and coracoid. That’s because Pterodaustro nests in a clade (Fig. 1) that does not fuse the scapulocoracoid.

So what’s the pattern?
Basal pterosaurs do not have a fused scapulocoracoid. Dimorphodon may have a fused s/c. Campyognathoides and basal Dorygnathus fuse the s/c. Basal Rhamphorhynchus specimens are smaller and lack fusion. Derived Rhamphorhynchus regain fusion. Dorygnathid pre-azhdarchids beginning with tiny TM 10341 lose fusion. Large azhdarchids regain fusion. No ctenochasmatid or dorygnathid pre-ctenochasmatid fuse the scapulocoracoid. Jianchangnathus and all subsequent scaphognathids lose fusion. Basal ornithocheirds, no matter how large their wings are do not fuse the s/c. Certain, but not all derived ornithocheirds regain fusion. On another branch of scaphognathids, certain germanodactylids regain fusion. Shenzhoupterids and basal tapejarids lose fusion. Derived tapejarids, the big ones, regain fusion. (Does anyone have a good dsungaripterid scapulocoracoid? I haven’t seen one yet.) Germanodactylids including Pteranodon have fusion (not sure about basal taxa because so many are known just by skulls), but eopteranodontids and nyctosaurs lack scapulocoracoid fusion.

A little pterosaur referred to Eudimorphodon, BsP 1994 has a fused s/c. Arthurdactylus a much larger, longer winged ornithocheirid, does nto fuse the s/c. So size is not the issue.

All known pterosaur embryos come from clades that do not fuse the scapulocoracoid. However, the  juvenile Pteranodon has a fused s/c.

Addendum
Once a clade began to fuse the s/c, then lack of fusion generally accompanied phylogenetic size reductions. Among azhdarchids, only Quetzalcoatlus fuses the s/c. This includes a smaller Pteranodon YPM2525 which may also represent a size reduction shown here.

Among the derived ornithocheirds you do get a more randomized on-off-on-off pattern.

So there you have it. All results subject to change with injections of new valid data.

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.

References
Maisano JA 2002. The potential utility of postnatal skeletal developmental patterns in squamate phylogenetics. Journal of Vertebrate Paleontology 22:82A.
Maisano JA 2002. Terminal fusions of skeletal elements as indicators of maturity in squamates. Journal of Vertebrae Paleontology 22: 268–275.

DNA vs. Morphology in Reptiles and Living Things

I’m always harping on taxon inclusion and the importance of large taxon lists –especially when you don’t have large studies to base your more focused studies on.

I recently came across the gigantic family tree (Fig. 1) of living things based on DNA by David Hillis. There are so many taxa that the individual listings are not visible unless the PDF is enlarged to about 54 inches in width. Carl Zimmer of Discover Magazine interviewed Hillis here.

Figure 1. The Dave Hillis tree of all living things.

Here’s the tree pruned to the vertebrates and colorized for simplification. Yes, its only a few degrees on the circle.

Figure 2. Segment from the David Hillis tree of life showing vertebrates. Here the mammals (ant their ancestors) were the first amniotes to branch off, confirming traditional thinking and countering the large reptile tree which shows mammals on the lineage toward archosaurs.

The Hillis Tree is both more inclusive (with bacteria, fungi and plants) and less inclusive (no extinct taxa, very few living vertebrates). Even so, the resulting tree echoes traditional vertebrate trees in finding mammals (synapsids) as the first of the living clades to branch off from the other amniotes. This is different from the large reptile tree which shows the basal diapsids that gave rise to archosaurs were derived from the basal synapsids that gave rise to mammals.

So, it’s morphology vs. DNA.
Most of the time morphology agrees with DNA in the large reptile tree and the Hillis tree. Only at one point are they distinct. This is troublesome as with living taxa DNA rules, but with extinct taxa morphology rules (no DNA).

The Origin of Archosaurs and/or Archosauriformes
The literature includes papers on the origin of Archosaurs and Archosauriformes, usually originating with Proterosuchus and Archosaurus. Fewer papers look deeper into time, finding the outgroup to these taxa in protorosaurs. Fewer still look yet deeper into time, finding the outgroup to protorosaurs in Younginiformes. Wiki reports, “Younginiformes (including Acerosodontosaurus, Hovasaurus, Kenyasaurus, Tangasaurus, Thadeosaurus, Youngina, et alia sensu Currie and other researchers in the 1980s) is probably not a clade. It appears to represent a grade of South African Permo-Triassic diapsids that are not more closely related to each other as a whole than they are to other reptiles.” Indeed these taxa do form a grade in the large reptile tree.

Palaeos.com reported, “the in-group relationships of “Younginiformes,” as well as their monophyletic status, are neither understood nor have they been tested in a modern phylogenetic framework.”

So, at this point on the reptile tree, things are getting murky in the literature. That’s why the large reptile tree exists, to test previously untested relationships and establish new topologies.

DNA has also linked turtles to archosaurs, and there’s no morphological correlate there. At least no one has announced one.

Figure 3. Click to enlarge. Platypus skeleton at Melbourne Museum. Photo credit: w:User:Pengo from Wikipedia.

Then there’s the Platypus (Ornithorhychus anatinus)
The Guardian reported on a recent DNA study of an egg-laying mammal “While humans have two sex chromosomes, the X and Y, the platypus has 10, with five of each kind.” The new study, published in Nature (2008), shows the platypus as both evolutionary relic and pioneer. Chris Ponting, at the Medical Research Council’s functional genomics unit at Oxford University, said scientists had had the first chance to see if the platypus’s weird appearance was reflected in its DNA: “Lo and behold, we saw genes like those in lizards and birds, as well as some like those in other mammals. It has retained many genes other mammals lost from a time when all mammals looked much like lizards.”

Hedges 1993
Hedges (1993) reported, “This classical phylogeny of amniotes has been challenged
by recent morphological studies of living forms. Traits such as a single aortic trunk, folded cerebellum, scroll-like turbinals, loop of Henle (kidney), adventitious cartilage, and endothermy are found only in birds and mammals and have been proposed as evidence for a close relationship. Analyses of fossil and recent morphological data indicated that support for a bird-crocodilian relationship rests primarily on the fossil data, and specifically with some mammal-like reptile fossils that place mammals at the base of the amniote tree. Molecular sequence data from three genes (myoglobin, (3hemoglobin, 18S rRNA) have supported a bird-mammal grouping, but sequence data from several other genes have not. These conflicting results have created uncertainty about our ability to resolve amniote phylogeny.” Ultimately Hedges recovered a vertebrate tree echoing the Hillis tree (Fig. 1), which echoes traditional thinking based on small prior studies.

So why don’t birds and crocs nest with mammals in DNA tests?
I’m stymied. Perhaps the platypus can tell us why. It’s venom DNA is most like that of venomous snakes, yet is clearly not related. Perhaps mammals are just that different and have been since the days of the basal therapsids. This is the prime mystery initiated by, rather than solved by the larger reptile tree.

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.

References
Hillis DNA Tree of Living Things

Testing Hill 2005

N.  Brocklehurst wrote,
“I think your repeated assertion that palaeontologists don’t test the relationships you suggest is a bit…well its just not true. Admittedly some havn’t been tested e.g. Tetraceratops (never tested outside synapsids), but a great many have. As just one example Hill (2005) uses a (mostly) genus-level taxon list which covers 80 taxa from almost all the major groups in Amniota with charater list almost 3 times the size of yours to show that Caseids do not go with Milleretids and Bolosaurids, Rhynchosaurs do not go with Rhynchocephalians, Ophiacodontids are not the sister to Therapsids, Synapsids do not go with Archosaurs, Captorhinids do not go with Lepidosaurs, Mesosaurids do not go with marine reptiles…I could go on.”

This is a topic worthy of a post. Coincidentally and several years ago I had studied Hill (2005) and submitted a manuscript describing its faults. It was rejected.

What Hill (2005) was looking for and how he did it
Hill (2005) sought to determine the phylogenetic position of turtles within the Amniota by increasing taxonomic sampling and including integumentary characters, like scutes on glyptodons and sauropods. Strange, mixing such taxa, only because they had scutes. But it was published, so hats off to Hill.

As in all supermatrices, no effort was made by Hill to cull the data or study the taxa. The data was presented ‘as is.’ Hill (2005) reported, “The morphological data set assembled here represents the largest yet compiled for Amniota.” He concluded with, “Turtles are here resolved as the sister taxon to a monophyletic Lepidosauria (squamates + Sphenodon), a novel phylogenetic position that nevertheless is consistent with recent molecular and morphological studies that have hypothesized diapsid affinities for this clade.”

I tested Hill (2005) back in 2006 (long before ReptileEvolution.com) in a three-step process.

1. Taking Hill (2005) as is. 
Hill 2005 created a supermatrix by combining published data and new data based on osteology and histology of the integument. Some strange pairings resulted (Fig. 1). Bulky Diadectomorpha nested as sisters to lithe marine Mesosauridae and as sister taxa to the Synapsida. None of these look very much alike. Round-faced Acleistorhinus nested with flat-faced Lanthanosuchus. Short-faced Trilophosaurus nested with long-faced Choristodera (Champsosaurus). The so-called ‘rib’ gliders nested with marine Sauropterygia. Turtles nest with Sphenodon and both are the sister taxa to Archosauria (dinos + crocs + parasuchia + aetosaurs). Parasuchians nest within aetosaurs. Aetosaurs nest within crocodylomorphs, derived from Protosuchus. Other than these misfits, the rest ain’t too bad. And we can’t blame Hill for this because, true to the method, he was just pulling together published trees (but without casting a critical eye on the data).

To one of Neil Brocklehurst points, note the sphenacodonts and basal therapsids are suprageneric in Hill (2005), so there was the opportunity for some cherry-picking of traits and key taxa. Stenocybus. a key taxon, was not included.

Figure 1. Hill 2005. See text for details. Suprageneric taxa are marked by black squares.

2. Hill 2005 revision #1
A thorough examination of Hill’s data matrix revealed that hundreds of blank matrix boxes could be scored. Hundreds of others could be more accurately rescored, sometimes with additional character states to more accurately reflect characters. Since this was a supertree compilation, such a critical eye was not part of Hill’s process or method. I don’t like to let things slide.

Taxa that were difficult to access and contributed to excess polytomies using Hill’s scoring, such as Coahomasuchus, the titanosaur sauropods, Glyptosaurinae, Akanthosuchus, Goniopholis, Simosuchus and Mahajangasuchus were deleted. None of these taxa are basal to their respective clades.

Characters that were difficult to determine (foramina, braincase, notochordal opening) were left as Hill’s predecessors had scored them.

The resulting cladogram (Fig. 2) shows more appropriate tree topology with most clades in a more reasonable (more parsimonious) order (sister taxa look more alike overall and in detail), but pareiasaurs and turtles still nest here between lizards and Crurotarsi, which appears untenable.

Figure 2. Hill (2005) revised by the addition of more character scores. Note the topology changes.

3. Hill 2005 revision #2
The addition of a just few taxa (in red) to Hill (2005) revision #1 (Fig. 3) recovers a tree topology very much like the large reptile tree, including the major dichotomy at the base of the Reptilia (a hypothesis totally unknown to Hill in 2005). This underscores the importance of a wide gamut in a taxon list when exploring untested relationships. Here turtles nest with pareiasaurs, Procolophon and other lepidosauromorphs. Casea nests with Millerettidae, far from the Synapsida. Kuehneosaurus nests with similarly-shaped arboreal Lepidosauriformes. Synapsids and mesosaurs nest with sauropterygians and archosauriforms.

Figure 3. Adding taxa to the revision of Hill (2005). Still not perfect, but a lot better.

It’s worthy to note that
the large reptile tree does not include any glyptodonts, derived crocodylomorphs or very many ornamented lizards. Instead the large reptile tree used more basal taxa to establish a wider gamut of relationships, leaving the above-mentioned highly derived taxa for other more focused studies.

It is also worthy to note that
even with so few taxa, and largely using Hill’s characters, the reptile tree dichotomy was recovered.

To Neil’s point about the Hill (2005) 3x larger character list
Once again: It’s the taxon list (not the number of characters) that needs to expand to figure out the amniote tree topology. As an example, see what just a few extra taxa can do to a tree? (Fig. 3). Any number of characters over 150 tends to flatten out the results from 95% consistency to 98% to 98.5% to 99.1%, never quite reaching, but very closely approaching 100%. On the other hand, every additional taxon provides an additional opportunity for any already included taxon to find a more parimonious partner somewhere on the tree. The larger the list, the better.

Assertion?
No. I tested that sucker. This is evidence.

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.

References
Hill RV 2005. Integration of Morphological Data Sets for Phylogenetic Analysis of Amniota: The Importance of Integumentary Characters and Increased Taxonomic Sampling. Systematic Biology 54(4):530–547.

Eosinopteryx – part 1 – Feathers, but no flapping

Eosinopteryx brevipenna (Godefroit et al. 2013, Middle-Late Jurassic, Tiaojishan Formation) is represented by a new complete skeleton. It was a feathered theropod dinosaur about 30 cm long. The forelimb feathers were quite long (Fig. 1), but the tail feathers were not.

Paravian? or Preavian?
We’ve been looking for a feathered theropod without elongated coracoids to precede Archaeopteryx. We also need this taxon to be not pre-oviraptorid or pre-alvarezsaurid. The authors argue, with a very extensive phylogenetic analysis, that this is a troodontid resembling Anchiornis, with less extensive feathers on the hind limbs and tail. Anchiornis greatly resembled Archaeopteryx and is, therefore, closely related. Of that, there is no doubt.

Why There is Doubt
I have not created a competing analysis. Checking out Greg Paul’s figure of Anchiornis (Paul 2010), I note his Anchiornis has the short torso and elongated coracoid also seen in Archaeopteryx, troodontids and deinonychosaurs.

Figure 1. Click to enlarge. Eosinopteryx reconstructed in lateral view. Soft tissue impressions preserved on the fossil are represented here in gray. Note the small size of the coracoid (yellow) and its curved lower rim, which indicates this specimen was a pre-flapping dinosaur. Pedal digit 2 was not modified as a “killing” claw. Elements figured with DGS.

What sets Eosinopteryx apart from these?
A short coracoid with a broad curved ventral rim - Therefore Eosinopteryx did not flap and was not descended from flappers. We haven’t seen a terrestrial descendant of Archaeopteryx yet without elongated coracoids. For more on this, compare Huaxiagnathus (with its short coracoid) to Velociraptor, (with its long, tall coracoid). Otherwise these two greatly resemble one another, with the former lacking sternal plates, a retroverted pubis and caudal rods. These traits are also lacking in Eosinopteryx.

A relatively smaller skull – Much smaller than in Anchiornis.

A relatively longer torso – Much longer than in Anchiornis.

A relatively shorter pubis – Much shorter than in Anchiornis.

All these traits are primitive for theropods.

Unfortunately, 
Huaxiagnathus was not included in the analysis of Godefroit et al. (2013). Neither were oviraptorids or alvarezsaurids. Eosinopteryx needs to be compared to these missing basal taxa along with the other taxa they previously tested. Once that’s done, let’s see if the topology of the tree doesn’t shift Eosinopteryx down below (more primitive than) Archaeopteryx. 

Addendum: The analysis of Godefroit et al. (2013) was based on and provided only a segment of an earlier analysis that DID include these more primitive taxa. Thus my doubt is reduced somewhat as all pertinent taxa were included.  Even so, I wonder why these two “sisters” don’t look more alike.

If anyone has details on why Godefroit et al. 2013 said the “bone structure would have limited its ability to flap its wings,” I’d like to see it.

Interesting that this birdy topic just came up a few days ago with Mahakala. Reminds me to be careful what I wish for.

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.

References
Godefroit P, Demuynck H, Dyke G, Hu D, Escuillié FO and Claeys P. 2013. Reduced plumage and flight ability of a new Jurassic paravian theropod from China. Nature Communications 4: 1394. doi:10.1038/ncomms2389
Paul GS 2010. The Princeton Field Guide to Dinosaurs. Princeton University Press 320 pp.

wiki/Eosinopteryx

Same or Different? When Should You Invent a new Genus? or Just Add a Species? Or Revise the Whole Clade?

Now that new pterosaurs are being added to the large pterosaur tree on a fairly constant basis, it’s time to figure out what to name them.

In the old days everything was named “Pterodactylus,” no matter what it was. Sharp-eyed observers soon figured out that there were differences that set certain specimens apart and these were then renamed. Others have not yet been widely recognized as distinct, but they need to be (Fig. 1).

Figure 1. Click to enlarge. The Pterodactylus lineage and mislabeled specimens formerly attributed to this “wastebasket” genus

Nowadays, new pterosaurs distinct from all others are being given new generic names, and that’s a good thing. However some of the new specimens nest within long lists of other genera. Others are given new species names within certain genera without nesting near those genera. The problem is a result of the incompleteness of all previously published pterosaur trees. They simply do not include enough taxa. They have a priori deleted all tiny specimens and all congeneric variations that, in the large pterosaur tree, provide clues to the evolution of more derived variations, some of which are distinct genera, as in the Campylognathoides/Rhamphorhynchus transition.

Some examples
MPUM6009 was considered a Eudimorphodon and a Carniadactylus despite nesting far from both genera. MCSNB 8950 was considered a Eudimorphodon, but nested with anurognathids.

Nesodactylus nested within the genus Campylognathoides. Bellubrunnus and Qinglongopterus nested within the genus Rhamphorhynchus.

Fenghuangopterus, Sericipterus and Cacibupteryx nested within the genus Dorygnathus.

Eosipterus and Cuspicephalus nest within the genus Germanodactylus.

Kellner (2010) renamed one Pteranodon, Dawndraco, but it remains surrounded by other Pteranodon specimens.

The question is, do we revise all the old genera and give them new names now that we know how distant some were from each other? Or do we retain those genera and take away the new generic names of the new specimens between them to reflect their traditional generic nesting?

Now all this doesn’t take into account marginal generic names, like Ningchengopterus at the base of the Pterodactylus clade or Muzquizopteryx at the base of the Nyctosaurus clade. These names are likely to be valid because they are distinct genera, but so are many of the species within Pterodactylus and Nyctosaurus. If they were modern birds, not prehistoric pterosaurs, their differences would be recognized.

Part of the historical problem, of course, goes back to Chris Bennett and others who considered smaller species to be immature forms of larger species without adequately describing them or placing them in analysis. It turns out that the vast majority of those where simply smaller forms that were evolving to become the larger forms – or vice versa.

It’s a problem. It needs to be recognized and dealt with. But it will only be recognized if pterosaur specimens are not a priori deleted from analysis for whatever reason.

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.

References
Kellner AWA 2010. Comments on the Pteranodontidae (Pterosauria, Pterodactyloidea) with the description of two new species. Anais da Academia Brasileira de Ciências 82(4): 1063-1084.

A Bremer Test for the Large Reptile Tree

Neil Brocklehurst was kind enough to take the data from the large reptile tree and add Bremer Support scores (Fig. 1, click to enlarge it, if that’s not enough find it here.)

Figure 1. Click to enlarge. Bremer Test of the large reptile tree as prepared by Neil Brocklehurst. I exchanged colors for his numbers.

Every node scored at least one because the tree is fully resolved. Not every published tree can say that. Only 185 traits are used here. That has been criticized for being too few, but several earlier tests and the current results indicate it is sufficient for this taxon list. Many more could be used to more distinctly separate and lump certain clades together (sails on pelycosaurs and shells on turtles were omitted, for instance.)

What high scores mean
High Bremer Support values in this context generally mean that two taxa are sufficiently alike to nest together and yet quite distinct from the outgroup sister. In other words, it takes many additional steps for one of the ingroup taxa to nest with the outgroup one. The two ingroup taxa are much more alike than the third.

What low scores mean
Low Bremer Support values generally mean that two taxa are sufficiently alike to nest together and yet together are just barely distinct from the outgroup taxon. It takes only one or a few additional steps for one of the ingroup taxa to nest with the outgroup taxon. All three are more like each other within the given character list.

High scores generally are regarded as sought after strong clades. High scores can occur when you have two closely related outliers, as in Chroniosaurus and Chroniosuchus or as in Acleistorhinus and Eunotosaurus that are distinct from their proximal outgroups.

In paleontology low scores can occur among less complete sisters, perhaps where one is a skull and one is the post-cranial material only. Sometimes just a few bones are known and this plays havoc with Bremer scores. Low scores can also occur in more complete evolutionary sequences in which all of the taxa are closely related, changing little from taxon to taxon, as in the theropod dinosaurs or the rauisuchians.

Raising Bremer Scores
Of course, in more focused studies we can apply more focused character traits to a more focused taxon list. Here, in the large reptile tree, every character is rather generalized so as to apply or not apply to a large subset. This strategy also contributes to the high homoplasy index.

You can read more about maximum parsimony here. Wikipedia describes Bremer support in this way, “This is simply the difference in number of steps between the score of the MPT(s), and the score of the most parsimonious tree that does not contain a particular clade (node, branch). It can be thought of as the number of steps you have to add to lose that clade; implicitly, it is meant to suggest how great the error in the estimate of the score of the MPT must be for the clade to no longer be supported by the analysis, although this is not necessarily what it does. Decay index values are often fairly low (one or two steps being typical), but they often appear to be proportional to bootstrap percentages.”

Collapsing nodes with low scores
You can collapse a tree step-by-step by collapsing low score nodes first. Not sure what this gives you except an unfocused picture (with lost resolution).

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data

Partitioning – part 2 – post-cranial characters only

Bennett (2012) discussed the value of partitioning a family tree into discrete parts (skull, limbs, post-crania, etc.) to help understand the the tree as a whole. Earlier I deleted all the post-cranial traits and recovered the same tree topology as without deletions. Here I delete all the cranial traits from the large reptile tree. I also delete any taxa for which only the skull iss known (otherwise there would be no scores at all for such taxa) and I delete taxa without limbs (snakes, etc.) for the same reason.

The results (Fig. 1) do not closely mirror the original large reptile tree. Only some (mostly derived) clades are retained. The base of this tree is unresolved, including a large number of unrelated taxa including mixing some amniotes with non-amniotes. Certainly more characters further describing various aspects of the post-crania would have resolved more of this tree. This makes sense as there is always a lack of diversity at the base of any clade.

Here, strangely, the arboreal drepanosaurids become sisters to the unrelated Triassic ‘rib’ gliders and together they are sisters to the marine Enaliosauria, sans Claudiosaurus, which nests correctly with other basal diapsids.

Here, strangely, turtles (Odontochelys and Proganochelys) and pareiasaurs nest with the burrowing lizard taxon, Bipes.

Here, hearkening back to tradition, the rhynchosaurs and trilophosaurs nest with basal archosauriforms and prolacertiforms.

Figure 1. The large reptile tree partitioned into post-cranial traits only. Light orange areas = new Archosauromorpha. Lt. green = new Lepidosauromorpha. Blue = non-reptiles. Greyed areas are shifts from the large reptile tree placements.

Pterosaurs still nest with basal fenestrasaurs and tritosaur lizards. Poposaurs still nest with dinosaurs. Vancleavea still nests with thalattosaurs. Pararchosauriforms are still distinct from Euarchosauriforms.

These results conclude our look at partitioning with the present character set. Smaller subsets (limbs, vertebrae, etc.) would do nothing to improve resolution.

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.

References
Bennett SC 2012. The phylogenetic position of the Pterosauria within the Archosauromorpha re-examined. Historical Biology. iFirst article, 2012, 1–19.

Partitioning the Large Reptile Tree – part 1 – cranial characters only

Short and sweet today.

Bennett (2012) discussed the value of partitioning a family tree into discrete parts (skull, limbs, post-crania, etc.) to help understand the tree as a whole. Here I deleted all the post-cranial traits from the large reptile tree. I also deleted any taxa for which the skull was unknown (otherwise there would be no scores at all for such taxa).

The results (Fig. 1) closely mirror the original large reptile tree.

Figure 1. The large reptile tree, all post-cranial traits deleted. A few taxa without skulls are also deleted. This is one demonstration of partitioning. Tomorrow comes another. Green areas mark loss of resolution. Gray = Reptilia. Orange = new Lepidosauromorpha. Yellow = new Archosauromorpha.

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.

References
Bennett SC 2012. The phylogenetic position of the Pterosauria within the Archosauromorpha re-examined. Historical Biology. iFirst article, 2012, 1–19.

re: Adding Taxa

A new phylogenetic study by Wiens and Tiu (2012) concluded: “We show that adding taxa that are highly incomplete can improve phylogenetic accuracy in cases where analyses are misled by limited taxon sampling. These surprising empirical results confirm those from simulations, and show that the benefits of adding taxa may be obtained with unexpectedly small amounts of data. These findings have important implications for the debate on sampling taxa versus characters, and for studies attempting to resolve difficult phylogenetic problems.”

Nice to hear. I agree.
All three of my trees (reptiles, pterosaurs and basal therapsids) include incomplete taxa. Some with skulls. Some without. Some just bits and pieces. Most, however, are complete. It’s important to include odd and incomplete taxa, just to know where they belong. Problems come when skull only taxa nest with skull-less taxa. Even so, that can be overcome with complete sisters.

Interestingly:
The Wiens and Tiu (2012) study found turtles closer to archosaurs than lizards with mammals the outgroup. Not sure why turtle DNA is closer to archosaurs when their morphology says otherwise. And I wonder why mammal DNA doesn’t nest closer to archosaurs when their morphology says otherwise. Bottom line: the morphology has to support the DNA and vice versa. Since prehistoric taxa will never give us DNA, we’re stuck with morphology, so long as a sufficient number of taxa are included. The more the better. 500+ taxa is not a bad start.

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.

References
Wiens JJ, Tiu J 2012. Highly Incomplete Taxa Can Rescue Phylogenetic Analyses from the Negative Impacts of Limited Taxon Sampling. PLoS ONE 7(8): e42925. doi:10.1371/journal.pone.0042925

Characters: More or Less

The large reptile tree has been criticized for having too few characters, 228, to adequately work with 305 taxa. Despite these objections one tree was recovered (full resolution).

To test the supposition that the ratio of characters to taxa should change the tree topology I undertook the following exercise which can be repeated by anyone using the present matrix, available by request.

1. The number of taxa was reduced (Fig. 1) by excluding all taxa more primitive than Thadeosaurus and all of the new Lepidosauromorpha (the other half of the entire Reptilia). That reduced the number of taxa from 305 to 85 and effectively increased the ratio of characters to taxa by 268%. To no one’s surprise, a single tree was recovered from this segment with no change in tree topology.

2. The number of characters was reduced by 10% by deleting every 10th character. Three trees were recovered with no change in tree topology other than a minor loss of resolution.

3. The number of characters was reduced by another 10% by deleting every 10th character. 27 trees were recovered with no change in tree topology other than a minor loss of resolution.

4. The number of characters was reduced by another 10% by deleting every 10th character. 1200+ trees were recovered when I stopped the process. There was a great loss of resolution at the base of the Archosauria due to the inclusion of the very incomplete taxa Trialestes and SMNS 12352.

5. Trialestes and SMNS 12352 were deleted and the test rerun. 165 trees were recovered with no change in tree topology other than a minor loss of resolution.

6. The number of characters was reduced by another 10% (for a total of 40%) by deleting every 10th character. 1260 trees were recovered with 136 characters. The tree topology remained quite similar to the original tree (see Fig. 1) with loss of resolution at the weaker nodes.

Figure 1. Test segment from the large reptile tree showing loss of resolution but no change in tree topology after exclusion of 40% of the original 228 characters and deletion of two very incomplete taxa. Only 136 characters recovers this tree.

You be the judge
Is this still a good tree? Or would a further increase in the number of characters substantially change the tree topology? Most of the traditional clades have been recovered here (dinosaurs, protorosaurs, parasuchia, ornithischianas, etc). The only novelty occurs when certain taxa not previously included are present. That is the key to this segment of the large reptile tree. If more characters pertinent to specific leaves were included, I’ll grant you that may shift one leaf with another. If anything substantial changes, please let me know about it.

GIGO
David Marjanovic is fond of saying GIGO (Garbage In, Garbage Out). Well, the present test got rid of 40% of “the garbage” and still recovered the same tree. On the same note, with 40% of the characters deleted that should have removed many of the so-called correlated characters. The purported power of those arguments has been reduced with these tests.

The key to the success of the large reptile tree is its size, specifically the large number of taxa. More taxa provide more nesting opportunities. The large reptile tree provides nearly a full gamut of opportunities, greatly reducing the possibility of a “by default” nesting. By comparison, the ratio of characters to taxa has little effect on the topology in this series of tests.

I hope this settles the issue. If not, please send your results or make them known.

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.