Taxa closest to the human lineage in the LRT

The large reptile tree is capable of providing a list of taxa closest to the lineage of any included taxon. And it is updated all the time…

For instance,
in the lineage of humans (Homo sapiens) the following taxa are closest to that main line. Read this list with the understanding that taxa closest to the main line have often evolved traits that we infer (from phylogenetic bracketing) were not present in the actual hypothetical ancestor. The chance of finding the actual ancestors in the fossil record are vanishingly small, so we do the best we can with what specimens we have. Also note that the rare appearance of key fossils may be tens to hundreds of millions of years after their likely first appearance in this lineage. Thus the the chronological order may not match the phylogenetic order, but it does provide a ‘window’ to that first appearance.

  1. Ichthyostegabasal tetrapod  365 mya
  2. Pederpes 350 mya
  3. Proterogyrinus 322 mya
  4. Seymouria 275 mya
  5. Utegenia – also basal to frogs 300 mya
  6. Silvanerpetonproximal to the basalmost reptile 335 mya
  7. Gephyrostegus bohemicusbasalmost reptile/amniote 310 mya
  8. Eldeceeonbasalmost archosauromorph 335 mya
  9. Romeriscus 306 mya
  10. Solenodonsaurus also basal to chroniosuchids 290 mya
  11. Casineria – 335 mya
  12. Brouffia 310 mya
  13. Coelostegus  310 mya
  14. Protorothyris MCZ 1532 290 mya
  15. Protorothyris CM 8617 290 mya
  16. Protorothyris MCA 2149 290 mya
  17. Vaughnictis – last common ancestor of mammals and dinosaurs 290 mya
  18. Apsisaurus –  basalmost synapsid 295 mya
  19. Varanosaurus FMNH PR 1760 280 mya
  20. Varanosaurus BSPHM 1891 XV20 280 mya
  21. Archaeothyris 306 mya
  22. Ophiacodon 290 mya
  23. Haptodus – also basal to pelycosaurs 305 mya
  24. Stenocybus – also basal to anomodontids 295 mya
  25. Cutleria basalmost therapsid 295 mya
  26. Hipposaurusbasalmost kynodont 260 mya
  27. Ictidorhinus 260 mya
  28. Biarmosuchus 260 mya
  29. Eotitanosuchus 260 mya
  30. Lycosuchus 260 mya
  31. Procynosuchusbasalmost cynodont 250 mya
  32. Thrinaxodon 245 mya
  33. Probainognathus 230 mya
  34. Haldanodon 145 mya
  35. Liaoconodon 120 mya
  36. Pachygenelus 195 mya
  37. Sinoconodonbasalmost mammal, also basal to living monotremes 195 mya
  38. Megazostrodon 200 mya
  39. Juramaia 160 mya
  40. Cronopio 98 mya
  41. Didelphisbasalmost metatherian extant
  42. Thylacinus – basal to many living marsupials recently extinct
  43. Monodelphisbasalmost eutherian extant
  44. Eomaia 125 mya
  45. Nandinia – also basal to carnivores extant
  46. Ptilocercus – basalmost primate/dermpteran/bat extant
  47. Notharctus – also basal to lemurs 54 mya
  48. Aegyptopithecus* 33 mya
  49. Proconsulbasalmost anthropoid 18 mya
  50. Ardipithecus* basalmost hominid 5 mya
  51. Australopithecus* 3 mya
  52. Homo sapiens extant

* not yet listed in the LRT, but documented at ReptileEvolution.com

With the recent addition of certain stem mammals,
like Haldanodon and Liaconodon, this list expands upon and refines the list that first appeared in Peters 1991. Each of the names links to further information. There is also a video that includes most of these taxa here on YouTube.com.

References
Peters D 1991. From the Beginning, the Story of Human Evolution. online PDF.

Carroll 1988

Figure 1. Vertebrate Paleontology by RL Carroll 1988.

Figure 1. Vertebrate Paleontology by RL Carroll 1988 is one of the starting points for this blog and ReptileEvolution.com

In 1988
Dr. Robert L. Carroll published a large work devoted to the study of fish and tetrapods: Vertebrate Paleontology and Evolution. Between its black covers and silver dust jacket there was – and is – an immense amount of data on just about every taxon known at the time… a time just before software driven phylogenetic analysis became de rigueur.

My copy
has been used so much it has a broken binder, which makes every section lighter, easier for scanning.

For its time, and for a few decades later
Vertebrate Paleontology and Evolution was the ‘go-to’ textbook for students and artists of this science. (See below).

A few quotes from the Amazon.com website:

  1. “This book was my textbook for Vertebrate Paleontology and Evolution at the University of Rochester back in 1992.”
  2. “the only easily available work that goes to any depth on this intensely interesting subject.”
  3. “The book is very daunting to look at if you just flip through it. However, it does a nice job of introducing concepts and terms to the reader. Its organization is straightforward, starting with the simplest vertebrates and eventually finishing with mammals.:
  4. “Just realize that some of the information may not reflect our current understanding since the book is over 10 years old and many new finds have come to light, new ideas have been introduced, and old ideas reexamined.”
  5. “It’s an essential for anyone building a library of paleo textbooks.”
  6. “I’m a working fossil preparator and this is the primary reference text used in paleontology labs at the American Museum of Natural History, Yale Peabody Museum and others I’m sure.”
  7. “I had Romer’s Vertebrate Paleontology, which is an excellent book, until a paleontologist friend directed me to Carroll’s book. He acknowledges Romer’s work in the field but this is an updated version (for the time of publication).”
  8. “If you want to chart the course of evolution up to the present – read this book!”

Carroll 1988 updated
Romer’s Vertebrate Paleontology  (1933, 1945, 1966) which was the ‘go-to’ textbook of its day.

ReptileEvolution.com
updates portions of Carroll 1988. Likewise and in due course, someone someday may want to update ReptileEvolution.com. I hope they do so.

Every so often
it’s good to give credit to one’s mentors and resources. Sometimes you learn by doing. Other times you learn by reading. I suppose everyone who writes such a large gamut book knows he/she is doing something to help future students and enthusiasts who they will never meet. I feel the same way, butI imagine both Carroll and Romer were additionally warmed by a healthy royalty check once or twice a year.

References
Carroll RL 1988. Vertebrate Paleontology and Evolution. W. H. Freeman and Co. New York.
Romer AS 1966. Vertebrate Paleontology. University of Chicago Press, Chicago; 3rd edition

wiki/Vertebrate_Paleontology_and_Evolution

Zhenyuanlong: Dromaeosaur? No. Tyrannosaur with wings? Yes.

Lü and Brusatte 2015
described a short-armed, winged Early Cretaceous Liaoning theropod, Zhenyuanlong suni (Fig. 1, JPM-0008 Jinzhou Paleontological Museum), as a dromaeosaur. Their published phylogenetic analysis included only dromaeosaurs but their text indicates a large inclusion set.

Figure 1. Zhenyuanlong in situ with colors applied to bones and feathers. These colors are transferred to create the reconstruction in figure 3.

Figure 1. Zhenyuanlong in situ with colors applied to bones and feathers. These colors are transferred to create the reconstruction in figure 3. The pelvic elements are reconstructed at right. The manus and pes are reconstructed at left.  Scale bars are 10cm.

From the Lü and Brusatte text
“We included Zhenyuanlong in the phylogenetic dataset of Han et al., based on the earlier analysis of Turner et al, which is one of the latest versions of the Theropod Working Group dataset. This analysis includes 116 taxa (two outgroups, 114 ingroup coelurosaurs) scored for 474 active phenotypic characters. Following Han et al., characters 6, 50, and 52 in the full dataset were excluded, 50 multistates were treated as ordered, and Unenlagia was included as a single genus-level OTU. The analysis was conducted in TNT v1.142 with Allosaurus as the outgroup.”

I reconstructed this theropod,
from published photographs (Figs. 1, 2) using (DGS digital graphic segregation), added it to the large reptile tree and found that it nested between tiny Compsognathus and gigantic Tyrannosaurus rex. Of course, Zhenyuanlong had the opportunity to nest with several dromaeosaurs, but it did not do so.

Figure 2. Skull of Zhenyuanlong in situ, as originally traced, colorized with skull, palate and mandible segregated.

Figure 2. Skull of Zhenyuanlong in situ, as originally traced, colorized with skull, palate and mandible segregated. Original quadrate may be a quadratojugal.

When you look at the reconstruction,
(Fig. 3) the similarity to T. rex becomes immediately apparent… except for those long feathered wings, of course.

I’ll run through several of the traits that link
Zhenyuanlong to Tyrannosaurus to the exclusion of dromaeosaurs here. It’s a pretty long list. Even so, if you see any traits that should not be listed, let me know and why.

  1. skull not < cervical series length
  2. skull not < half the presacral length
  3. premaxilla oriented up
  4. lacrimal not deeper than maxilla
  5. naris dorsolateral
  6. naris at snout tip, not displaced dorsally
  7. orbit length < postorbital skull
  8. orbit not > antorbital fenestra
  9. orbit no > lateral temporal fenestra
  10. orbit taller than wide
  11. frontal with posterior processes
  12. posterior parietal inverted ‘B’ shape
  13. jugal posterior process not < anterior
  14. parietal strongly constricted
  15. quadratojugal right angle
  16. majority of quadrate covered by qj and sq
  17. postorbital extends to minimum parietal rim
  18. maxillary teeth at least 2x longer than wide
  19. mandible tip rises
  20. angular not a third of mandible depth
  21. retroarticular process expands dorsally and ventrally
  22. cervicals taller than long
  23. cervicals decrease cranially
  24. mid cervical length < mid dorsal
  25. caudal transverse processes present beyond the 8th caudal
  26. humerus/femur ratio < 0.55
  27. metacarpals 2 & 3 do not align with manual one joints
  28. pubis angles ventrally – not posteriorly
  29. 4th trochanter of femur sharp
  30. metatarsals 2 & 3 align with p1.1
Figure 3. Zhenyuanlong reconstructed in lateral view. Something behind the pelvis could be the remains of an egg, but needs further study. Both sets of wing feathers are superimposed here. Click to enlarge.

Figure 3. Zhenyuanlong reconstructed in lateral view. Something behind the pelvis could be the remains of an egg, but needs further study. Both sets of wing feathers are superimposed here. Click to enlarge. Note the pubis is not oriented posteriorly. Note the longer legs of Zhenyuanlong compared to tested dromaeosaurs.

Shifting
Zhenyuanlong to the dromaeosaurs adds a minimum of 127 steps to the large reptile tree. There is one clade of theropods that nests between the current tyrannosaur and dromaeosaur clades.

Figure 3. Cladogram subset of the large reptile tree showing the strong nesting of Zhenyuanlong as the current sister to Tyrannosaurus. Obviously many more theropod taxa are missing here. They have not been tested yet.

Figure 4. Cladogram subset of the large reptile tree showing the strong nesting of Zhenyuanlong as the current sister to Tyrannosaurus. Obviously many more theropod taxa are missing here. They have not been tested yet.

Note
I have not tested as many theropods as there are in several theropod cladograms.

The possible faults with the Lü and Brusatte study were

  1. a lack of reconstructions to work with, rather than just a score sheet that others had created and they trusted. Reconstructions test identifications by making sure the puzzle pieces actually fit, both morphologically and cladisitically.
  2. I think they were fooled by the apparent posterior orientation of the pubis in situ when in vivo it was not oriented posteriorly
  3. I’m guessing that the traits they used could be used on in situ fossils without making reconstructions. The traits I use require reconstructions.

With this nesting
the origin of long pennaceous wing feathers is evidently more primitive than earlier considered, developed twice. And perhaps this is why T. rex had such tiny arms. They were former wings, not grasping appendages.

References
Lü J and Brusatte SL 2015. A large, short-armed, winged dromaeosaurid (Dinosauria: Theropoda) from the Early Cretaceous of China and its implications for feather evolution. Scientific Reports 5, 11775; doi: 10.1038/srep11775.

Variation in Archaeopteryx and basal bird radiation

Updated October 30, 2015 with a new GIF animation that reveals the furcula of this specimen on the newly added counter-plate. 

The basal bird
Archaeopteryx lithographica  (Meyer 1861, Late Jurassic, Solnhofen Formation ~150 mya, 30-50 cm in length) is known from 12 skeletal specimens, 11 of which are published. Two of those are shown here (Fig. 1). Bennett (2008) reports, over the years workers have split these specimens into six generic and ten species names, while others have lumped them all into a single species.

Figure 1. The six tested Solnhofen birds currently named Archaeopteryx, Jurapteryx and Wellnhoferia.

Figure 1. The six tested Solnhofen birds currently named Archaeopteryx, Jurapteryx and Wellnhoferia.

In typical and traditional bird cladograms
only one Archaeopteryx is ever employed. Perhaps the subtle differences between the Solnhofen specimens are considered inconsequential in phylogenetic analyses that attempt to reveal early bird interrelationships. At least that is the tradition.

In like fashion
Solnhofen pterosaurs are also known from hundreds of specimens, but in typical pterosaur analyses only a single specimen from the most common genera, Rhamphorhynchus, Pterodactylus andScaphognathus are ever employed. At least that is the tradition.

As readers know,
I have added dozens of Solnhofen pterosaur specimens to my analysis and found that:

  1. no two tested taxa were identical (except the juvenile/adult pairing in Rhamphorhynchus)
  2. variations in genera are phylogenetic rather than ontogenetic; and
  3. those variations are the overlooked keys to understanding the interrelationships of pterosaurs in general

For instance,
from those results the widely accepted clade, “Pterodactyloidea,” was found to have not one, but four origins, all developing a complete set (rather than a partial set as in wukongopterids) of pterodactyloid-grade traits all by convergence.

So, with the presence of Galapagos-like variation in Solnhofen pterosaurs…
I wondered if there was Galapagos-like variation in the Archaeopteryx specimens. And if so,  what were those variations? Would they be substantial enough to appear in an analysis not focused on birds, like the large reptile tree? (It now includes 594 taxa.)

A little history on Archaeopteryx lumping and splitting
Houck et al (1990) found evidence in scatter plot analysis of immaturity in the six specimens then known and interpreted the specimens as a growth series of a single species.

Elzanowski (2002) rejected the notion that any of the specimens were immature and so recognized the London, Berlin and Munich specimens as three distinct species and the Solnhofen specimen (BSP 1999) as a new genus, Wellnhoferia grandis.

Senter and Robins (2003) repeated the Houck et al analysis with one added and one excluded specimen agreed with Houck et al. on a single species documenting an ontogenic series.

Mayr et al. (2007) described the Thermopolis specimen and lumped all specimens into two species.

Bennett (2008) likewise used statistical analysis in a study of Alligator to document variation a single species, concluding that “lengths of skeletal elements in a sample of a single species can have high correlation coefficients, and that such high correlation coefficients are not indicative of multi-species samples.”

This is all well and good
but where are the phylogenetic analyses? Bennett (1995) lumped all of his Rhamphorhynchus specimens together using statistics, but missed the speciation recovered in phylogenetic analysis. Even the feet show variation! Perhaps the same is true of Archaeopteryx?

I start with just two Archaeopteryx taxa,
the large London and small Eichstaett specimens (Fig.1). I added both to the large reptile tree and was mildly surprised by the unconventional results. The London specimen nested at the base of the few specimens currently tested in the Enantiornithes clade (Fig. 2). The Eichstaett specimen nested at the base of the few specimens tested in the Euornithes clade.

Figure 4. Here I add the Munich specimen of Archaeopteryx to the large reptile tree and recover it basal to the Scansoropterygidae, the clade of basal birds that shares a long finger 3.

Figure 2. Here I add the Munich specimen of Archaeopteryx to the large reptile tree and recover it basal to the Scansoropterygidae, the clade of basal birds that shares a long finger 3.

These are novel nestings
Typically other specimens nest between Archaeopteryx and Enantiornithes. The classic transitional taxa include  RahonavisXiaotingia and Confuciusornis. In the large reptile Rahonavis nests with Velociraptor, Xiaotingia (together with Eosinopteryx) is the proximal outgroup taxon for Archaeopteryx, and Confuciusornis nests as a basal euornithine,

Remember, this is small list of pertinent taxa
with far fewer pre-birds and birds included than are usually found in bird origin cladograms. Likewise, there are also far fewer theropod and bird specific characters employed here.

The key differences
between this study and prior studies are simply the inclusion of one more Archaeopteryx specimen into the matrix, the use of reconstructions, and a set of 228 generic characters that work for reptiles at large, but are not bird or theropod specific.

London specimen enantiornithine traits from the large reptile tree:

  1. snout constricted in dorsal view
  2. nasal shape parallel in dorsal view
  3. premaxilla ascending process not beyond naris
  4. nasals subequal to frontals
  5. maxilla with antorbital fossa
  6. pineal foramen/cranial fontanelle absent
  7. frontal parietal suture not straight
  8. no temporal ledge
  9. quadrate posterior not concave
  10. squamosal + quadratojugal indented, no contact
  11. jaw joint descends
  12. premaxillary teeth: > 4
  13. retroarticular angle: straight
  14. mandible ventrally: straight then convex

Eichstaett specimen euornithine traits from the large reptile tree:

  1. snout not constructed in dorsal view
  2. nasal shape, premaxilla invasion and separation
  3. premaxilla ascending process beyond naris
  4. nasals shorter than frontals
  5. maxilla without antorbital fossa
  6. pineal foramen/cranial fontanelle present
  7. frontal parietal suture straight and wider than n/f suture
  8. squamosal temporal ledge
  9. quadrate posterior concave
  10. only squamosal indented
  11. jaw joint in line with maxilla
  12. premaxillary teeth: 4 or fewer
  13. retroarticular angle: ascends
  14. mandible ventrally: straight then concave

Plus
There are several other traits that are not universal among derived taxa in both clades. These help to lump and split the derived taxa. Request the .nex file here.

Figure 2. London Archaeopteryx pectoral area with a focus on the scapula, coracoid and clavicles.

Figure 2. GIF animation London Archaeopteryx pectoral area with a focus on the scapula, coracoid and clavicles.

And then
there are several enantiornime-euornithine splitting traits not listed as traits in the large reptile tree.

Enantiornithine traits in the London specimen of Archaeopteryx:

  1. coracoid with convex articulation with scapula
  2. coracoid with convex lateral shape
  3. Y-shaped clavicles
  4. metatarsals fused proximally
Figure 2. Pectoral girdle of the Eichstaett specimen of Archaeopteryx.

Figure 2. Pectoral girdle of the Eichstaett specimen of Archaeopteryx. Two frames, each 5 seconds long.

Euornithine traits in the Eichstaett specimen of Archaeopteryx:

  1. coracoid with concave articulation with scapula
  2. coracoid with straight lateral shape
  3. clavicle not preserved
  4. metatarsals: fusion patterns not clear

As mentioned previously
this addition of one more Archaeopteryx to a phylogenetic analysis will not settle any issues. Paleontology rarely settles any issues. But hopefully others will take the time to trace the bones, create the reconstructions and add several Archaeopteryx specimens to future phylogenetic analyses. As has been demonstrated several times now, statistical analyses of Solnhofen taxa don’t reveal what phylogenetic analyses seem to.

References
Bennett SC 1995. A statistical study of Rhamphorhynchus from the Solnhofen Limestone of Germany: Year-classes of a single large species. Journal of Paleontology 69:569–580.
Bennett SC 2008. Ontogeny and ArchaeopteryxJournal of Vertebrate Paleontology 28 (2): 535-542.
Houck MA, Gauthier JA and Strauss RE 1990. Allometric scaling in the earliest fossil bird, Archaeopteryx lithographica. Science 247: 195–198.

Trees of Life: Birds and Pterosaurs

Yale’s Richard Prum recently announced that the Tree of Life of Birds is almost complete. A genomic analysis of 198 species of birds was published in the Oct. 7 edition of the journal Nature. Prum reported, ““In the next five or 10 years, we will have finished the tree of life for birds.” I presume that means fossil taxa will also be included and scored by morphological traits because genes (genomic traits) are not available.

It is not the first time…
Trees of Life for Birds were announced earlier here, here, here and here.

Having been through a similar study, I support all such efforts. AND I will never attempt to add any but a few sample birds to the large reptile tree. Others have better access to specimens and they have a big head start on the process.

Unfortunately,
some workers have ignored the pterosaur tree of life. Recently Mark Witton ignored isometric growth patterns in pterosaurs to agree with Bennett (2013) that the genus Pterodactylus includes tiny short-snouted forms, mid-sized long-snouted forms (including the holotype, of course) and large small-heron-like forms. Witton reports, “Speaking of adulthood, it was also only recently that we’ve obtained a true sense of how large Pterodactylus may have grown. We typically imagine this animal as small bodied – maybe with a 50 cm wingspan – but a newly described skull and lower jaw makes the first unambiguous case for Pterodactylus reaching at least 1 m across the wings (Bennett 2013).”

We looked at Bennett’s paper earlier in a three part series that ended here. The taxon Witton refers to is actually just a wee bit larger than the holotype and is known from a skull, so wingspread can only be guessed. The tiny short-snouted forms are actually derived from the short-snouted scaphognathids as shown here.

The Pterodactylus lineage and mislabeled specimens formerly attributed to this "wastebasket" genus

Figure 1. Click to enlarge. The Pterodactylus lineage and mislabeled specimens formerly attributed to this “wastebasket” genus. Others have split the largest specimens of Pterodactylus from the others without employing a phylogenetic analysis.

You might recall
that one of the largest complete Pterodactylus specimens (Fig. 1) recovered by the large pterosaur tree was mistakenly removed from this genus and lumped with Ardeadactylus, a basal pre-azhdarchid, all without phylogenetic analysis.

Agreeing with Bennett,
Witton deletes some taxa that actually belong to this genus, while accepting others that do not belong, all based on eyeballing specimens without a phylogenetic analysis that includes a large gamut of specimens (that does not delete the tiny forms). Eyeballing taxa is not the way to handle lumping and splitting. Phylogenetic analysis is. We looked at the Pterodactylus wastebasket problem here.

References
Bennett  SC 2012 [2013]. New information on body size and cranial display structures of Pterodactylus antiquus, with a revision of the genus. Paläontologische Zeitschrift (advance online publication) doi: 10.1007/s12542-012-0159-8
http://link.springer.com/article/10.1007/s12542-012-0159-8

Evolution basics – starring Jon Stewart and Babe Ruth

Evolution does not work in mysterious ways.
The basics (small variations leading over dozens of generations to larger changes) are simple:

GIF movie 1. Skull width as a variable demonstrated by Babe Ruth and John Stewart in this animated GIF file.

GIF movie 1. Skull width as a variable demonstrated by Babe Ruth and John Stewart in this animated GIF file.

  1. wider / narrower (skull, body, feet, etc.)
  2. taller-larger / smaller-shorter
  3. longer (more ribs) / shorter (fewer ribs)
  4. longer limbs / shorter limbs
  5. larger skull / smaller skull
  6. longer preorbital region / longer postorbital region
  7. longer neck / shorter neck
  8. sharp claws / rounded claws
  9. etc. / etc.

At left 
are extinct baseball star, Babe Ruth, and extant comedian/commentator, Jon Stewart, graphically demonstrating #1 on the above list, wider / narrower in the skull shape. Both are male members of the species Homo sapiens.

Other traits
one can add to this list include various perforations or fenestrae (which have several and often convergent origins and disappearances:

  1. fenestra between the naris and orbit (antorbital fenestra)
  2. fossa surrounding antorbital fenestra
  3. one or more fenestrae between the orbit and occiput
  4. fenestra in the mandible
  5. occipital fenestrae expand over braincase
  6. acetabulum perforated or not

And once fenestrae are formed:

  1. Loss of lower temporal arch
  2. Loss of upper temporal arch
  3. Loss of both

Then add
the size and shape of various bones and their processes compared to other bones and you have yourself a long character list. Enough of these (150+) provide a good matrix of characters and character states that can produce the menagerie of reptiles found in the large reptile tree, now numbering 566 taxa for 228 characters.

The wider / narrower and smaller / larger dichotomies 
can also be seen in the variety of specimens attributed to Proterosuchus and Chasmatosaurus (Fig. 2, Broom 1903). Some paleontologists (Welman 1998, Ezcurra  and Butler 2015) consider these taxa congeneric. They think this variety constitutes an ontogenetic series. On the other hand, the large reptile tree recovered these taxa in distinct nodes and clades. Narrower-skulled forms nest together. So do wider-skulled forms and they lead to other even more distinct taxa, including some once again tiny forms. The tall-skulled proterosuchids do not lead to more derived taxa.

Figure 3. The many faces of Proterosuchus to scale and in phylogenetic order, among with their closest known relatives. Note the phylogenetic miniaturization, reduction of the drooping premaxilla and loss of the antorbital fenestra after the TM 201 specimen of Chasmatosaurus. Click to enlarge.

Figure 2. The many faces of Proterosuchus to scale and in phylogenetic order, among with their closest known relatives. Note the phylogenetic miniaturization, reduction of the drooping premaxilla and loss of the antorbital fenestra after the TM 201 specimen of Chasmatosaurus. Click to enlarge.

The smallest taxon
shown here (Fig. 2), Youngoides romeri, leads to euparkeriids and then to a long list of archosauriforms including dinosaurs, crocs and birds. This last common ancestor of proterosuchids and euparkeriids (all archosauriforms) also had a small antorbital fenestra.

Have a great weekend!
Keep those cards and letters coming.

References
Broom R. 1903. On a new reptile (Proterosuchus fergusi) from the Karroo beds of Tarkastad, South Africa. Annals of the South African Museum 4: 159–164.
Ezcurra MD and Butler RJ 2015. Post-hatchling cranial ontogeny in the Early Triassic diapsid reptile Proterosuchus fergusi. Journal of Anatomy. Article first published online: 24 APR 2015. DOI: 10.1111/joa.12300
Welman J 1998. The taxonomy of the South African proterosuchids (Reptilia, Archosauromorpha). Journal of Vertebrate Paleontology 18 (2): 340–347.

Bird stem evolved faster than other theropods

Figure 1. The evolution of birds as a consequence of miniaturization. Artist: Davide-Bonnadonna

Figure 1. The evolution of birds as a consequence of miniaturization. Artist: Davide-Bonnadonna. I think the last taxon on the right has been enlarged somewhat for clarity. See Figure 2.

A new paper by Lee et al. (2014)
reported on their phylogenetic analysis of theropods employing 1549 characters. They found two drivers underlying the dinosaur-bird transition: 1) sustained miniaturization; and 2) the evolution of skeletal adaptations 4x faster than other dinosaurs.

They found that miniaturization facilitated the evolution of morphological novelties associated with small size: reorientation of the body mass; 2) increased aerial ability; and 3) paedomorphic skulls with enlarged eyes and brains along with a reduced snout and smaller teeth without serrations.

I heartily endorse this work.
It supports a paradigm that miniaturization that produces new clades. Cope’s Rule generally does not. We see similar miniaturization at the genesis of reptiles, amphibians, therapsids, mammals, dinosaurs, crocodylomorphs (together the archosaurs), pterosaurs, diapsids and several clades within each of these, like bats and pterodactyloid-grade pterosaurs. We knew for a long time that sustained miniaturization also produced birds. So that’s not news. It just has never been so well laid out before.

Lee et al. conclude: 
“Because size reduction, feather elaboration, paedomorphism, and other anatomical novelties permitted by small size all evolved in concert along the bird stem, identifying the primary driver of this sustained trend is probably impossible. It is likely that all traits influenced and provided the context for the evolution of others.” 

Actually not so impossible.
The authors make no mention of the fact that smaller taxa generally mature more quickly, breed more often and die sooner. In other words, generational turnover happens more quickly in smaller taxa, as everyone know. The rate of evolution over time is accelerated by the rate of reproduction over time (all other things being equal). That’s the primary driver. 

In the family tree of amniotes,
you see this all the time, not just in bird origins. Good to see this get the press it deserves.

References
Lee MSY, Cau A, Naish D and Dyke GJ 2014. Sustained miniaturization and anatomical innovation in the dinosaurian ancestors of birds.