Monday, October 2, 2017

Beginning at the same time

Paravians and the type of dinosaur they are purported to have evolved from, appear in the fossil record beginning at the same time.
The purported dinosaur ancestor is not dated earlier than the descendant. 

Scansoriopteryx (Paraves)
Temporal range: Callovian to Kimmeridgian, 165-156 Ma
Serikornis (Paraves)
Temporal range: Middle Jurassic, 165-162 Ma

These Paravians line up (165 ma) with the type of dinosaur they are purported to have evolved from. For example with:

Proceratosaurus (basal Tyrannoraptora)
Temporal range: Middle Jurassic, 165 Ma

Proceratosaurus is a genus of small-sized (~3 metres (9.8 ft) long) carnivorous theropod dinosaur from the Middle Jurassic (Bathonian) of England.[1] It was originally thought to be an ancestor of Ceratosaurus, due to the similar small crest on its snout.[2] Now, however, it is considered a coelurosaur, specifically one of the earliest known members of Tyrannosauroidea,[3] the clade of basal relatives of the tyrannosaurs.[4]
So the actual fossil evidence has paravians and the type of dinosaur they are thought to have evolved from appearing in the fossil record beginning at the same time.



For reference:
Notice the imaginary lines running from Coleurosaur to Paraves.

Friday, August 25, 2017

Mapping the wing transition
Check out Figures 16.3 and 16.4 Pages 455 - 456.

Flight feathers are not variants of thermal insulating material, but structures that have evolved over a long period, enabling the span and aspect ratio of the wings to be progressively increased as they developed.
The structure of a flight feather is dedicated wholly to withstanding the bending and torsional moments caused by aerodynamic forces, and delivering these moments through the follicle to the wing skeleton.
This shows that feathers in basal Paraves are not exaptations.
The main features of the hypothetical patagial glider can still be seen, albeit much modified, in the modern bird wing, which has a patagium that joins the side of the body to the elbow joint, and continues as a narrow strip along the posterior side of the ulna, and of the reduced hand skeleton.

Very recently Xu and Mackem (2013) introduced a digit reduction scheme called the
lateral shift hypothesis (LSH). Like the TDH, it follows the II, III, IV
identication of tetanuran digits, and values the positional criteriaover the compositional ones. Unlike the TDH, however, it argues
that a (partial) homeotic anterior shift took place and that digit IV
was completely reevolved, following in both points the FSH

Also see:

Sunday, July 23, 2017

The digits transition

I am working with the following hypothesis of the transition of fingers and phalanges:
     I   II   III   IV  V
P-2-3-4-5-x Pterosaur
x-2-3-4-x-x Basal Paraves 
x-1-2-1-x-x Modern bird

P = pteroid. Numbers represent the number of phalanges. For example, this shows 5 phalanges in the fourth finger (IV) of pterosaur. Roman numerals represent fingers (digits). 

Summary of Changes:
The first step occurred in pterosaur with the loss of the fifth finger (V).
Then in the transition to basal paraves:
The pteroid was lost. (It is a prepollex).
The fourth finger (IV) was progressively shortened and finally lost 

This transition (from pterosaur digits to the digits of a creature like Scansoriopteryx) is consistent with all the known evidence. It is not contrary to any known evidence.

The set of changes I am proposing is similar to that descibed in the hexadactyl origin hypothesis (HOH).

The early pantetrapod Devonian ancestors of
birds were polydactylous, with up to eight fingers
on both forelimbs and hindlimbs (Clack, 2002). In
the embryos of several modern tetrapods, it is
possible to observe the presence of mesenchymal
condensations other than those of digits 1-5 that
may represent vestiges of the additional digits of
early pan-tetrapods. A mesenchymal condensation
in front of digit 1 is called a prepollex
, and a
mesenchymal condensation found posterior to
digit 5 is called a postminimus.
Although only three digits develop in the wing of
the chicken, recent work on the expression of the
Sox9 gene allowing the visualization of mesenchymal
condensations has revealed that six mesenchymal
condensations can be found in the
developing wing that can be compared to digital
condensations (Welten et al., 2005). Remarkably,
it is possible to interpret these condensations as
consistent with the proposal that the anteriormost
condensation may be a prepollex the following
three digital condensations are digits 1, 2 and 3,
the next condensation is a vestige of digit 4, and
the posteriormost condensation (named ''element
X'' by Welten et al., 2005) can be assumed to be a
vestige of digit 5
In fact, there is no evidence of phalangeal initiation of any kind in the
brief Sox9-expressing region anterior to the first digit in the wing. It remains plausible that this is a prepollex
The prepollex has no phalanges just like the pteroid which has no phalanges!

Good reference:

With the influence of SOCS2 there can be a graduated reduction of digit IV in the transition to basal Paraves:

To discover genes that specifically contribute to the second and third wing digit identities, we performed differential expression analysis of the mRNA-seq data between samples LFb and LFc. We found two genes, Tbx3 and Socs2, with high expression in sample LFc (Supplementary Fig. 9 and Fig. 3a). To our knowledge no studies have been published indicating a role for Socs2 in limb development. ISH confirms its strong expression in the third forelimb digit to the exclusion of all other digits in forelimb and hindlimb(Fig.3b-g). Recently it has been shown that the third forelimb digit has a unique mode of development in birds8. This, combined with our gene expression survey,supports the idea that the third wing digit has a unique derived identity in birds.
Suppressor of cytokine signaling (SOCS)-2 regulates normal postnatal growth and its deficiency in mice causes gigantism with increased bone length and proportional enlargement in skeletal muscles.
SOCS2 negatively regulates growth hormone action in vitro and in vivoMice deficient in SOCS2 display an excessive growth phenotype characterized by a 30-50% increase in mature body size. Here we show that the SOCS2-/- phenotype is dependent upon the presence of endogenous growth hormone (GH) and that treatment with exogenous GH induced excessive growth in mice lacking both endogenous GH and SOCS2. This was reflected in terms of overall body weight, body and bone lengths, and the weight of internal organs and tissues.
Welten et al
The bird wing is of special interest to students of homology and avian evolution. Fossil and developmental data give conflicting indications of digit homology if a pentadactyl "archetype" is assumed. Morphological signs of a vestigial digit I are seen in bird embryos, but no digit-like structure develops in wild-type embryos. To examine the developmental mechanisms of digit loss, we studied the expression of the high-mobility group box containing Sox9 gene, and bone morphogenetic protein receptor 1b (bmpR-1b)-markers for precondensation and prechondrogenic cells, respectively. We find an elongated domain of Sox9 expression, but no bmpR-1b expression, anterior to digit II. We interpret this as a digit I domain that reaches precondensation, but not condensation or precartilage stages. It develops late, when the tissue in which it is lodged is being remodeled. We consider these findings in the light of previous Hoxd-11 misexpression studies. Together, they suggest that there is a digit I vestige in the wing that can be rescued and undergo development if posterior patterning cues are enhanced
We observed Sox9 expression in the elusive ‘‘element X’’ that is sometimes stated to represent a sixth digit. Indeed, incongruity between digit domains and identities in theropods disappears if birds and other archosaurs are considered primitively polydactyl. Our study provides the first gene expression evidence for at least five digital domains in the chick wing. The failure of the first to develop may be plausibly linked to attenuation of posterior signals.


For instance, it is predictable that the loss of digit I from the manus should be simultaneous with loss of a single phalanx from one or more of the three remaining digits: as those losses are attributed to a single cause, the depletion of mesenehyme in the limb bud, they should not occur sequentially.  

We argue that limb bud cells that would normally form the digit II condensation proliferate toward a more anterior direction, into the space made physically available by the loss of digit I. This causes the presumptive digit II to leave the Shh activity zone. At the same time its cells do not express hoxD12 (and other posterior digit markers) any more, and therefore its transcriptome becomes characteristic of digit I.
Its phalangeal number is reduced--whether due to weaker anterior FGF8 signaling, caused by lower Shh levels that would otherwise stabilize the expression, or because of different hox and downstream gene expression.
Experiments with molecular signaling pathways in early limb development have shown that  modulating interdigital bmp signaling (Dahn and Fallon, 2000) or blocking bmp with a dominant negative receptor (Zou and Niswander, '96) is able to remove one phalanx from each digit
A comparison of digits within each limb shows that the posterior forelimb digits are more strongly differentiated than the posterior hindlimb digits, and forelimb digit III exhibits a unique expression of Socs2. 
Larry Febo
Xu and Mackem

Also see:

Wednesday, June 7, 2017

More evidence against the dinosaur to bird theory that needs to be explained away.
Tyrannosauroid integument reveals conflicting patterns of gigantism and feather evolution
Phil R. Bell, Nicolás E. Campione, W. Scott Persons, Philip J. Currie, Peter L. Larson, Darren H. Tanke, Robert T. Bakker
Recent evidence for feathers in theropods has led to speculations that the largest tyrannosaurids, including Tyrannosaurus rex, were extensively feathered. We describe fossil integument from Tyrannosaurus and other tyrannosaurids (Albertosaurus, Daspletosaurus, Gorgosaurus and Tarbosaurus), confirming that these large-bodied forms possessed scaly, reptilian-like skin. Body size evolution in tyrannosauroids reveals two independent occurrences of gigantism; specifically, the large sizes in Yutyrannus and tyrannosaurids were independently derived. These new findings demonstrate that extensive feather coverings observed in some early tyrannosauroids were lost by the Albian, basal to Tyrannosauridae. This loss is unrelated to palaeoclimate but possibly tied to the evolution of gigantism, although other mechanisms exist.
In fact, the early tyrannosauroids did not have feathers either.

With the benefit of hindsight it is easy to see that if fossils of the small flying dromaeosaurs from China had only been discovered before the larger flightless dromaeosaurs like Deinonychus or Velociraptor were found, the interpretations of the past three decades on how birds are related to dinosaurs would have been significantly different. If it had already been established that dromaeosaurs were birds that could fly, then the most logical interpretation of larger flightless dromaeosaurs found afterwards would have to be that they represented birds, basically like the prehistoric equivalent of an Ostrich, which had lost their ability to fly. 

Friday, May 26, 2017

Following the logic

Cladograms which are based on bootstrap/jackknife, show that there is a large polytomy at the base of the coelurosaur clade. This is because the hypothesized nodes (Maniraptoriformes, Maniraptora, Pennaraptora) are not supported and have been collapsed. The result is a much more correct picture of the taxa involved.

Xu et al
See Figure S9


Brusatte et al

This means (to take one example) that we cannot tell what the relationship is between oviraptorids and Paraves from the cladistic analysis. The two possibilities are that creatures similar to oviraptorids were transitional between dinosaurs and Paraves OR that oviraptorids were secondarily flightless members of  Paraves. It is important to note that the cladistic analyses cannot tell us which one is more credible.
Consequently we have to look for other indicators as to which is more credible.

We can immediately see that the alternative that oviraptorids were secondarily flightless members of Paraves means that we do not need to postulate ghost lineages. Also we do not need to postulate the pre-adaptations that are required by the idea that creatures similar to oviraptorids were transitional between dinosaurs and basal Paraves.

The same applies to ornithomimosaurs and alvarezsaurids.

Wednesday, April 26, 2017

Secondarily flightless paravians (2015)
Cau et al
Furthermore, phylogenetic analyses that incorporate sufficient character data are able to differentiate the members of such paravian lineages as Dromaeosauridae, Troodontidae and Avialae, as demonstrated by our present study. Nevertheless, reinterpretation of Balaur as a flightless avialan reinforces the point that at least some Mesozoic paravian taxa, highly similar in general form and appearance to dromaeosaurids, may indeed be the enlarged, terrestrialised descendants of smaller, flighted ancestors, and that the evolutionary transition involved may have required relatively little in the way of morphological or trophic transformation. (2017)
Most significantly, the taxon has the earliest known asymmetrical troodontid feathers, suggesting that feather asymmetry was ancestral to Paraves. (2007)
A Basal Dromaeosaurid and Size Evolution Preceding Avian Flight
Alan H. Turner
1,*, Diego Pol2, Julia A. Clarke3,4,1, Gregory M. Erickson5, Mark A. Norell1
Fossil evidence for changes in dinosaurs near the lineage leading to birds and the origin of flight has been sparse. A dinosaur from Mongolia [Mahakala] represents the basal divergence within Dromaeosauridae. The taxon's small body size and phylogenetic position imply that extreme miniaturization was ancestral for Paraves (the clade including Avialae, Troodontidae, and Dromaeosauridae), phylogenetically earlier than where flight evolution is strongly inferred. In contrast to the sustained small body sizes among avialans throughout the Cretaceous Period, the two dinosaurian lineages most closely related to birds, dromaeosaurids and troodontids, underwent four independent events of gigantism, and in some lineages size increased by nearly three orders of magnitude.  
So there is evidence of secondarily flightless paravians.
Let us now tie this with the issue of the statistically poorly supported core nodes. (See earlier posts).
Let's look at Pennaraptora. Pennaraptora is particularly poorly supported. Consequently Oviraptors may not be related to Paraves as sister taxa as commonly presented. Instead, Oviraptorids (that are dated 10's of millions of years later than basal Paraves) may well be secondarily flightless members of Paraves. And in fact that idea has been proposed over the years.
Halszka Osmólska et al. (2004) ran a cladistic analysis that came to a different conclusion. They found that the most birdlike features of oviraptorids actually place the whole clade within Aves itself, meaning that Caudipteryx is both an oviraptorid and a bird. In their analysis, birds evolved from more primitive theropods, and one lineage of birds became flightless, re-evolved some primitive features, and gave rise to the oviraptorids. This analysis was persuasive enough to be included in paleontological textbooks like Benton's Vertebrate Paleontology (2005).[11] The view that Caudipteryx was secondarily flightless is also preferred by Gregory S. Paul,[12]  et al.,[13] and Maryańska et al.[14]

The evidence all points to Oviraptorids being secondarily flightless members of derived Paraves. And of course, a few established researchers had already come to that conclusion.
The same logic applies to Ornithomimosaurs (as being secondarily flightless members of derived Paraves).
Paul (2002) has argued that the reason some maniraptoran taxa possess so many derived avian apomorphies is that they are, in fact, secondarily flightless birds that are more derived than basal avian taxa like Archaeopteryx. Although Paul (2002) retained a theropod ancestry for birds, support for his hypothesis would clearly complicate the consensus BMT view. A few cladistic analyses have retrieved Alvarezsauridae (e.g., Perle et al. 1993, 1994; Chiappe et al. 1998) and Oviraptorosauria (Lü et al. 2002, Marya´nska et al. 2002) as birds more derived than Archaeopteryx, and other noncladistic studies have proposed avian status for various oviraptorosaur (Elzanowski 1999, Lü et al. 2005) and dromaeosaur taxa (Czerkas et al. 2002, Burnham 2007). These studies have provided support for elements of Paul’s (2002) hypothesis.
Oviraptorosaurs, like deinonychosaurs, are so bird-like that several scientists consider them to be true birds, more advanced than ArchaeopteryxGregory S. Paul has written extensively on this possibility, and Teresa Maryańska and colleagues published a technical paper detailing this idea in 2002.[5][16][17]Michael Benton, in his widely respected text Vertebrate Paleontology, also included oviraptorosaurs as an order within the class Aves.[18] However, a number of researchers have disagreed with this classification, retaining oviraptorosaurs as non-avialan maniraptorans slightly more primitive than the deinonychosaurs.[19]
Avialan status for Oviraptorosauria
This analysis places Oviraptorosauria within Avialae, in a sister−group relationship with Confuciusornis. Oviraptorosaurs are hypothesized to be secondarily flightless. 
The status of oviraptorosaurs as secondarily flightless birds, more advanced than is Archaeopteryx, has already been suggested (Paul 1988; Olshevsky 1991; Elżanowski 1999; Lü 2000)
Paul's phylogeny from his influential book.
You can see the saltation from Tyrannoraptora to Paraves.
Constraining Epidexipteryx as a basal oviraptorosaur requires only one additional step in our dataset (fig. 75)
The great similarity that exists among basal paravians, basal oviraptorosaurs, and Epidexipteryx  (2015)
Testing the neoflightless hypothesis: propatagium reveals flying ancestry of oviraptorosaurs 
Alan Feduccia1 • Stephen A. Czerkas2
The presence of numerous flight features reveal that Caudipteryx, like the extant flightless ratites, originated from volant ancestors (de Beer 1956; Feduccia 2012, 2013), most likely via the evolutionary process of heterochrony, specifically paedomorphosis (arrested development), by which the adult retains the morphology of a younger stage of development (Livesey 1995).

(O'Connor and Sullivan 2014)
Reinterpretation of the Early Cretaceous maniraptoran (Dinosauria: Theropoda) Zhongornis haoae as a scansoriopterygid-like non-avian, and morphological resemblances between scansoriopterygids and basal oviraptorosaurs The condition present in Zhongornis resembles that seen in scansoriopterygids (Epidendrosaurus, Epidexipteryx) and basal oviraptorosaurs (Caudipteryx), which also have proportionately short tails compared to basal paravians
Dyke G J, Norell M A, 2005. Caudipteryx as a non-avialan theropod rather than a flightless bird. Acta Palaeont Pol, 50(1): 101–116
There is no reason—phylogenetic, morphometric or otherwise—to conclude that Caudipteryx is anything other than a small non−avialan theropod dinosaur.
"Non-avian theropod" could still be a member of Paraves.
Oviraptors were either secondarily flightless avialae or secondarily flightless non-avialae paraves.
They descended from flying ancestors. They are not transitional between dinosaurs and paraves.

Gregory Paul
The past two decades have witnessed the collapse of every single classical autapomorphy (or “holodiagnostic” to use Charig’s term) of Aves, from furculae, to feathers. Accordingly, distinguishing a par-avian theropod from a bona fide bird, is increasingly a matter of subjectivity. Add into this mess the argument that taxa traditionally classified as lying outside Aves are in fact neoflightless forms closer to Neornithes than is Archaeopteryx, and you have enough to drive the prospective student of avian phylogenetics to despair.
In order to test previous suggestions that oviraptorosaurs might be basal avialans, we ran two additional analyses. The first of these analyses was constrained to produce a monophyletic group comprising all oviraptorosaurian and non-archaeopterygid avialan species, whereas the second was constrained to produce a monophyletic group comprising all oviraptorosaurian and avialan including archaeopterygid species. The first analysis resulted in 1096 most parsimonious trees, each having a length of 1410 steps. Figure S10 shows the strict consensus of the 1096 trees. The second analysis resulted in 216 most parsimonious trees, each having a length of 1413 steps. Figure S11 shows the strict consensus of the 216 trees. These analyses indicate that the hypotheses that recover an Oviraptorosauria-Avialae clade are considerably less parsimonious than the hypothesis shown in Figure 6. However, one reason that the Oviraptorosauria-Avialae hypotheses are worse supported by our dataset might be the large amount of missing data from the palates and braincases of the basal oviraptorosaurs and basal avialans, regions that represent important sources for oviraptorosaurian synapomorphies. 
In oviraptorosaurs and basal avialans the supraacetabular crest is absent.
The remaining maniraptorans form the clade Pennaraptora ("feathered raptors"). These comprise the oviraptorosaurs, the scansoriopterygids, and the eumaniraptorans. These groups are united by several important characteristics:
  • Another increase of brain size
  • Laterally directed shoulder joint
  • Honest-to-goodness pennaceous feathers on at least the arms and tail (their presence on arms at least are documented further down the tree, at least shared with ornithomimosaurs)
  • Brooding on nests of eggs (may have been present in more basal coelurosaurs)

Thursday, April 20, 2017

"An abnormally rapid period of morphological evolution"

We have seen that the statistical support values do not support the dinosaur to bird hypothesis. See earlier posts.
Consequently we are justified in looking at the dinosaur to bird hypothesis much more critically and to re-assess the explanations that have been given. For example, the "explanation" that the evolution rate was abnormally rapid.
Stephen Brusatte et al
[R]ecent studies converge in identifying the dinosaur-bird transition as an abnormally rapid period of morphological evolution.

"Relative to the femur, the humerus is significantly longer and thicker in basal paravians than in non-paravian theropods." (Xu et al)
"The significant lengthening and thickening of the forelimbs indicates a dramatic shift in forelimb function at the base of the Paraves." (Xu et al)
"We find an increase in rates of body size and body size dependent forelimb evolution leading to small body size relative to forelimb length in Paraves." (Puttick et al)
"We were really surprised to discover that the key size shifts [body size and forelimb length] happened at the same time, at the origin of Paraves," said Mr Puttick of Bristol's School of Earth Sciences.  "This was at least 20 million years before the first bird, the famous Archaeopteryx, and it shows that flight in birds arose through several evolutionary steps."                     
As the quotes above show, basal paravians are very different than non-paravian theropods. The researchers who believe they are related, explain this as an "abnormally rapid period of morphological evolution". Of course that is not an explanation. It is an acknowledgement that they have no explanation.
Stephen Brusatte et al
Gradual Assembly of Avian Body Plan Culminated in Rapid Rates of Evolution across the Dinosaur-Bird Transition
Our results demonstrate that the rise of birds was a complex process: birds are a continuum of millions of years of theropod evolution, and there was no great jump between nonbirds and birds in morphospace, but once the avian body plan was gradually assembled, birds experienced an early burst of rapid anatomical evolution. 
Although birds are clearly distinct compared to all other living vertebrates, the avian bauplan isn’t especially distinct relative to other coelurosaurs, particularly their closest relatives.
There is growing evidence that changes in discrete character evolution, body size, and limb anatomy occurred quickly in the vicinity of the origin of birds, either at the node Avialae, in close avialan outgroups [basal paraves] or beginning with slightly more derived birds [3, 4, 5, 6, 19, 20, 21, 22]. It is likely that different types of data will pinpoint changes at slightly different positions on phylogeny, but in general, recent studies converge in identifying the dinosaur-bird transition as an abnormally rapid period of morphological evolution.
The initial results of the branch (Dryad Fig. S4-13) and clade (Fig. S3; Dryad Fig. S14-23) tests strongly support significantly high rates in Avialae, and to a lesser degree Tyrannosauroidea.
Other clades show significantly low or non-significant rates, with the exception of two smaller clades: Graciliraptor + Microraptor + Shanag + Sinornithosaurus + Tianyuraptor (within Dromaeosauridae), and Anchiornis + Aurornis + Eosinopteryx + Xiaotingia (within Troodontidae) [basal Paraves] which frequently show high rates. 
Important note:
Anchiornis + Aurornis + Eosinopteryx + Xiaotingia (within Troodontidae) [basal Paraves] frequently show high rates. 
In other words, the authors recognize the gap at Paraves. (2011)
An Archaeopteryx-like theropod from China and the origin of Avialae
Xing Xu1,2, Hailu You3 , Kai Du4 & Fenglu Han2
The discovery of Xiaotingia further demonstrates that many features
previously regarded as distinctively avialan actually characterize the
more inclusive Paraves. For example, proportionally long and robust
forelimbs are optimized in our analysis as a primitive character state
for the Paraves (see Supplementary Information). The significant
lengthening and thickening of the forelimbs indicates a dramatic shift
in forelimb function at the base of the Paraves, which might be related
to the appearance of a degree of aerodynamic capability.
We use the ratios of humeral length to femoral length, and humeral diameter to femoral diameter, as indicators of forelimb length and robustness. Relative to the femur, the humerus is significantly longer and thicker in basal paravians than in non-paravian theropods, derived dromaeosaurids and troodontids (the relatively short and slender forelimbs in the last two groups are secondarily evolved according to the current phylogenetic analysis). (2014)
Sustained miniaturization and anatomical innovation in the dinosaurian ancestors of birds
Michael S. Y. Lee,1,2* Andrea Cau,3,4 Darren Naish,5 Gareth J. Dyke5,6
Although there is no overall theropod-wide trend (fig. S7 and SM, part D), there is an exceptional trend within the single lineage that comprises much of the avian stem.
Our study quantifies rates of evolutionary innovation in dinosaurs using 1549 (data set 1) and 421 (data set 2) skeletal and other anatomical traits distributed across the entire body. A clear pattern emerges: Branches along the bird stem undergo substantially faster morphological evolution than those of the rest of the tree. 
Recent discoveries have highlighted the dramatic evolutionary transformation of massive, ground-dwelling theropod dinosaurs into light, volant birds. Here, we apply Bayesian approaches (originally developed for inferring geographic spread and rates of molecular evolution in viruses) in a different context: to infer size changes and rates of anatomical innovation (across up to 1549 skeletal characters) in fossils. These approaches identify two drivers underlying the dinosaur-bird transition. The theropod lineage directly ancestral to birds undergoes sustained miniaturization across 50 million years and at least 12 consecutive branches (internodes) and evolves skeletal adaptations four times faster than other dinosaurs. The distinct, prolonged phase of miniaturization along the bird stem would have facilitated the evolution of many novelties associated with small body size, such as reorientation of body mass, increased aerial ability, and paedomorphic skulls with reduced snouts but enlarged eyes and brains.

These results reconcile contradictory studies identifying presence (4–8) or absence (9–11) of a trend toward size reduction in theropods. Although there is no overall theropod-wide trend (fig. S7 and SM, part D), there is an exceptional trend within the single lineage that comprises much of the avian stem.
Also see Figure S6.

Let's look at body size and forelimb length.
Notice that the changes appear for the first time at the origin of Paraves (not earlier).

That shrinkage sped up once bird ancestors grew wings and began experimenting with gliding flight. Last year, Benton’s [Puttick] team showed that this dinosaur lineage, known as paraves, was shrinking 160 times faster than other dinosaur lineages were growing. “Other dinosaurs were getting bigger and uglier while this line was quietly getting smaller and smaller,” Benton said. “We believe that marked an event of intense selection going on at that point.”
Mark Puttick and colleagues investigated the rates of evolution of the two key characteristics that preceded flight: body size and forelimb length.  In order to fly, hulking meat-eating dinosaurs had to shrink in size and grow much longer arms to support their feathered wings. 
"We were really surprised to discover that the key size shifts happened at the same time, at the origin of Paraves," said Mr Puttick of Bristol's School of Earth Sciences.  "This was at least 20 million years before the first bird, the famous Archaeopteryx, and it shows that flight in birds arose through several evolutionary steps."                                     
High rates of evolution preceded the origin of birds (2014)
Puttick, M.N., Thomas, G.H., and Benton, M.J. in Evolution: DOI: 10.1111/evo.12363 
The origin of birds (Aves) is one of the great evolutionary transitions. Fossils show that many unique morphological features of modern birds, such as feathers, reduction in body size, and the semilunate carpal, long preceded the origin of clade Aves, but some may be unique to Aves, such as relative elongation of the forelimb. We study the evolution of body size and forelimb length across the phylogeny of coelurosaurian theropods and Mesozoic Aves. Using recently developed phylogenetic comparative methods, we find an increase in rates of body size and body size dependent forelimb evolution leading to small body size relative to forelimb length in Paraves, the wider clade comprising Aves and Deinonychosauria. The high evolutionary rates arose primarily from a reduction in body size, as there were no increased rates of forelimb evolution. In line with a recent study, we find evidence that Aves appear to have a unique relationship between body size and forelimb dimensions. Traits associated with Aves evolved before their origin, at high rates, and support the notion that numerous lineages of paravians were experimenting with different modes of flight through the Late Jurassic and Early Cretaceous.
The taxon's small body size and phylogenetic position imply that extreme miniaturization was ancestral for Paraves (the clade including Avialae, Troodontidae, and Dromaeosauridae), phylogenetically earlier than where flight evolution is strongly inferred. In contrast to the sustained small body sizes among avialans throughout the Cretaceous Period, the two dinosaurian lineages most closely related to birds, dromaeosaurids and troodontids, underwent four independent events of gigantism, and in some lineages size increased by nearly three orders of magnitude.

Also see: