Advertisement
Research Article

On the Size and Flight Diversity of Giant Pterosaurs, the Use of Birds as Pterosaur Analogues and Comments on Pterosaur Flightlessness

  • Mark P. Witton mail,

    mark.witton@port.ac.uk

    Affiliation: School of Earth and Environmental Sciences, University of Portsmouth, Portsmouth, United Kingdom

    X
  • Michael B. Habib

    Affiliation: Department of Sciences, Chatham University, Pittsburgh, Pennsylvania, United States of America

    X
  • Published: November 15, 2010
  • DOI: 10.1371/journal.pone.0013982
  • Published in PLOS ONE

Reader Comments (2)

Post a new comment on this article

Comments on Witton and Habib 2010

Posted by DavidPeters1247 on 27 Nov 2010 at 10:36 GMT

New and revised figures can be accessed at:
http://www.davidpetersstu...

Witton and Habib [1] made an important and well-reasoned assessment of the flight-worthiness of giant pterosaurs and I agree with their basic conclusions. Giant pterosaurs were indeed fully capable of flight. I write to attempt to correct certain data and hypothetical problems including the figures (see Figs. 1, 2).

Pterosaur origins

Witton and Habib [1] considered pterosaurs “animals of controversial phylogenetic affinities” then, ironically, referenced three papers, all of which concluded pterosaurs were related to Scleromochlus.

While this placement is the vast majority view, it is a problem because Scleromochlus is a basal crocodylomorph sharing a suite of synapomorphies with Gracilisuchus [2]. Otherwise in the literature, pterosaurs have never been considered crocodylomorphs. Scleromochlus had a wide flat skull, an antorbital fenestra with a deep fossa, tiny hands with greatly reduced lateral digits, a diverging pubis and ischium, it lacked a fifth toe and had several other traits [2] that rule it out as a suitable pterosaur sister taxon. Two studies [3, 4] found parasuchians to be sister taxa to pterosaurs, but no one ever ventured to produce a list of pterosaurian synapomorphies within or just outside of the Parasuchia. One study [5] found pterosaurs + Scleromochlus to be sisters to Dinosauromorpha and this clade to be the sister to Parasuchia and Proterochampsidae. The point is: no one has ever produced a series of generic taxa with an increasing list of pterosaurian synapomorphies from among the Archosauromorpha. Hone and Benton [6] put it best when they cited Bennett [7] in reporting, “Pterosaurs have been allied to virtually every basal and crown-group archosaur clade as well as to the dinosaurs, but few characters can be found that unite them with any other clade among the archosaurs.”

The only competing hypothesis [2, 8, 9] does document a series of three outgroup generic taxa (Cosesaurus, Sharovipteryx and Longisquama) of increasing similarity to basal pterosaurs, sharing an antorbital fenestra without a fossa, an attenuated tail, hands with unreduced lateral digits, an elongated ilium with an expanded sacral series, a fused puboischium, a pteroid [10], a preaxial carpal, an elongated fifth toe and soft tissues that only pterosaurs also have. More taxa precede these three including the basal squamate, Huehuecuetzpalli [10] which itself is a better candidate for pterosaur ancestry than any known archosaur with its retracted naris, unfused and simple hinge tarsals, unreduced lateral digits and ossified sternum. This phylogenetic genesis is relevant to the Witton and Habib [1] results (discussed below).

Bone fusion in pterosaurs

Witton and Habib [1] reported: “All pterosaurs bore robust, fused scapulacoracoids and, in derived, fully grown pterodactyloids, the anterior dorsal vertebrae fused into a rigid notarium…”

This statement is not supported by data. Several specimens of pterosaurs did not have fused scapulocoraocids. Moreover, several clades of pterosaurs did not have fused scapulocoracoids either, including many, but not all, members of the Ornithocheiridae, many of which grew to be quite large. Similarly, most fully-grown pterodactyloids did not have a rigid notarium. The Witton and Habib [1] bone fusion hypothesis comes from the false precept that pterosaurs were archosaurs (see above), which do fuse certain bones at maturity. In counterpoint, pterosaurs followed fenestrasaur patterns. While several clades of pterosaurs did fuse certain bones (whether at maturity or prior to hatching is not known), others did not, so the presently known pattern of fusion was phylogenetic rather than ontogenetic. With this new data, the number of valid juvenile pterosaurs in the fossil record drops considerably.

Bipedal walking

Witton and Habib [1] reported: “…bipedal walking, has been criticised with suggestions that the hindlimbs are comparatively diminutive compared to the rest of the body, that the hindlimb musculature would achieve poor mechanical advantage if the pterosaur body was elevated to an erect bipedal pose, that the anteriorly-positioned centre of gravity (induced by the large forelimbs, flight muscles and skull) would render the animal unstable and that the wings could not be folded away neatly [88]. As such, it seems unlikely that any ornithocheiroid could sustain a bipedal stance for a great length of time and would have had to overcome the hindlimb-forelimb length dichotomy inherent in their quadrupedal gait for sustained terrestrial locomotion.”

Their criticism is unfounded and the opposing viewpoint [2, 8, 9] was not referenced. No matter which ancestor one proposes, Scleromochlus or Sharovipteryx, both were obligate bipeds. Thus, from a phylogenetic point-of-view, quadrupedality in pterosaurs would have been secondary coming from either precursor.

Pterosaur hindlimbs were not diminutive, just the opposite. Pterosaurs are relatively long-legged with a tibia longer than the femur (a cursorial indicator) and a hind limb often longer than the torso. Pterosaurs and their fenestrasaur sisters had anteriorly elongated ilia used as anchors for expanded thigh muscles. Pterosaur precursors created a new bone, the prepubis, to anchor additional pubic muscles. Fenestrasaurs, including pterosaurs, had an expanded sacrum, a response to the stresses of bipedal locomotion. Larger pterosaurs incorporated more vertebrae into their sacrum. Pterosaurs had gastralia to stiffen the belly against the cantilever moment of the entire upper body working to bend it downward.

There would have been no mechanical disadvantage to a bipedal pose in pterosaurs just as there is no mechanical disadvantage to the 19 living species of lizards that are able to rise and/or locomote bipedally [8].

There would be no instability induced because elevation to a bipedal configuration places the center of balance, the shoulder, directly over the toes, as in birds.

The wing would have been easy to fold away by tucking it proximal to the elbow (with a reduced wing shape, Fig. 1).

I agree that in ornithocheirids a bipedal stance was unlikely for a great length of time. They did have a combination of the most slender hind limbs and smallest feet coupled with the most robust forelimbs of all pterosaurs. Members of this clade must have rarely landed and during such times, would have been clumsy at best. Such extremes are not the case, however, in most pterosaurs.

Facultative bipedalism in pterosaurs is supported on the basis of occasionally bipedal ichnites matched to the fenestrasaur Cosesaurus and the inability of Sharovipteryx to assume a quadrupedal configuration due to the disparate lengths of its fore and hindlimbs [9]. Virtually all pterosaurs extended the forelimbs far enough to contact the substrate while balanced with their shoulders over their toes.

That pterosaurs were secondarily quadrupedal is supported by pterosaur manual ichnites, which are directed laterally to posteriorly [9]. This hypothesis is also indicated by the inability of all pterosaurs to generate thrust from their forelimbs because the vector from hand to shoulder never had an anterior component, except when dubiously vaulting [1], an activity that has never been recorded as an ichnite. Essentially pterosaurs that walked quadrupedally used their forelimbs like ski poles for support, not for thrust.

Bipedal launching in pterosaurs has been criticized [1] by noting the relative strengths and cross-sections of pterosaur humeri and femora. Only one femur, that of Quetzalcoatlus sp., was displayed in Table 2 of Witton and Habib [1]. The authors’ noted it would fail in pure cantilever bending. Lucky for the pterosaur, such forces were rarely, if ever, applied to it. Substantial supporting tissues in the form of muscles, tendons and ligaments, surrounded the femur. By comparison, the femur of Sharovipteryx was proportionately much more gracile, yet it was an obligate biped [8] and would have launched bipedally.

Pterosaur bounding and leaping

Witton and Habib [1] reported: “…both pterosaur limb sets are adapted for powerful leaping, and at least bounding gaits were probably attainable in all pterosaurs.”
In counterpoint, absolutely no pterosaur tracks show a bounding gait.

Azhdarchid trackways and pedes

Witton and Habib [1] reported: “Azhdarchid trackways indicate that their feet were short (a mechanically advantageous trait for walking animals), possessed soft-tissue pads around the metatarsal heads and heels and that their limbs were held directly under the body when walking. Their skeletons reveal atypically long femora and wing metacarpals that serve to lengthen the limbs for increased stride efficiency, while their pedal bones are unusually robust.”

This statement is not supported by data. Azhdarchid pedes were long (Quetzalcoatlus) to extremely long and narrow (Zhejiangopterus) in comparison to the feet of other pterosaurs. The observation of Witton and Habib [1] that the limbs were largely under the body underscores their ability to elevate the forelimbs off the substrate without losing balance [2, 9].

Flightlessness possibilities in the Dimorphodontidae

Witton and Habib [1] remarked, “In our view…the pterosaur lineage closest to abandoning flight may not be giant at all but, rather, the considerably smaller basal pterosaur clade Dimorphodontidae (wingspans of 0.6–1.3 m). Dimorphodon has been found to be a particularly heavyset pterosaur with relatively high wing loading, attributes found in modern fliers like rails and galliforms that find flight particularly energetically expensive. Given that Dimorphodon also possesses an unusually robust skeleton – including long limbs and well-developed appendages - it was probably also a competent terrestrial (or, more likely, scansorial – see Unwin) animal that spent much of its time grounded (Hyder et al., in prep).

Dimorphodon was not “heavyset,” just the opposite. The skull, with its slender strut morphology, was more air than bone. The torso was much smaller than the skull and further shortened by a rather unique reduction in the dorsal vertebral count. The pelves were relatively small and shallow. The limbs were relatively robust because Dimorphodon was phylogenetically close to the base of all pterosaurs. With such robust unguals clearly this was an animal adept at clinging (to trees), with no clear indications of a terrestrial lifestyle seen in later, more derived small-clawed pterosaurs, such as Ctenochasma and Pterodaustro.

Quadrupedal launch hypothesis

Witton and Habib [1] remarked, “There is good evidence that pterosaurs launched from a standing, quadrupedal start in a superficially vampire bat-like fashion, vaulting over their forelimbs and using powerful flapping to gain altitude.”

First, there is no ichnite evidence for a quadrupedal launch.

Second, comparisons to vampire bats (Desmodus rotundus) leave certain doubts. In vampires (≤30 grams) the pectoralis creates 85% of the launch power [10]. The hind limbs produce just enough thrust to keep the forelimb thrust vector between the thumbs. The feet leave the substrate before the thumbs do. At apogee the bat’s back is three times higher than at loading and the forelimbs trail beneath the body. Afterwards the limbs rise and flex, creating a parachute to slow the descent. The wing fingers do not deploy until the wings continue their rise (and the body falls) for the first downstroke 0.10 of a second after becoming airborne and 0.20 of a second after initiating the launch sequence. In Desmodus the humerus is longer than the femur, tibia, pelvis or skull and 22% of the forelimb length, which also includes an antebrachium and slender elongated digits.

By contrast an azhdarchid, such as Zheijangopterus [12], does not have a similar morphology to Desmodus (but certain anurognathid pterosaurs do). In Zheijangopterus the humerus is a third the tibia, less than half the femur, shorter than the pelvis, about 13% of the skull length and less than 10% of the forelimb length. This pterosaur is a magnitude larger than Desmodus, but comparatively underpowered. Zhejiangopterus would have found it difficult, if not impossible, to launch to twice or thrice its standing height. Snapping open its large wings at the apogee of its brief leap would also be daunting.

As an alternative, the relatively long hind limbs and large pelvis could have provided relatively more leverage and power for running or leaping and launch. They were much more heavily muscled. Virtually all other tetrapod saltators, including the pterosaur precursors Sharovipteryx, also used heavily muscled hind limbs anchored on elongated hips for leaping. This coupled with extended wings providing lift and thrust from the start, would appear to be the mode of launch in pterosaurs, as in birds.

A pterosaur landing sequence has been preserved [13]. It simply reverses this take-off sequence as the pterosaur comes in for a two-point landing, then folds its wings while bipedal before initiating quadrupedal foraging.

A quadrupedal launch creates two great risks for the wings. First there is the danger from ground contact. The hyper-elongated wing fingers start the sequence incompletely folded posterior to the metacarpus and oriented toward the substrate. Sometime between apogee and the first downstroke the entire wing must rebound from the initial leap, which keeps it in the vertical plane absorbing the momentum of the leap (as in Desmodus). Near apogee the proximal wing must rotate transversely and rapidly rise past the horizontal plane, while snapping open the wing finger. This would be quite a feat for ornithocheirids in which the wing fingers are larger than any other bone in the body and may total several meters in length. Any leap that was not perfect would result in damage to the wing against the substrate and a crash. This assumes that every pterosaur was capable of leaping beyond the perimeter of its opening wings, which appears doubtful. It would be far less risky if the wings were already deployed and raised above the horizon ready for the first downstroke before the initiation of the leap, as in birds. With huge open wings providing much of the take-off power, the forces applied to the femora would have been greatly reduced.

Second, the thumbs of Desmodus are larger and more robust than those of other bats due to the magnified stresses involved at lift-off. At launch the thumbs transmit 9-10 times more force to the substrate than at rest [11]. Pterosaurs did not have such robust fingers. They were relatively small and weak, though not nearly as small as Witton and Habib [1] suggest (Fig. 2)! In pterosaurs metacarpal IV and the wing finger never touched the substrate, as indicated in all pterosaur manus tracks. So launch forces would not have traveled through the large fourth metacarpal, contra Witton and Habib [1] and Habib [14], but rather through the comparatively slender medial metacarpals I-III. In some taxa, such as Anhanguera (Fig. 2) these did not even ossify to the carpus, demonstrating that large axial stress forces did not travel through them.

Wing scaling in giant pterosaurs

Witton and Habib [1] reported, “The scaling allometry of the wing metacarpal is further evidence of this launch strategy: larger pterosaurs have disproportionately long wing metacarpals, a trait echoed in pterosaur ontogeny as well as phylogeny. During quadrupedal launching, the increased length of these elements would increase the mechanical advantage of the vaulting pterosaur to assist takeoff, possibly of particular importance to relatively large, heavy pterosaurs.”

These statements are not supported by data. Certain Nyctosaurus specimens (e.g. UNSM 93000, University of Nebraska State Museum, Lincoln) have such long metacarpals they cannot be positioned under the body and they are not the largest of pterosaurs. The JZMP-03-03-2 (Jinzhou Museum of Paleontology, Jinzhou City, Liaoning Province, China) embryo ornithocheirid has a metacarpus as long as the antebrachium, which is longer than in any other ornithocheirid adult. Given the example of Desmodus, the mechanical advantage in a quadrupedal launch would increase with increased humeral length, not metacarpal length. During a hypothetical quadrupedal launch, the short hind limbs and relatively short humerus of Nyctosaurus could only elevate the pterosaur a few inches at best, not nearly enough to clear the enormous opening wing fingers from contact with the substrate.

Procellariiform homologies

Witton and Habib [1] reported, “… there is little other quantitative evidence that any pterosaurs were specifically procellariiform-like in life. Comparing the procellariiform body plan to that of pterosaurs may show why such data is scarce: procellariiform bodies are not particularly pterosaur-like with longer, narrower wings that act independently of the hindlimbs, shorter necks, smaller heads and an entirely different pelvic and hindlimb morphology.”

This statement is not supported by data. Certain pterosaurs, such as Nyctosaurus were indeed procellariiform-like in several aspects such as size, wing shape and leg length. Pteranodon was larger, but otherwise similar (Fig. 1). The problem developed for Witton and Habib [1] due to their invention or adoption of a wing shape that was not supported by the data [2] (discussed below).

Wing proportions

Witton and Habib reported, “Pterosaur wings are constructed with different functional proportions than those of procellariiforms, with the brachial region (corresponding to the humeral length) relatively longer in procellariiforms, but the antebrachial region (ulna length in procellariiforms; unlna+syncarpus+wing metacarpal in pterosaurs) proportionally longer in pterosaurs.”

Unfortunately the wing shaped created by Witton and Habib [1] is not supported in the fossil record. No specimens preserve a membrane attached to the lateral tibia [2, contra 1, 14]. All specimens that preserve a wing membrane indicate a shape stretched only between the elbow and wingtip with a short fuselage fillet medial to the elbow extending to the anterior thigh [2]. These were all ignored or overlooked in a prior study [14]. Reports [14] of “membrane shrinkage” are inventions and mistakes based on data that otherwise supports the Zittel [2] wing shape (Fig. 1). Some were based on torn and twisted membranes, which preserve the brachiopatagium in front of the wing finger! The Zittel wing shape [2] permits folding of the wing finger proximal to the elbow. It also permits a virtual disappearance of the membrane when folded, as fossils [14] indicate.

Wing shape

Witton and Habib [1] reported, “The longer body and legs of Quetzalcoatlus could create a deep, low-aspect wing that would generate greater lift during takeoff (assuming ankle-attached brachiopatagia), while the smaller Pteranodon body and wings were narrower and produced less lift when launching but were more glide-efficient.”

Based on the Witton and Habib [1] figure (Fig. 1) this is true only in the wing shape they invented for both Quetzalcoatlus and Pteranodon. Neither taxon includes any specimens that preserve wing shape. However, by applying the wing shape patterns of other pterosaurs in which the membrane is well-preserved [2], results present very little difference other than the relative lengths of the wings and necks (Fig. 1b2, 1c2). In dorsal view it is clear that pterosaurs were shaped like high-aspect ratio airplanes, complete with horizontal stabilizers created by the sprawling hind limbs (yet another trait arguing against an archosaur ancestry and supporting affinity with Sharovipteryx).

Anhanguera forelimb illustration problems

Witton and Habib’s figure 5 (Fig. 2a, b) presents the left forelimb of Anhanguera with hypothetical muscles in place. This is a model used to present the hypothesis of forelimb launching in pterosaurs and as such it puts the extensor tendon process in contact with the substrate (the lowest point in the left illustration) and provides robust hypothetical muscles over the bones.

There are many problems with this reconstruction, many of which pertain to the supporting hypotheses of Witton and Habib [1]. The free fingers and their metacarpals were illustrated much too small and they were improperly placed of the dorsal surface of metacarpa IV (Fig. 2b). No other pterosaurs have this configuration. In all pterosaurs metacarpal III is subequal to metacarpal IV. In Anhanguera metacarpals I and II are also subequal to III (Fig. 2c). Whether or not the ossified portions reach proximally to the carpus does not affect this distal placement. With the metacarpals correctly placed, fingers I-III extended beyond metacarpal IV.

Only the medial three fingers created known ichnites. Metacarpal IV was elevated off the substrate protecting the wing finger from ground contact (contra Witton and Habib [1]. In all pterosaurs fingers I-III retained their primitive orientation of palmar side ventral (in flight) or medial (when quadrupedal) with the antebrachia in the neutral position (neither supinated nor pronated). No pterosaur preserves fingers I-III on the dorsal surface of the wing as Witton and Habib [1] illustrated them. Specimens that preserve the slender metacarpals pressed against the anterior face of the large metacarpal did so post-mortem due to the fragile hinge at the metacarpal III-IV interface. Pterosaurs preserved dorsal side down may preserve fingers I-III on the ventral surface of metacarpal IV due to the same taphonomic hinge effect.

Of lesser importance, the flexor tendon on the wing finger in Witton and Habib [1] figure 5 (Fig. 2a, b) should not have inserted on the flexor process, but rather a short distance beyond it, as in living tetrapods. In that way the flexor could fully flex the wing against the metacarpus, rather than just pull it to a right angle where the pulling vector is reduced to zero.

Figure 5 in Witton and Habib [1] (Fig. 2a) also appears to be much too heavily muscled. The slenderness of the antebrachial bones and the need to reduce weight for flight would tend to indicate a more gracile muscle set (Fig. 2c).

Summary

Witton and Habib [1] were correct with regard to their hypotheses of giant pterosaur flight. I provided evidence they were incorrect with regard to wing shape, free finger configuration and quadrupedal launching. There is no inchnite evidence indicative of a quadrupedal launch in pterosaurs and such a launch would risk damage to a rapidly opening wing finger. A launch sequence involving flapping at the instant of leaping would greatly reduce this danger. The Witton and Habib [1] hypothesis was based on the model provided by the bat, Desmodus, but pterosaurs much smaller than the large azhdarchids they employed were much more similar in overall size, weight and wing segment proportions.

References

1. Witton MP Habib MB (2010) On the Size and Flight Diversity of Giant Pterosaurs, the Use of Birds as Pterosaur Analogues and Comments on Pterosaur Flightlessness. PLoS ONE 5(11): e13982. doi:10.1371/journal.pone.0013982.

2. Peters D (2002) A New Model for the Evolution of the Pterosaur Wing—with a twist. Historical Biology 15: 277–301.

3. Irmis RB, Nesbitt S, Padian K, Smith ND, Turner AH (2007) A Late Triassic Dinosauromorph Assemblage from New Mexico and the Rise of Dinosaurs. Science 317: 358–361.

4. Nesbitt S (2007) The anatomy of Effigia okeeffeae (Archosauria, Suchia), theropod-like convergence, and the distribution of related taxa. Bulletin of the American Museum of Natural History 302: 1–84.

5. Sereno PC 1991. Basal Archosaurs: phylogenetic relationships and functional implications. Society of Vertebrate Paleontology, Memoire 2, Journal of Vertebrate Paleontology 11, Supplement (4), 1-53.

6. Hone DWE, Benton MJ (2007) An evaluation of the phylogenetic relationships of the pterosaurs to the archosauromorph reptiles. Journal of Systematic Paleontology 5: 465–469.

7. Bennett SC (1996) The phylogenetic position of the Pterosauria within the Archosauromorpha. Zoological Journal of the Linnnean Society London 118: 261–309.

8. Peters D (2000a) A Redescription of Four Prolacertiform Genera and Implications for Pterosaur Phylogenesis. Rivista Italiana di Paleontologia e Stratigrafia, 106: 293–336.

9. Peters D (2000b). Description and Interpretation of Interphalangeal Lines in Tetrapods. Ichnos 7(1):11–41.

10. Peters D (2009). A reinterpretation of pteroid articulation in pterosaurs. Journal of Vertebrate Paleontology 29:1327–1330.

11. Schutt WA Jr, Altenbach JS, Chang YH, Cullinane DM, Hermanson JW, Muradali F, Bertram JE (1997) The dynamics of flight-initiating jumps in the common vampire bat Desmodus rotundus. Journal of Experimental Biology 200(Pt 23):3003–12.

12. Cai Z, Wei F (1994) On a new pterosaur (Zheijangopterus linhaiensis gen. et sp. nov.) from Upper Cretaceous in Linhai, Zhejiang, China. Vertebrata Palasiatica 32: 181–194.

13. Mazin J-M, Billon-Bruyat J-P, Padian K (2009) First record of a pterosaur landing trackway. Proceedings of the Royal Society of London B. doi: 10.1098/rspb.2009.1161

14. Habib MB (2008) Comparative evidence for quadrupedal launch in pterosaurs. Zitteliana B28: 161–168.

14. Elgin RA, Hone DWE, Frey E (In press) The extent of the pterosaur flight membrane. Acta Palaeontologica Polonica.


No competing interests declared.

RE: Comments on Witton and Habib 2010

MHabib replied to DavidPeters1247 on 03 Dec 2010 at 01:35 GMT

The authors of the manuscript (M. Witton and M. Habib) wish to reply to the thoughtful comments of David Peters to clarify several points of interest. We find the comments of D. Peters to be thought-provoking and deserving of proper scientific reply, but we do find several errors in the reasoning he presents.

Regarding Pterosaur origins

Original comments by D. Peters: "Witton and Habib [1] considered pterosaurs “animals of controversial phylogenetic affinities” then, ironically, referenced three papers, all of which concluded pterosaurs were related to Scleromochlus.

While this placement is the vast majority view, it is a problem because Scleromochlus is a basal crocodylomorph sharing a suite of synapomorphies with Gracilisuchus [2]. Otherwise in the literature, pterosaurs have never been considered crocodylomorphs…

…The only competing hypothesis [2, 8, 9] does document a series of three outgroup generic taxa (Cosesaurus, Sharovipteryx and Longisquama) of increasing similarity to basal pterosaurs, sharing an antorbital fenestra without a fossa, an attenuated tail, hands with unreduced lateral digits, an elongated ilium with an expanded sacral series, a fused puboischium, a pteroid [10], a preaxial carpal, an elongated fifth toe and soft tissues that only pterosaurs also have."

**Author response: While we appreciate the counterpoint provided by D. Peters, we note that the phylogenetic affinities of Pterosauria are, in fact, not particularly relevant to the issue of flight capacity in giant pterosaurs. All of the taxa we included in our analysis are highly derived members of the in-group Pterosauria, and our arguments of a functional nature, drawing on evidence from comparative biomechanics and direct evidence from pterosaur anatomy and trackways. We also note that the phyogenetic hypothesis that Peters prefers has been called into question based upon character identification, much of which relied on photographic studies. Chris Bennett has challenged this character state reconstruction method, and has made one of his pieces on the subject freely available at: bigcat.fhsu.edu/biology/cbennett/Bennett-PT-article.pdf Future analyses may validate the phylogeny of D. Peters, but we feel it is premature to draw firm conclusions regarding pterosaur origins at this time, and we have therefore opted for a biomechanical approach that does not rely on that information.

It is our observation that many, if not most, of the concerns D. Peters raises with our manuscript seem to be linked to a supposition that bipedality was the ancestral condition for Pterosauria, and that this condition must have therefore had mechanical consequences for pterosaur locomotion. We make no specific suppositions about the gait of pterosaur ancestors - pterosaurs could have been primarily quadrupedal or secondarily quadrupedal, but our results are not dependent upon this differentiation.


Regarding Bone fusion in pterosaurs
Original comments by D. Peters: "Several specimens of pterosaurs did not have fused scapulocoraocids. Moreover, several clades of pterosaurs did not have fused scapulocoracoids either, including many, but not all, members of the Ornithocheiridae, many of which grew to be quite large. Similarly, most fully-grown pterodactyloids did not have a rigid notarium. The Witton and Habib [1] bone fusion hypothesis comes from the false precept that pterosaurs were archosaurs (see above), which do fuse certain bones at maturity."

**Author response: We welcome the alternative view of D. Peters but note that neither he, nor any other author, has demonstrated that an archosaur placement for pterosaurs is a "false precept", and as such, we stand by our assessment that the elements in question fuse in adulthood, and that many pterosaurs known from the fossil record were simply juveniles.


Regarding Bipedal walking
Original comments by D. Peters: "Pterosaur hindlimbs were not diminutive, just the opposite. Pterosaurs are relatively long-legged with a tibia longer than the femur (a cursorial indicator) and a hind limb often longer than the torso. Pterosaurs and their fenestrasaur sisters had anteriorly elongated ilia used as anchors for expanded thigh muscles. Pterosaur precursors created a new bone, the prepubis, to anchor additional pubic muscles. Fenestrasaurs, including pterosaurs, had an expanded sacrum, a response to the stresses of bipedal locomotion. Larger pterosaurs incorporated more vertebrae into their sacrum. Pterosaurs had gastralia to stiffen the belly against the cantilever moment of the entire upper body working to bend it downward."

**Author response: D. Peters seems to have mis-understood our use of the term "diminutive". While some pterosaurs did have relatively long hind limb elements, these elements are also slender in most species, with modest resistance to bending. The total mass and volume of the hind limb elements is diminutive in most pterosaurs when compared to the forelimbs, head, and cervical series. While it is true that pterosaurs had elongated ilia and an expanded sacrum, there is no evidence at present that this must have been a response to the stresses of bipedal locomotion. Pterosaur trackways repeatedly demonstrate that pterosaurs were quadrupedal (for example: Unwin, 1997; Lockley, 2001; Mazin et al., 2009).

Original comments by D. Peters: "There would have been no mechanical disadvantage to a bipedal pose in pterosaurs just as there is no mechanical disadvantage to the 19 living species of lizards that are able to rise and/or locomote bipedally [8]. There would be no instability induced because elevation to a bipedal configuration places the center of balance, the shoulder, directly over the toes, as in birds."

**Author response: The mechanical disadvantage for pterosaurs in a bipedal posture would be substantial based on our data. The hind limb structure of pterosaurs differed substantially from that of living lizards, and unlike facultatively bipedal lizards, the forelimbs of pterosaurs were much more robust than their hind limbs (see Table 2 in original manuscript). Furthermore, our bending strength analysis demonstrates that the femora of large pterosaurs had insufficient strength in bending to sustain bipedal running. We observe that a pterosaur balancing in the bipedal position Peters suggests would force the wings into an inappropriate angle of attack, well beyond the stall limit for any flying vertebrate, which contradicts his assertions on launch (see below). A pterosaur that rose and ran in bipedal posture would lose the substantial power of the forelimbs, which were considerably more robust than the hind limbs (see Table 2 in original manuscript).

Original comments by D. Peters: "That pterosaurs were secondarily quadrupedal is supported by pterosaur manual ichnites, which are directed laterally to posteriorly [9]. This hypothesis is also indicated by the inability of all pterosaurs to generate thrust from their forelimbs because the vector from hand to shoulder never had an anterior component, except when dubiously vaulting [1], an activity that has never been recorded as an ichnite. Essentially pterosaurs that walked quadrupedally used their forelimbs like ski poles for support, not for thrust. "

**Author response: Pterosaurs may very well have been secondarily quadrupedal; we make no claims to the locomotion of the pterosaur outgroup. However, the proper propulsion component cannot be accurately calculated by simply drawing a line from hand to shoulder, and in any case, pterosaur trackways indicate significant body weight on the forelimbs (Lockley, 2001; Mazin et al., 2009).

Original comments by D. Peters: "Bipedal launching in pterosaurs has been criticized [1] by noting the relative strengths and cross-sections of pterosaur humeri and femora. Only one femur, that of Quetzalcoatlus sp., was displayed in Table 2 of Witton and Habib [1]. The authors’ noted it would fail in pure cantilever bending. Lucky for the pterosaur, such forces were rarely, if ever, applied to it. Substantial supporting tissues in the form of muscles, tendons and ligaments, surrounded the femur. By comparison, the femur of Sharovipteryx was proportionately much more gracile, yet it was an obligate biped [8] and would have launched bipedal."

**Author response: We supplied only the femoral strength of Quetzalcoatlus because our manuscript deals primarily with azhdarchid pterosaurs. However, MBH has calculated bending strength in other pterosaurs (including Anhanguera, Pteranodon, Anurognathus, and others) and in all such sampled taxa the ratios of humeral to femoral strength suggest quadrupedal launching. Soft tissue surrounding the femur would improve its resistance to impacts, but have only a limited effect on bending, because such tissues are highly elastic and passively deform - the Young's modulus for muscle being stretched is about 20kPa while that of bone is 20GPa - 1000 times greater (Lieber, 2002). It is true that pure cantilever bending was unlikely for pterosaur femora (a note we make in the paper directly). However, even with the true moment arm during leaping accounted for (about 33% of the full femur length for azhdarchids), the safety factor under the multiple body weights required for launching would be less than one, and therefore impossible for the element to sustain.


Pterosaur bounding and leaping
Original comments by D. Peters: "This statement is not supported by data. Azhdarchid pedes were long (Quetzalcoatlus) to extremely long and narrow (Zhejiangopterus) in comparison to the feet of other pterosaurs. The observation of Witton and Habib [1] that the limbs were largely under the body underscores their ability to elevate the forelimbs off the substrate without losing balance [2, 9]."

**Author response: We are uncertain as to where D. Peters is receiving his data on the foot proportions of Quetzalcoatlus, as these have never been published. Drawings of the animal have previously used other pterosaur pedal proportions as estimates. The authors have accessed the material personally and while it remains unpublished data, did not find support for the statement that the proportions were long in that animal. Placing the limbs under the body does not necessarily suggest that an animal can balance in a bipedal posture.


Flightlessness possibilities in the Dimorphodontidae
Original comments by D. Peters: "Dimorphodon was not “heavyset,” just the opposite. The skull, with its slender strut morphology, was more air than bone. The torso was much smaller than the skull and further shortened by a rather unique reduction in the dorsal vertebral count. The pelves were relatively small and shallow. The limbs were relatively robust because Dimorphodon was phylogenetically close to the base of all pterosaurs. With such robust unguals clearly this was an animal adept at clinging (to trees), with no clear indications of a terrestrial lifestyle seen in later, more derived small-clawed pterosaurs, such as Ctenochasma and Pterodaustro."

**Author response: We suggest that Dimorphodon was scansorial in our own analysis, and therefore are uncertain of the disagreement in that regard. We are unaware of any detailed analysis of pneumaticity in Dimorphodon which could validate the assertion by D. Peters that the skull was "more air than bone". For an open-access review and analysis of pterosaur pneumaticity, see: http://www.plosone.org/ar...


Quadrupedal launch hypothesis
Original comments by D. Peters: "First, there is no ichnite evidence for a quadrupedal launch."

**Author response: there is no trackway evidence of any launch (bipedal or quadrupedal) from any pterosaurs currently known. We do note, however, that researchers have previously searched for bipedal launch tracks, and found none. Very few individuals have actually searched for quadrupedal launch tracks, and Unwin (pers. com.) has suggested that launch trackways may have been missed because individuals working on pterosaur ichnites did not have the correct search image.

Original comments by D. Peters: "Second, comparisons to vampire bats (Desmodus rotundus) leave certain doubts. In vampires (≤30 grams) the pectoralis creates 85% of the launch power [10]. The hind limbs produce just enough thrust to keep the forelimb thrust vector between the thumbs. The feet leave the substrate before the thumbs do. At apogee the bat’s back is three times higher than at loading and the forelimbs trail beneath the body. Afterwards the limbs rise and flex, creating a parachute to slow the descent. The wing fingers do not deploy until the wings continue their rise (and the body falls) for the first downstroke 0.10 of a second after becoming airborne and 0.20 of a second after initiating the launch sequence. In Desmodus the humerus is longer than the femur, tibia, pelvis or skull and 22% of the forelimb length, which also includes an antebrachium and slender elongated digits."

**Author response: We note that Peters omitted the source for the data on Desmodus launch. Those data come from Schutt et al. (1997), and readers can utilize that source for further information on vampire launch. We do not claim that pterosaur launch would be similar to vampire bat launch in detail, but only that it would be broadly similar in being quadrupedal, with the forelimbs supplying most of the power. The long humerus and thumb of Desmodus are utilized in launch, but are not required for quadrupedal launch. Other morphologies can also be consistent with quadrupedal launch, and we suggest the vault-ballistic model for pterosaurs because it is consistent with both the quantitative mechanics and general anatomy of those animals. The quadrupedal launch systems of pterosaurs and bats are not homologous and should not be expected to show point-to-point similarity.

Original comments by D. Peters: "By contrast an azhdarchid, such as Zheijangopterus [12], does not have a similar morphology to Desmodus (but certain anurognathid pterosaurs do). In Zheijangopterus the humerus is a third the tibia, less than half the femur, shorter than the pelvis, about 13% of the skull length and less than 10% of the forelimb length. This pterosaur is a magnitude larger than Desmodus, but comparatively underpowered. Zhejiangopterus would have found it difficult, if not impossible, to launch to twice or thrice its standing height. Snapping open its large wings at the apogee of its brief leap would also be daunting."

**Author response: In fact, because pterosaurs could utilize a large number of muscle groups in powering launch, had disproportionately large forelimb and pectoral apparatus, and were likely utilizing a large amount of anaerobic muscle with large power generation (see Marden 1994), an animal such as Zhejiangopterus would have been relatively overpowered relative to the average bat. It also need not launch three times its standing height to launch; vampire bats specifically use a highly vertical launch trajectory which we find unlikely for pterosaurs. Again, we emphasize that the similarities between pterosaur launch and bat launch would be general in nature. A brief quantitative analysis of power generation in azhdarchid pterosaurs demonstrates that they would produce sufficient power during a quadrupedal launch to takeoff (see below). Opening of the wing finger in a rapid manner would have also been sufficiently efficient because of the large moment arm provided by the carpal structure in most pterosaurs (Prondvai and Hone, 2009).

With regards to power generation for launch, we submit the following as a basic demonstration of the error in assuming low power in large pterosaurs, using a 260kg Quetzalcoatlus northropi as an example.

If the tendon material in the Flexor Digitorum system was avian to croc grade, and the locking system was in place as expected, then the power mag factor could easily be 3 to 4. To be conservative, we originally used only a range of 1.5-2 for our assessment, and the animals still achieved sufficient speed to launch. For a quad launch with the expected 3.5 power magnification (still only half that of a frog, by way of comparison - see Biewener 2003) we estimate: 50kg flight muscle*390W/kg*3.5=68,250 Watts of power from the forelimb impulse.

This is excluding the fact that some of non-flapping oriented upper limb musculature can also power launch (like the relatively massive Latissimus dorsi and teres muscles). Adding power from the hindlimbs (likely aerobic, and therefore more on the order of 175W/kg, as compared to the 390W/kg for anaerobic flight muscle in archosaurs - see Askew and Marsh, 2001), brings the total power to well over 70,000Watts. This is emphatically not the amount of power available during flapping - one of the advantages of building most of the takeoff speed during the launch leap (which essentially all living flyers do) is the massive power magnification animals get during impulse maneuvers when starting from a preload and/or using a lock-and-release system. Frogs get about 700% magnification, for example. Pterosaurs were unlikely to have been frog-like leapers, but we note that they did very likely have huge tendons for energy storage in the forelimb. At this power level, a 200 kg Quetzalcoatlus with a 2.5 meter shoulder height at rest would achieve nearly a roughly 5 meter shoulder height at the top of the launch leap, accelerate at nearly 25m/s^2 during launch, and already be traveling at steady state speed by the time the wings unfolded, with a launch time of around half a normal flapping stroke cycle (the timing of vertebrate launch has a direct proportionality to flapping time in most living species, and is faster in quadrupedal launchers than bipedal launchers).

Original comments by D. Peters: "As an alternative, the relatively long hind limbs and large pelvis could have provided relatively more leverage and power for running or leaping and launch. They were much more heavily muscled. Virtually all other tetrapod saltators, including the pterosaur precursors Sharovipteryx, also used heavily muscled hind limbs anchored on elongated hips for leaping. This coupled with extended wings providing lift and thrust from the start, would appear to be the mode of launch in pterosaurs, as in birds."

**Author response: The hind limbs of pterosaurs were long, but weak. They could not support launch forces as the bending strength was insufficient to sustain multiple body weights during takeoff (this is explained in the manuscript to which this commentary refers). Given the low strength in bending of the hind limb elements, it is highly unlikely that the muscles attached to them produced greater forces than those of the forelimb, a conclusion which is further supported by the relatively greater volume available for muscle attachment and expansion in the pectoral girdle and forelimb of pterosaurs, as compared to their hind limb. The leverage for leaping was considerably less advantageous in the hind limb, as it could not be preloaded in the same manner as the forelimb, and the hind limb bone elements would break under the loads needed for acceleration to flight speed in a large pterosaur. For a pterosaur to balance bipedally in the manner described by Peters, the wings must be placed at a very high angle of attack, at least 25 degrees, and possibly greater. At this steep angle, the wings would almost certainly stall (stall could be delayed to some extent for small species) - large vertebrate flyers today have a maximum angle of attack around 12 degrees, sometimes slightly more (Alexander, 2002): readers should note that larger flyers usually fly at a lower angle of attack than small ones.

Original comments by D. Peters: A pterosaur landing sequence has been preserved [13]. It simply reverses this take-off sequence as the pterosaur comes in for a two-point landing, then folds its wings while bipedal before initiating quadrupedal foraging.

**Author response: Bats launch quadrupedally when taking off from the ground (a behavior not limited to vampires, incidentally, flying foxes use quadrupedal launch from the ground and can do so even with heavy loads (MacAyeal et al., in press)), but bats land hind limb first (Riskin et al., 2009) - there is nothing contradictory about the prospect of pterosaurs being quadrupedal launchers with a forelimb-driven takeoff while touching down with the hind limbs first during landing. In fact, it is the expected behavior based on living comparative examples and mechanical principles.

Original comments by D. Peters: A quadrupedal launch creates two great risks for the wings. First there is the danger from ground contact. The hyper-elongated wing fingers start the sequence incompletely folded posterior to the metacarpus and oriented toward the substrate. Sometime between apogee and the first downstroke the entire wing must rebound from the initial leap, which keeps it in the vertical plane absorbing the momentum of the leap (as in Desmodus). Near apogee the proximal wing must rotate transversely and rapidly rise past the horizontal plane, while snapping open the wing finger. This would be quite a feat for ornithocheirids in which the wing fingers are larger than any other bone in the body and may total several meters in length. Any leap that was not perfect would result in damage to the wing against the substrate and a crash. This assumes that every pterosaur was capable of leaping beyond the perimeter of its opening wings, which appears doubtful.

**Author response: we have applied ballistic equations to estimate the maximum height after launch during quadrupedal takeoff. For all pterosaurs so far sampled (12 species) we find that the launch height is sufficient for wing clearance assuming that the wing is still mostly folded as it is raised into the first upstroke prior to the initial power stroke. That the wing fingers are several meters in length is actually not a major impediment, because the wing finger is assumed to have remained folded against the rest of the forelimb during the initial upstroke (as in bats, Schutt et. al. 1997). Quadrupedal launch does not, in fact, present a clearance deficit. Instead the tradeoff is simply the time of the additional upstroke at the end of the leap. The time required for this can be easily estimated using the flapping frequency equations developed by Pennycuick (2008). Even for large pterosaurs, this additional time turns out to be sufficiently brief (approximately 0.12 seconds for Quetzalcoatlus northropi) that the added time in air from the powerful quadrupedal launch more than makes up for the difference.

In fact, contrary to the assertion of D. Peters, quadrupedal launch reduces the problems of clearance for the wing, because the animal achieves a much greater height (as a result of greater launch power). Vampire bats leap much higher than similarly-sized birds, for example (Schutt et al., 1997). Animal flyers drive themselves into the air with the stance limbs first, prior to engaging the wings (Heppner and Anderson, 1985; Bonser and Rayner, 1996; Earls, 2000; Tobalske et al. 2004; Nachtigall and Wilson, 1967; Nachtigall, 1968; 1978; Schouest et al., 1986; Trimarchi and Schneiderman, 1993; 1995). As a result, the clearance for the wings is determined by the total power available for leaping. Bipedal launch in pterosaurs would provide insufficient clearance for the wings.

Original comments by D. Peters: It would be far less risky if the wings were already deployed and raised above the horizon ready for the first downstroke before the initiation of the leap, as in birds. With huge open wings providing much of the take-off power, the forces applied to the femora would have been greatly reduced.

**Author response: This is an inaccurate portrayal of animal launch. The wings of birds are raised during the launch leap, not prior (Earls, 2000) and in no flying vertebrates do the wings provide the primary launch power (Heppner and Anderson, 1985; Bonser and Rayner, 1996; Earls, 2000; Tobalske et al. 2004). Even in insects, the walking limbs usually provide the takeoff impulse (Nachtigall and Wilson, 1967; Nachtigall, 1968; 1978; Schouest et al., 1986; Trimarchi and Schneiderman, 1993; 1995). In fact, the wings of a large vertebrate cannot provide much power during launch, because of the low speed regime at the onset of launch and the power and time required to build full circulation on the wing (an effect originally described by Wagner, 1925). No animal alive today is known to take off in the manner D. Peters describes, and in fact, fluid theory predicts that no flying animal much larger than the smallest birds should be able to do so (as it turns out, even the smallest birds do not - hummingbirds get most of their launch power from the hind limbs, see Tobalske et al., 2004).

Original comments by D. Peters: Second, the thumbs of Desmodus are larger and more robust than those of other bats due to the magnified stresses involved at lift-off. At launch the thumbs transmit 9-10 times more force to the substrate than at rest [11]. Pterosaurs did not have such robust fingers. They were relatively small and weak, though not nearly as small as Witton and Habib [1] suggest (Fig. 2)! In pterosaurs metacarpal IV and the wing finger never touched the substrate, as indicated in all pterosaur manus tracks. So launch forces would not have traveled through the large fourth metacarpal, contra Witton and Habib [1] and Habib [14], but rather through the comparatively slender medial metacarpals I-III. In some taxa, such as Anhanguera (Fig. 2) these did not even ossify to the carpus, demonstrating that large axial stress forces did not travel through them.

**Author response: pterosaur trackways are not clear on this factor: it is possible that the center of the manus tracks from pterosaur ichnites preserve some impression from the wing finger pivot joint (MCIV to PhIV-1) because the central region of the prints is often broad. However, regardless of personal interpretation in this regard, there is no evidence that any pterosaur was incapable of pressing the MCIV:PhIV-1 joint against the substrate during a transient motion. Given the substantial reinforcement of that joint, we currently consider it highly likely that the MCIV:PhIV-1 joint was subjected to high loads, which implies that it contacted the substrate in some way - the reinforcement of this joint, it should be noted, is largely perpendicular to the primary direction of bending in flight but aligned with the primary bending direction for quadrupedal launching and walking.


Wing scaling in giant pterosaurs
Original comments by D. Peters: Witton and Habib [1] reported, “The scaling allometry of the wing metacarpal is further evidence of this launch strategy: larger pterosaurs have disproportionately long wing metacarpals, a trait echoed in pterosaur ontogeny as well as phylogeny. During quadrupedal launching, the increased length of these elements would increase the mechanical advantage of the vaulting pterosaur to assist takeoff, possibly of particular importance to relatively large, heavy pterosaurs.”

These statements are not supported by data. Certain Nyctosaurus specimens (e.g. UNSM 93000, University of Nebraska State Museum, Lincoln) have such long metacarpals they cannot be positioned under the body and they are not the largest of pterosaurs. The JZMP-03-03-2 (Jinzhou Museum of Paleontology, Jinzhou City, Liaoning Province, China) embryo ornithocheirid has a metacarpus as long as the antebrachium, which is longer than in any other ornithocheirid adult.

**Author response: We only stated that relatively long metacarpals would be of utility to large pterosaurs, not that only large pterosaurs have long metacarpals. Furthermore, while the metacarpals in Nyctosaurids are indeed very long, the wing finger in those same animals is also very long, such that the fourth metacarpal does not actually occupy a much greater proportion of the wing than in many other pterodactyloids. By contrast, the metacarpal comprises a disproportionate component of the length of the wing in azhdarchids.

Original comments by D. Peters: "Given the example of Desmodus, the mechanical advantage in a quadrupedal launch would increase with increased humeral length, not metacarpal length. During a hypothetical quadrupedal launch, the short hind limbs and relatively short humerus of Nyctosaurus could only elevate the pterosaur a few inches at best, not nearly enough to clear the enormous opening wing fingers from contact with the substrate."

**Author response: We believe that D. Peters has misunderstood the comparative example of Desmodus. Bats in general have elongate humeri. While in vampires the robust humerus is very important mechanically during launch, it need not be the primary moment arm for all quadrupedal launchers. Elongation of other elements in the forelimb will also improve excursion distance for leaping as well as the arc produced during vault phase. It is important to stress that bats do not represent a unique solution to quadrupedal launching.


Procellariiform homologies
Original comments by D. Peters: “Certain pterosaurs, such as Nyctosaurus were indeed procellariiform-like in several aspects such as size, wing shape and leg length. Pteranodon was larger, but otherwise similar (Fig. 1). The problem developed for Witton and Habib [1] due to their invention or adoption of a wing shape that was not supported by the data [2] (discussed below)."

**Author response: Some pterosaurs had similar overall mass to procellariiform birds, but this is a relatively trivial comparison. Wing shape in Nyctosaurus is unknown, and while it was almost certainly a very high aspect ratio wing, we note that the body shape, head proportions, and hind limb strength are all markedly different between Nyctosaurus and procellariiform birds (for bone strength of albatrosses, see Ruff and Habib, 2008; Nyctosaurus hind limb strength currently in a set of unpublished data - contact authors directly for more information). Furthermore, the maximum lift coefficient and lift:drag ratio for Nyctosaurus was likely not similar to that of a procellariiform bird (Palmer, 2010). While we appreciate D. Peters suggesting comparison with additional pterosaur species beyond those in the manuscript, we find that direct analogy between pterosaurs and procellariiform birds is weak.


Wing proportions
Original comments by D. Peters: "Unfortunately the wing shaped created by Witton and Habib [1] is not supported in the fossil record. No specimens preserve a membrane attached to the lateral tibia [2, contra 1, 14]. All specimens that preserve a wing membrane indicate a shape stretched only between the elbow and wingtip with a short fuselage fillet medial to the elbow extending to the anterior thigh [2]. These were all ignored or overlooked in a prior study [14]. Reports [14] of “membrane shrinkage” are inventions and mistakes based on data that otherwise supports the Zittel [2] wing shape (Fig. 1). Some were based on torn and twisted membranes, which preserve the brachiopatagium in front of the wing finger! The Zittel wing shape [2] permits folding of the wing finger proximal to the elbow. It also permits a virtual disappearance of the membrane when folded, as fossils [14] indicate."

**Author response: for a comprehensive look at pterosaur membrane extent, we suggest that readers consult Elgin et al. (in press). The manuscript is also discussed and linked for free viewing here: http://pterosaur-net.blog... We understand that Peters disagrees with the current majority conclusion that the pterosaur wing attached to the hind limb. However, current evidence supports such a conclusion, and we cite several examples in the original paper. If evidence comes to light to change this conclusion, we will be happy to update our planform models. We appreciate that Peters takes a different view of a complex topic, but do note that Peters' model is currently preferred by him alone, and therefore we feel it is premature to use his wing model given the extensive literature to the contrary.


Wing shape
Original comments by D. Peters: "In dorsal view it is clear that pterosaurs were shaped like high-aspect ratio airplanes, complete with horizontal stabilizers created by the sprawling hind limbs (yet another trait arguing against an archosaur ancestry and supporting affinity with Sharovipteryx)."

**Author response: Again, as above, while we understand that pterosaur wing shape has been somewhat controversial in the past, the current literature does not support the model that Peters suggests. Regardless, the aspect ratio for Pteranodon is greater than that for Quetzalcoatlus, even in reconstructions where the wing is free of the hind limb. The sprawling hind limbs do not, as best we can tell, suggest a specific ancestry for pterosaurs, but regardless, they would be unlikely to act like the horizontal stabilizers on an aircraft, and instead, we observe that the uropatagia of pterosaurs would have been capable of producing powerful yawing moments that would improve maneuverability. The stabilizers on modern aircraft are largely related to the construction of a passively stable vessel; there is no evidence at present that flying animals are regularly passively stable in flight, though some large soaring birds may be passively stable in pitch (even then, the mechanism differs from that used in conventional aircraft; see Krus 1997). Readers interested in seeing examples of soft tissue preservation in pterosaurs can take a look at: http://pterosaur.net/foss...


Anhanguera forelimb illustration problems
Original comments by D. Peters: "Only the medial three fingers created known ichnites. Metacarpal IV was elevated off the substrate protecting the wing finger from ground contact (contra Witton and Habib [1]. In all pterosaurs fingers I-III retained their primitive orientation of palmar side ventral (in flight) or medial (when quadrupedal) with the antebrachia in the neutral position (neither supinated nor pronated). No pterosaur preserves fingers I-III on the dorsal surface of the wing as Witton and Habib [1] illustrated them. Specimens that preserve the slender metacarpals pressed against the anterior face of the large metacarpal did so post-mortem due to the fragile hinge at the metacarpal III-IV interface. Pterosaurs preserved dorsal side down may preserve fingers I-III on the ventral surface of metacarpal IV due to the same taphonomic hinge effect."

**Author response: We follow the reconstruction of Bennett (2003). If other reconstructions with similar rigor are later presented in the literature then we can update our mechanical model. Regardless, we note again that the elevation of Metacarpal IV from the substrate during walking is unclear, and our model only assumes that it could contact the ground during launch, which all anatomical data currently supports.

Original comments by D. Peters: "Figure 5 in Witton and Habib [1] (Fig. 2a) also appears to be much too heavily muscled. The slenderness of the antebrachial bones and the need to reduce weight for flight would tend to indicate a more gracile muscle set (Fig. 2c)."

**Author response: We understand that the musculature presented in Figure 5 of our manuscript may seem extreme compared to prior reconstructions. However, it is entirely consistent with the strength of the bone elements, known relationships between bone and muscle volume/mass in living vertebrates (Prange et al., 1979), and muscle attachment sites. We suspect that D. Peters may be unfamiliar with the bone to muscle relationships, in terms of volume and mass, within living tetrapods. When that comparative context is taken into account, the musculature we present is actually relatively slender given the breadth of the bones contained within the muscular compartments. However, because the long bones of Anhanguera were heavily pneumatized and expanded (as is the case in many other pterosaurs), the expectation is that the diameter of the bone should be relatively large compared to the muscular volume overlying it. A similar phenomena has been noted in sauropod necks; for a discussion readers can go here: http://svpow.wordpress.co...

We note that the issue of weight reduction for flight has been often misunderstood. For example, the skeletons of flying animals are not lighter than the skeletons of non-flying species (Prange et al., 1979; Dumont, 2010). Flying vertebrates also carry relatively large muscle masses (Pennycuick, 2008), rather than having much reduced musculature as is sometimes erroneously illustrated. Therefore, the comment by D. Peters that the muscles must be slender to reduce weight for flight is not consistent with the known mechanics of flying taxa.


References
Alexander DE. 2002. Nature's Flyers: Birds, Insects, and the Biomechanics of Flight. Johns Hopkins University Press, Baltimore. 358 pp.

Bennett SC. 2003. Morphological evolution of the pectoral girdle of pterosaurs myology and function, in Buffetaut, E., and Mazin, J-M. (eds) 2003, Evolution and Palaeobiology of Pterosaurs, Geological Society of London, 2003 217 pp. 191-215

Biewener. 2003. Animal Locomotion. Oxford University Press 294 pp.

Bonser RHC and Rayner, JVM. 1996. Measuring leg thrust forces in the common starling. J. Exp. Biol. 199: 435-439
Earls, K. D. 2000. Kinematics and mechanics of ground take-off in the starling Sturnis vulgaris and the quail Coturnix coturnix. Journal of Experimental Biology. 203: 725 -739
Lieber RL 2002. Skeletal Muscle Structure, Function & Plasticity. Lippincott Williams and Wilkins

Dumont. ER. 2010. Bone density and the lightweight skeletons of birds. Proceedings of The Royal Society B Biological Sciences DOI: 10.1098/rspb.2010.0117

Elgin, R.A., Hone, D.W.E and Frey, E. in press. The extent of the pterosaur flight membrane. Acta Palaeontologica Polonica 5X(X): xxx-xxx. doi: 10.4202/app.2009.0145

Heppner, FH. and Anderson JGT. 1985. Leg thrust important in flight take-off in the pigeon. J. Exp. Biol. 114: 285 -288

Krus P. 1997. Natural Methods for Flight Stability in Birds. AIAA Paper 97-5653

Lockley MG., Wright JL, Langston W. Jr. and West ES. 2001. New pterosaur track specimens from the Late Jurassic of Oklahoma and Colorado: their paleobiological significance and regional ichnological context. Modern Geology 20: 179–203

MacAyeal LC, Riskin DK, Swartz SM, and Breuer KS. (in press). Vertical flight performance and load carrying in lesser dog-faced fruit bats (Cynopterus brachyotis). Journal of Experimental Biology. Accepted 11 November, 2010

Marden JH. 1994. From damselflies to pterosaurs: how burst and sustainable flight performance scale with size. American Journal of Physiology, 266, 1077-1084

Mazin JM, Billon-Bruyat JP, and Padian K. 2009. First record of a pterosaur landing trackway. Proc. R. Soc. B doi:10.1098/rspb.2009.1161

Nachtigall and D.M. Wilson. 1967. Neuro-muscular control of dipteran flight. J. exp. Biol. 47: 77–97

Nachtigall. 1968. Elektrophysiologische und kinematische Untersuchungen über Start und Stop des Flugmotors von Fliegen. Z. vergl. Physiol. 61: 1-20

Nachtigall. 1978. Der Startsprung der Stubenfliege Musca domestica. Ent. Germ. 4: 368-373

Palmer C. 2010. Flight in slow motion: aerodynamics of the pterosaur wing Proceedings of the Royal Society B : 10.1098/rspb.2010.2179

Pennycuick CJ. 2008. Modelling the Flying Bird. Elsevier. 496 pp.

Prange HD, Anderson JF, and Rahn H. 1979. Scaling of skeletal mass to body mass in birds and mammals. Am. Nat. 113:103-122

Prondvai E, and Hone DWE. 2009. A model for wing extension in pterosaurs. Historical Biology, 20: 237-254

Riskin DK, Bahlman JW, Hubel TY Ratcliffe JM, Kunz TH, and Swartz SM. (2009). Bats go head-under-heels: the biomechanics of landing on a ceiling. Journal of Experimental Biology 212, 944-953

Schouest LP, Anderson M, Miller TA. 1986. The ultrastructure and physiology of the tergotrochanteral depressor muscle of the housefly, Musca domestic. J. Exp Zool. 239: 147-158

Schutt WA Jr, Altenbach JS, Chang YH, Cullinane DM, Hermanson JW, Muradali F, Bertram JE. 1997. The dynamics of flight-initiating jumps in the common vampire bat Desmodus rotundus. Journal of Experimental Biology. 200(Pt 23): 3003-12.

Tobalske, B.W., Altshuler, D. and Powers, D. 2004. Take-off mechanics in hummingbirds (Trochilidae). J. Exp. Biol. 207, 1345-1352

Trimarchi JR and Schneiderman AM. 1993. Giant fiber activation of an intrinsic muscle in the mesothoracic leg of Drosophila melanogaster. J. Exp. Biol. 177: 149-167

Trimarchi JR and Schneiderman AM. 1995. Initiation of flight in the unrestrained fly, Drosophila melanogaster. J. Zool. Lond. 235: 211-222

Unwin DM. 1997. Pterosaur tracks and the terrestrial ability of
pterosaurs. Lethaia, 29: 373–386.

Wagner HA. 1925. "Über die Entstehung des dynamischen Auftriebes von Tragflügeln", Zeitschrift für angewandte Mathematik und Mechanik 5: 17-35

Competing interests declared: Authors of the manuscript