95Vertebrate Anatomy Morphology Palaeontology 9:95–104 ISSN 2292-1389 Vertebrate Anatomy Morphology Palaeontology is an open access journal http://ejournals.library.ualberta.ca/index.php/VAMP Article copyright by the author(s). This open access work is distributed under a Creative Commons Attribution 4.0 International (CC By 4.0) License, meaning you must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use. No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits. INTRODUCTION Body mass is an important characteristic of organisms, as it relates to many important life functions such as metabolic rate (Strotz et al. 2018), relative maturity, and biomechan- ics (Hutchinson et al. 2011; Sander et al. 2011; Schmidt- Nielsen 1984). However, it is much harder to estimate body mass of extinct taxa due to a variety of factors including a lack of soft tissues, and taphonomic distortion of bones. A wide range of techniques have been developed to try to solve this conundrum, reviewed comprehensively by Brassey (2017) and Campione and Evans (2020). Currently, there are two major categories into which mass estimation tech- niques fall - volumetric mass estimates, that use various Constraining the body mass range of Anzu wyliei (Theropoda: Caenagnathidae) using volumetric and extant–scaling methods Kyle L. Atkins-Weltman1*, Eric Snively2 , and Patrick O’Connor 3,4 1660 Gateway Court, Apt N2, Lawrence, KS, 66049, USA; flarginblarg@gmail.com 2OSU College of Osteopathic Medicine at the Cherokee Nation, 19500 E. Ross Street, Tahlequah, OK ,74464, USA; eric.snively@okstate.edu 3Department of Biomedical Sciences, 228 Irvine Hall, Ohio University Heritage College of Osteopathic Medicine, Athens, OH, 45701, USA; oconnorp@ohio.edu 4Ohio Center for Ecological and Evolutionary Studies, Irvine Hall, Ohio University, Athens, Ohio, 45701, USA Published September 28, 2021 *corresponding author. © 2021 by the authors; submitted July 8, 2021; revisions received September 7, 2021; accepted September 7, 2021. Handling editor: Jordan Mallon. DOI 10.18435/vamp29375 Abstract: The ability to accurately and reliably estimate body mass of extinct taxa is a vital tool for inter- preting the physiology and even behavior of long-dead animals. For this reason, paleontologists have de- veloped many possible methods of estimating the body mass of extinct animals, with varying degrees of success. These methods can be divided into two main categories: volumetric mass estimation and extant scaling methods. Each has advantages and disadvantages, which is why, when possible, it is best to perform both, and compare the results to determine what is most plausible within reason. Here we employ volumetric mass estimation (VME) to calculate an approximate body mass for previously described specimens of Anzu wyliei from the Carnegie Museum of Natural History. We also use extant scaling methods to try to obtain a reliable mass estimate for this taxon. In addition, we present the first digital life restoration and convex hull of the dinosaur Anzu wyliei used for mass estimation purposes. We found that the volumetric mass estimation using our digital model was 216–280 kg, which falls within the range predicted by extant scaling techniques, while the mass estimate using minimum convex hulls was below the predicted range, between 159–199 kg. The VME method for Anzu wyliei strongly affirms the predictive utility of extant-based scaling. However, volumetric mass estimates are likely more precise because the models are based on comprehensive specimen anatomy rather than regressions of a phylogenetically comprehensive but disparate sample. ways of estimating body volume and density, and extant scaling methods, which use relationships between measured osteological characters and body mass in modern taxa, and attempt to reconcile these relationships with extinct organ- isms. Each method has advantages and drawbacks, which is why using both can be informative (Campione and Evans 2020) for identifying potential errors in one method or the other and to provide what may be a more realistic range of values, and for tradeoffs of comparative sample size versus time investment. We wanted to determine whether volumetric mass estima- tion could create a narrower range of body mass estimates than extant scaling for a taxon such as Anzu wyliei based on known specimens, using a class of techniques known Vertebrate Anatomy Morphology Palaeontology 9:95–104 96 as volumetric mass estimation (or VME), as well as using expanded extant scaling methods than those used in the original description (Lamanna et al. 2014). Many differ- ent methods have been used to estimate body mass from volume. Such methods have included using a scaled-down physical model of the animal of interest (Alexander 1985; Colbert 1962; Gregory 1905), and more recent techniques include 3D mathematical slicing (Henderson 1999; Snively et al. 2019), photogrammetry and parameterized comput- er modeling (Bates et al. 2009; Gunga et al. 2008, 1995; Hutchinson et al. 2007), and using minimum convex hulls to wrap around a digitized skeletal frame (Brassey and Sellers 2014; Sellers et al. 2012). See Brassey (2017) and Campione and Evans (2020) for a thorough review of each of these approaches. Current published mass estimates for Anzu wyliei yield a range of 200‒300 kg (Lamanna et al. 2014), based both on femoral circumference (Anderson et al. 1985) and femoral length measures (Christiansen and Fariña 2004; Zanno and Makovicky 2013). However, since the original publication more refined techniques for estimating body mass using stylopodial circumference have been developed that poten- tially allow for more rigor and further have the added benefit of yielding confidence ranges even for single-point estimates (Campione et al. 2014; Campione and Evans 2020). Such methods also do not resort to using multiple different techniques of mass estimation. Here, we derive new, updated mass estimates utilizing extant scaling methods developed by Campione et al. (2014) and further refined by Campione and Evans (2020). We then compare these to volumetric mass estimates obtained by digital scanning of Anzu speci- mens as well as manual digital modeling, similar to proced- ures presented by Romano et al. (2021) for the pareiasaur Scutosaurus. Furthermore, we estimated the animal’s mass with minimum convex hulling, to see which of the applied methods yielded a range that was more congruent with that generated from the extant scaling method. MATERIALS AND METHODS Volumetric Mass Estimation To conduct volumetric mass estimation, multiple elements of the paratype specimen of Anzu wyliei (CM 78001) and a second caenagnathid from the Hell Creek Formation (cur- rently being described) were digitized using photogram- metry. Digital photographs were captured with a Canon EOS Rebel T3i DSLR, with point clouds derived from the photos using Agisoft Metashape. Imperfections in the resulting meshes were corrected manually in ZBrush 2020. To minimize unnecessary corrections related to taphonomic distortion, only the more complete and/or better preserved of paired elements was digitized, with the exception of the femur. Elements of CM 78001 digitized using photogram- metry included the left ilium, left pubis, right ischium, both femora, left tibia, and right fibula. The elements of the second caenagnathid (CM 96523) that were digitized included the left metatarsal III and the right metatarsal IV. While not pertaining to Anzu wyliei, these elements were scaled to appropriate proportions using data from a cur- rently unpublished specimen (M. Lamanna, pers. comm. 2020). 3D scans of a series of presacral vertebrae from both the holotype (CM 78000) and referred specimen (CM 78001) were provided by L. Roberts. To surmount time constraints, the rest of the skeleton was not directly digitized using photogrammetry – instead, the remaining elements were manually sculpted using reference images of preserved elements from anterior, posterior, lateral, dorsal, and ventral perspectives. Manually sculpted elements included cervical and dorsal ribs, sacral and caudal vertebrae and chevrons, the entire pectoral girdle and forelimbs, and the proximal phalanges and unguals of the pes (Fig. 1). An approximate life-restoration of Anzu wyliei was first constructed in ZBrush 2020 using the skeletal outline drawings created by Scott Hartman as a starting point, though the tail of the life restoration was straightened to make it easier to match the articulation of the caudal ver- tebrae once they were digitized. The skeletal reconstruction represents a sufficiently accurate starting point as it was used to illustrate in the original description (Lamanna et al. 2014). The digitized skeleton was then placed within this life restoration to mimic the position of the skeleton within the living organism, with corrections made to fit the skel- eton where necessary (Fig. 2). To reconstruct the major pul- monary tissues, Dynamesh spheres were imported into the model space, then manually sculpted using Inflate, Move, and Dynamesh tools to shape into the lungs, trachea, and air sacs (Fig. 1). To calculate volume and subsequently body mass, the model was exported to 3DS Max 2021, where the volume of the entire life model was calculated in addition to the volume of the lungs and air sacs. The composition, pos- ition, and estimated arrangement of lungs and air sacs were constrained using osteological correlates for air sac invasion and placement in theropods (O’Connor 2006; Sereno et al. 2008; Wedel 2006) (Fig. 1) – specifically, lungs, cervical and abdominal airs sacs, and pneumatic diverticula (e.g., lateral cervical diverticula) were modeled along the cervical and dorsal vertebral series (Benson et al. 2012; O’Connor 2006; O’Connor and Claessens 2005). Since there are no well constrained osteological correlates for thoracic and clavicular air sacs, we utilized two calculations of body mass. First, we did a calculation using the most conserva- tive model, with only the cervical air sacs, lungs, abdominal air sacs, and trachea present. Second was a more exten- Atkins-Weltman et al. — Body mass range of Anzu wyliei 97 Figure 1. Digitized skeleton of Anzu wyliei with air sacs (“avian” model). Model shown in A, lateral; B, dorsal; C, ventral; D, anterior; and E, posterior views. Key: Green: cervical air sacs; orange: lungs; blue: abdominal air sacs; purple: thoracic air sacs; aquamarine: clavicular air sacs; pink: trachea. NOTE: Clavicular and thoracic air sacs were removed in the more conservative reconstruction. Vertebrate Anatomy Morphology Palaeontology 9:95–104 98 Figure 2. Life restoration of Anzu wyliei, with skeleton and air sacs semi-visible through transparency. Model shown in A, lateral; B, dorsal; C, ventral; D, anterior; and E, posterior views. Lungs and air sacs as in Figure 1. Atkins-Weltman et al. — Body mass range of Anzu wyliei 99 sive air sac system (an “avian model”) that included both anterior and posterior thoracic air sacs, and a clavicular air sac, as are present in Avialae (Benson et al. 2012; Duncker 1971; O’Connor 2004, 2006). In this way we account for the uncertainty in presence or absence of such air sacs, and importantly, determine how modeling them affects our estimate of body mass. Since the exact and overall tissue densities of extinct or- ganisms are impossible to calculate with certainty, we used two different values for density of the body excluding pul- monary air space to create an upper and lower bound for the model. The lower bound used an estimated density of 800 kg/m3, as has been applied to large sauropods because the pneumaticity in their bones was assumed to lower their body density (Gunga et al. 2008), an issue that has been brought up for highly pneumatic saurischian taxa (Benson et al. 2012; Brassey and Sellers, 2014; Campione and Evans 2020). However, increasing skeletal pneumaticity does not appear to change the total mass of the bones relative to whole body mass—the skeletons of highly pneumatic birds weigh the same relative to total body mass as less pneumat- ic birds (Martin-Silverstone et al. 2015; Prange et al. 1979). This could mean that skeletal pneumaticity probably does not directly have an effect on whole body density. The negative relationship between density and body mass in birds is probably related to an increase in size of the air sacs relative to total body volume (S. Gutherz pers. comm.). Thus, to allow for these uncertainties, we used a value of 1000 kg/m3, which has been used in many previous studies aiming to estimate body mass through volumetric meth- ods (Bates et al. 2009; Henderson 1999; Hutchinson et al. 2007, 2011). The non-respiratory value of 1000 kg/m3 is closer to recent estimates for extant and extinct saurischian dinosaurs (Larramendi et al. 2020). Since previous researchers have used convex hulling meth- ods to estimate body mass of both extant and extinct taxa (Brassey and Sellers 2014; Sellers et al. 2012), we also used this method with Anzu for comparative purposes. To con- struct minimal convex hulls, first the skeleton was exported from 3DS Max into MeshLab. The skeleton was divided into multiple segments, each given its own convex hull (Fig. 3), with each hull used to estimate volume. Because exact scaling parameters did not transfer between programs, the Transform:Scale:Normalize function we employed to scale the skeleton as close as possible to the known actual size of the animal. Furthermore, MeshLab’s volume output was in cm3, and thus we converted these volumes to m3. The total volumes of all segments (Tab. 1) were summed together and multiplied by the two extremes of density used for the manually modeled estimate, to establish a range of body masses. While feathers are known to be present in oviraptorosaurs based on direct preservation (Funston and Currie 2020; Qiang et al. 1998; Xu et al. 2010; Zhou et al. 2000), in addition to indirect inference based on quill knobs on the ulna (Kurzanov 1982), in many extant birds, feathers do not contribute significantly to overall body mass (Brassey and Sellers 2014; Hopps 2002; Larramendi et al. 2020; Table 1. Volumes of the convex hulls used to generate a minimum convex hull estimate of body mass for Anzu wyliei. Body segment Volume (cm3) Volume (m3) Mass (800 kg/m3 ) Mass (1000 kg/m3 ) Torso 125274.625 0.125274625 100.2197 kg 125.274625 kg Tail 12802.94043 0.01280294 10.24235234 kg 12.80294043 kg Skull 10121.46387 0.010121464 8.097171094 kg 10.12146387 kg Neck 12408.58691 0.012408587 9.926869531 kg 12.40858691 kg L humerus 797.968811 0.000797969 0.638375049 kg 0.797968811 kg L antebrachium 464.968689 0.000464969 0.371974951 kg 0.464968689 kg L manus 1855.232666 0.001855233 1.484186133 kg 1.855232666 kg R humerus 744.117065 0.000744117 0.595293652 kg 0.744117065 kg R antebrachium 466.130646 0.000466131 0.372904517 kg 0.466130646 kg R manus 4343.450684 0.004343451 3.474760547 kg 4.343450684 kg L stylopod 4343.450684 0.004343451 3.474760547 kg 4.343450684 kg L zeugopod 2814.483398 0.002814483 2.251586718 kg 2.814483398 kg L autopod 7956.745605 0.007956746 6.365396484 kg 7.956745605 kg R stylopod 4343.51416 0.004343514 3.474811328 kg 4.34351416 kg R zeugopod 2792.091309 0.002792091 2.233673047 kg 2.792091309 kg R autopod 7956.643555 0.007956644 6.365314844 kg 7.956643555 kg Total 199486.4135 cm3 0.199486413 m3 159.5891308 kg 199.4864135 kg L = left; R = right Vertebrate Anatomy Morphology Palaeontology 9:95–104 100 RESULTS Volumetric Mass Estimation The volume of the complete life model was 0.30 m3 (Tab. 2). The air sac volume differed between a more conservative (i.e., lungs and other pulmonary structures modeled at 0.02 m3) and less conservative (i.e., lungs and other pulmonary structures modeled at 0.03 m3). From these results, the conservative model is heavier at an estimated 224-280 kg, whereas the more speculative model is slightly lighter, ran- ging from 216–270 kg. Minimum convex hulling yielded a volume of 0.199 m3 (Tab. 1), with a mass estimate between 159 and 199 kg, depending on which body density value was applied. Extant Scaling Methods Because the femoral circumference of CM 78001was not available, we were only able to use femoral circumference of the Anzu holotype (CM 78000) for the extant scaling re- gression. However, while the two differ slightly in size, the difference is slight enough that the obtained femoral cir- cumference is a somewhat reasonable proxy for CM 78001 (M. Lamanna, pers. comm. 2021), although see Table 3 for comparative measurements. The resulting regression yielded a body mass range of between 202 and 342 kg (Fig. 4), whereas the original point estimate based on femoral circumference was 193 kg (Lamanna et al. 2014). DISCUSSION The results obtained from volumetric mass estimates of Anzu wyliei fit within both the original mass range proposed in the original description (Lamanna et al. 2014) and that predicted by the corroboration plot using Campione’s method (Campione et al. 2014; Campione and Evans 2012). Thus, current volumetric mass estimate of 216-280 kg provides a more precise range of body masses for this organism than those that use extant data alone (Tab. 2). Wecke et al. 2017). In large, extant ratites, feathery integu- ment comprises less than 2% of total body mass (Brassey and Sellers 2014). For this reason, we decided to construct our volumetric model without feathers, and infer that feathers would add up to an additional 2% to the volumet- ric body mass estimates reported in this paper. Extant Scaling Methods We provided data on femoral circumference of the holotype Anzu wyliei (CM 78000) to Dr. N. Campione, to use in his body mass regression analyses as has been done in previous works (Campione et al. 2014; Campione and Evans 2020) (Fig. 4). We also compared these to the values originally obtained by Lamanna et al. (2014), as using femoral length as an estimator could still be useful because actual femoral circumference can be easily distorted by taphonomic factors and be made impossible to measure. Figure 3. Convex Hull model of Anzu wyliei. Model shown in A, lateral; B, dorsal; C, ventral; D, anterior; and E, posterior views. Table 2. Mass estimates resulting from different methods of estimation. Note the increasing precision relative to earlier studies provided by the results presented herein. Further note the discrepancy of the min- imum convex hull estimate relative to all other estimates. Source: Lamanna et al. 2014 Nicolás Campione This paper This paper Estimation Type: Femoral length Log stylopodial circumference VME from digitally VME from minimum regression reconstructed model convex hulls Result: 200–300 kg 202–342 kg 216–280 kg 159–199 kg Atkins-Weltman et al. — Body mass range of Anzu wyliei 101 We also found that the minimum convex hulling method seems to lead to an underestimation of total body mass in this case, though this could be due to scaling errors when transferring between programs. However, if this is indeed a reflection upon minimum convex hulling as a method, it appears when comparing the convex hull model to the manually sculpted life restoration that much of the missing body mass relates to the extremely low volume of both the forelimbs and hind limbs using the hull approach (Figs. 2, 3). The convex hull model does not account for the M. iliotbialis, M. iliofibularis, or many of the other mus- cles connecting the pelvic bones to the femur, or the M. caudofemoralis spanning between the tail and the thigh. The same is true for many of the muscles in the upper forelimb. This could explain the much lower estimate and suggests that aspects of the convex hull method may make implaus- ible biological assumptions. Feathers would add less than 2% (Brassey and Sellers 2014) to our estimate of the body mass of Anzu, regardless of method. Moreover, it seems likely that with convex hulling, the choice of which elements to include in a particular convex hull may have a strong influence on the final result. Perhaps if the pelvic bones were included in the same hull as the femur, for example, the resulting hull would have added the muscles to the proximal hind limb. Yet, even if this change were made, there is still the issue of the incredibly thin zeugopodia of both forelimbs and hind limbs, largely underrepresenting both the gastrocnemius and, to a lesser extent, the antebrachial muscles. However, whereas these areas are clearly given less volume than they would ac- count for in life, both the head and the autopodia of both forelimbs and hind limbs account for a greater volume than they would in life. This is because separate elements are often hulled together - since the skull is modeled as a single element despite consisting of an upper and lower jaw (and the mouth being open in articulation), the hull included the gap between the jaws in its total volume which is bio- logically inaccurate. Similarly, since the individual fingers and toes were not added as separate hulls, the generated hull connected them in a single, wide structure (e.g., like a duck’s foot), which is not correct. Despite these three areas of greater-than-expected volume, they are not sufficient to balance out the underestimation of volume in other regions, resulting in a lower-than-expected body volume regardless of the density value selected. Whereas the volumetric mass estimate using a manually constructed life restoration has worked well to narrow the plausible range of body masses for Anzu wyliei, we continue to argue for the use of integrated methods making use of both volumetric and extant scaling methods when possible. Figure 4. Logistic regression of stylopodial circumference versus body mass in dinosaurs, and residual deviance. Red lines represent the upper and lower bounds of the 95% confidence interval. The pink dots represent the upper and lower esti- mates for CM 78000 based on femoral circumference. Gray “x” marks represent other specimens of various taxa used to create the regression. Regressions and figure courtesy Nicolás Campione. Vertebrate Anatomy Morphology Palaeontology 9:95–104 102 Volumetric body mass is only possible for taxa known from relatively complete remains (Campione and Evans 2020), which greatly reduces its utility in the vertebrate fossil record because many taxa are known only from incomplete skeletons. Furthermore, there are still great unknowns about body density, lung and air sac size and structure, and other soft tissue systems that no doubt convey sources of variation within volumetric mass estimates if not careful- ly accounted for a priori. This may be done by creating more than one model, to account for differing amounts of soft tissue as has been implemented by some researchers (Hutchinson et al. 2011), or as we have done here, simply by using a differing possible body density in the same mod- el. We chose the latter, as the relatively small total volume of the model (0.30 m3) and the sensitivity of the software calculating the volume (± 0.01 m3) meant that it would take the addition or subtraction of an immense amount of extra soft tissue relative to the model size to modify the body mass of the model by more than 10–20 kg in either direction. However, with larger taxa, even seemingly small changes may register due to the larger total volume, and thus, the smaller proportion of total volume 0.01 m3 rep- resents. For this reason, we suggest that anyone attempting to replicate this method with larger taxa should use a lower sensitivity to volume changes, perhaps using this study as an approximation of what proportion of total body volume to use for the sensitivity. For example, for an animal with a body volume of 6.0 m3, the sensitivity would need to be twenty times lower (i.e., ± 0.20 m3). Researchers can use this in conjunction with a maximum and minimum body density to create a range of body mass estimates using body volume (Campione and Evans 2020). We reiterate that Table 3. Comparative measurements of skeletal elements preserved in both CM 78000 and CM 78001, as reported in Lamanna et al. 2014. As reported in the original description, measurements greater than 205 mm were taken with tape meas- ure and are therefore less precise, and provided only to the nearest 5 mm. Element/dimension CM 78000 CM 78001 Skull and mandible Braincase Height, occipital condyle midline, dorsoventral 10.4 12.3 Width of occipital condyle, transverse 20.9 21.7 Height of foramen magnum, dorsoventral 18.0 15.3 Width of foramen magnum, transverse 15.1 13.1 Femur Length, proximodistal 525R 505L, 500R Width of proximal end, mediolateral 121.7L*, 127.4R 136.3L, 135.5R Width of distal end, mediolateral 94.1L*, 112.4R 110.0R Tibia Length, proximodistal 660L 595L Width of proximal end, mediolateral 93.5L, 101.7R* 86.1L Depth of proximal end, anteroposterior 104.6L 88.4L Width of distal end, mediolateral 101.7L 110.3L Fibula Length, proximodistal 585L, 580R 570L Width of proximal end, anteroposterior 70.0L, 65.0R 74.7L, 67.9R Width of proximal end, mediolateral 29.2L, 27.5R 25.0L, 31.8R Astragalocalcaneum Length, proximodistal 133.1L*, 141.8*R 220L Width across distal condyles, mediolateral 98.5L, 98.0R 109.7L Metatarsal V Depth of proximal end, anteroposterior 16.0L* 20.7L Abbreviations: aofe, antorbital fenestra; L, left; R, right; *, element incomplete, measurement as preserved Atkins-Weltman et al. — Body mass range of Anzu wyliei 103 these caveats apply to the precision of the specific software used, which yields lower relative error with larger volumes and greater relative error with smaller volumes. Software with greater precision, or consistent relative precision regardless of volume, will yield more equivalent relative precision with large and small volumes. CONCLUSIONS Using carefully sculpted digital models based on actual specimens, along with modeling a range of possible body densities, allows for a more realistic and accurate range of possible body masses than extant scaling alone. However, we note that this is only possible when specimens are sufficiently complete and well-known enough to reliably infer basic soft tissue anatomy, and as such, extant scaling is still critical for providing a bracket of reasonable values against which to compare those estimated by volumetric methods. Further, we show that there is a great sensitivity of minimum convex hulling to the selection of elements within the hull, making it difficult to determine reliability on animals with lower estimated body sizes. ACKNOWLEDGEMENTS We would like to thank Matt Lamanna and Ami Henrici of the Carnegie Museum of Natural History (Pittsburgh, PA, USA) for allowing access to the original Anzu wyliei type material, and Heinrich Mallison for providing ad- vice on photogrammetry. We also thank Lucy Roberts for providing scans of Anzu wyliei presacral vertebrae, Evelyn Volmer for sculpting missing skeletal elements and articu- lating the skeleton, and Emma Schachner, who provided photographs of the axial skeleton of the CM Anzu speci- mens. Scott Hartman’s skeletal diagrams proved invaluable for sculpting the initial basis of the life restoration and in- forming basic articulation of the skeletal elements. Nicolás Campione deserves acknowledgement for his assistance with extant-scaling regressions. Lastly, we acknowledge Samuel Gutherz for additional input regarding the air sac reconstruction. LITERATURE CITED Alexander, R.M. 1985. Mechanics of posture and gait of some large dinosaurs. Zoological Journal of the Linnean Society 83:1–25. https://doi.org/10.1111/j.1096-3642.1985.tb00871.x Anderson, J.F., A. Hall‐Martin, and D.A. Russell. 1985. Long-bone circumference and weight in mammals, birds and dinosaurs. Journal of Zoology 207:53–61. https://doi. org/10.1111/j.1469-7998.1985.tb04915.x Bates, K.T., P.L. Manning, D. Hodgetts, and W.I. Sellers. 2009. Estimating mass properties of dinosaurs using laser imaging and 3D computer modelling. PLoS ONE 4:e4532. https://doi. org/10.1371/journal.pone.0004532 Benson, R.B.J., R.J. Butler, M.T. Carrano, and P.M. O’Connor. 2012. Air-filled postcranial bones in theropod dinosaurs: physio- logical implications and the 'reptile'-bird transition. Biological Reviews of the Cambridge Philosophical Society 87:168–193. https://doi.org/10.1111/j.1469-185X.2011.00190.x Brassey, C.A., and W.I. Sellers. 2014. Scaling of convex hull vol- ume to body mass in modern primates, non-primate mammals and birds. PLoS ONE 9:e91691. https://doi.org/10.1371/jour- nal.pone.0091691 Campione, N.E., and D.C. Evans. 2012. A universal scaling re- lationship between body mass and proximal limb bone dimen- sions in quadrupedal terrestrial tetrapods. BMC Biology 10:60. https://doi.org/10.1186/1741-7007-10-60 Campione, N.E., and D.C. Evans. 2020. The accuracy and preci- sion of body mass estimation in non-avian dinosaurs. Biological Reviews 95:1759–1797. https://doi.org/10.1111/brv.12638 Campione, N.E., D.C. Evans, C.M. Brown, and M.T. Carrano. 2014. Body mass estimation in non-avian bipeds using a theoretical conversion to quadruped stylopodial proportions. Methods in Ecology and Evolution 5:913–923. https://doi. org/10.1111/2041-210X.12226 Christiansen, P., and R.A. Fariña. 2004. Mass prediction in theropod dinosaurs. Historical Biology 16:85–92. https://doi.or g/10.1080/08912960412331284313 Colbert, E.H. 1962. The weights of dinosaurs. American Museum Novitates 2076:1–16. Duncker, H.R. 1971. The lung air sac system of birds. A contri- bution to the functional anatomy of the respiratory apparatus. Ergebnisse der Anatomie und Entwicklungsgeschichte 45:7–171. Funston, G.F., and P.J. Currie. 2020. New material of Chirostenotes pergracilis (Theropoda, Oviraptorosauria) from the Campanian Dinosaur Park Formation of Alberta, Canada. Historical Biology 15 pp. https://doi.org/10.1080/08912963.2 020.1726908 Gregory, W.K. 1905. The weight of the Brontosaurus. Science 22:572–572. https://doi.org/10.1126/science.22.566.572 Gunga, H.-C., T. Suthau, A. Bellmann, S. Stoinski, A. Friedrich, T. Trippel, K. Kirsch, and O. Hellwich. 2008. A new body mass estimation of Brachiosaurus brancai Janensch, 1914 mount- ed and exhibited at the Museum of Natural History (Berlin, Germany). Fossil Record 11:33–38. https://doi.org/10.1002/ mmng.200700011 Gunga, H.-Chr., K.A. Kirsch, F. Baartz, L. Röcker, W.-D. Heinrich, W. Lisowski, A. Wiedemann, and J. Albertz. 1995. New data on the dimensions of Brachiosaurus brancai and their physiological implications. Naturwissenschaften 82:190–192. https://doi.org/10.1007/BF01143194 Henderson, D.M. 1999. Estimating the masses and centers of mass of extinct animals by 3-D mathematical slicing. Paleobiology 25:88–106. Vertebrate Anatomy Morphology Palaeontology 9:95–104 104 Hopps, E.C. 2002. Information on waterfowl feather charac- teristics. Transactions of the Illinois State Academy of Science 95:229–237. Hutchinson, J.R., K.T. Bates, J. Molnar, V. Allen, and P.J. Makovicky. 2011. A computational analysis of limb and body dimensions in Tyrannosaurus rex with implications for loco- motion, ontogeny, and growth. PLoS ONE 6 e26037. https:// doi.org/10.1371/journal.pone.0026037 Hutchinson, J.R., V. Ng-Thow-Hing, and F.C. Anderson. 2007. A 3D interactive method for estimating body segmental par- ameters in animals: application to the turning and running per- formance of Tyrannosaurus rex. Journal of Theoretical Biology 246:660–680. https://doi.org/10.1016/j.jtbi.2007.01.023 Kurzanov, S. 1982. Structural characteristics of the fore limbs of Avimimus. Paleontological Journal 1982:108–112. Lamanna, M.C., H.-D. Sues, E.R. Schachner, and T.R. Lyson. 2014. A new large-bodied oviraptorosaurian theropod dinosaur from the latest Cretaceous of western North America. PLoS ONE 9:e92022. Larramendi, A., G.S. Paul, and S. Hsu. 2020. A review and reappraisal of the specific gravities of present and past multi- cellular organisms, with an emphasis on tetrapods. Anatomical Record 2021:1–56. https://doi.org/10.1002/ar.24574 Martin-Silverstone, E., O. Vincze, R. McCann, C.H.W. Jonsson, C. Palmer, G. Kaiser, and G. Dyke. 2015. Exploring the rela- tionship between skeletal mass and total body mass in birds. PLoS ONE 10:e0141794. https://doi.org/10.1371/journal. pone.0141794 O’Connor, P.M. 2006. Postcranial pneumaticity: an evaluation of soft-tissue influences on the postcranial skeleton and the reconstruction of pulmonary anatomy in archosaurs. Journal of Morphology 267:1199–1226. O’Connor, P.M. 2004. Pulmonary pneumaticity in the post- cranial skeleton of extant aves: a case study examining Anseriformes. Journal of Morphology 261:141–161. O’Connor, P.M., and L.P.A.M. Claessens. 2005. Basic avian pulmonary design and flow-through ventilation in non- avian theropod dinosaurs. Nature 436:253–256. https://doi. org/10.1038/nature03716 Prange, H.D., J.F. Anderson, and H. Rahn. 1979. Scaling of skel- etal mass to body mass in birds and mammals. The American Naturalist 113:103–122. https://doi.org/10.1086/283367 Qiang, J., P.J. Currie, M.A. Norell, and J. Shu-An. 1998. Two feathered dinosaurs from northeastern China. Nature 393:753– 761. https://doi.org/10.1038/31635 Sander, P.M., A. Christian, M. Clauss, R. Fechner, C.T. Gee, E.-M. Griebeler, H.-C. Gunga, J. Hummel, H. Mallison, S.F. Perry, H. Preuschoft, O.W.M. Rauhut, K. Remes, T. Tütken, O. Wings, and U. Witzel. 2011. Biology of the sauropod dinosaurs: the evolution of gigantism. Biological reviews of the Cambridge Philosophical Society 86:117–155. https://doi. org/10.1111/j.1469-185X.2010.00137.x Schmidt-Nielsen, K. 1984. Scaling: Why is Animal Size so Important? Cambridge University Press, Cambridge, 256 pp. https://doi.org/10.1017/CBO9781139167826 Sellers, W.I., J. Hepworth-Bell, P.L. Falkingham, K.T. Bates, C.A. Brassey, V.M. Egerton, and P.L. Manning. 2012. Minimum convex hull mass estimations of complete mounted skel- etons. Biology Letters 8:842–845. https://doi.org/10.1098/ rsbl.2012.0263 Sereno, P.C., R.N. Martinez, J.A. Wilson, D.J. Varricchio, O.A. Alcober, and H.C.E. Larsson. 2008. Evidence for avian intra- thoracic air sacs in a new predatory dinosaur from Argentina. PLoS ONE 3:e3303. https://doi.org/10.1371/journal. pone.0003303 Snively, E., H. O’Brien, D.M. Henderson, H. Mallison, L.A. Surring, M.E. Burns, T.R. Holtz Jr., A.P. Russell, L.M. Witmer, P.J. Currie, S.A. Hartman, and J.R. Cotton. 2019. Lower rota- tional inertia and larger leg muscles indicate more rapid turns in tyrannosaurids than in other large theropods. PeerJ 7:e6432. https://doi.org/10.7717/peerj.6432 Strotz, L.C., E.E. Saupe, J. Kimmig, and B.S. Lieberman. 2018. Metabolic rates, climate and macroevolution: a case study using Neogene molluscs. Proceedings of the Royal Society B, Biological Sciences 285:20181292. https://doi.org/10.1098/ rspb.2018.1292 Wecke, C., D.R. Khan, A. Sünder, and F. Liebert. 2017. Age and gender depending growth of feathers and feather-free body in modern fast growing meattype chickens. Open Journal of Animal Sciences 7:376–392. Wedel, M.J. 2006. Origin of postcranial skeletal pneumaticity in dinosaurs. Integrative Zoology 1:80–85. https://doi.org/10.1111/ j.1749-4877.2006.00019.x Xu, X., X. Zheng, and H. You. 2010. Exceptional dinosaur fossils show ontogenetic development of early feathers. Nature 464:1338–1341. https://doi.org/10.1038/nature08965 Zanno, L.E., and P.J. Makovicky. 2013. No evidence for direc- tional evolution of body mass in herbivorous theropod dino- saurs. Proceedings of the Royal Society, Biological Sciences 280:1–8. Zhou, Z.-H., X.-L. Wang, F.-C. Zhang, and X. Xu. 2000. Important features of Caudipteryx - evidence from two nearly complete new specimens. Vertebrata PalAsiatica 38:242–254.