Report. The Earliest Pterodactyloid and the Origin of the Group. Brian Andres, 1, * James Clark, 2 and Xing Xu 3 1

Transcription

1 Current Biology 24, , May 5, 2014 ª2014 Elsevier Ltd All rights reserved The Earliest Pterodactyloid and the Origin of the Group Report Brian Andres, 1, * James Clark, 2 and Xing Xu 3 1 School of Geosciences, University of South Florida, Tampa, FL 33620, USA 2 Department of Biological Sciences, George Washington University, Washington, DC 10024, USA 3 Key Laboratory of Vertebrate Evolution and Human Origins, Institute of Vertebrate Paleontology and Paleoanthropology, Chinese Academy of Sciences, Beijing , China Summary The pterosaurs were a diverse group of Mesozoic flying reptiles that underwent a body plan reorganization, adaptive radiation, and replacement of earlier forms midway through their long history, resulting in the origin of the Pterodactyloidea, a highly specialized clade containing the largest flying organisms. The sudden appearance and large suite of morphological features of this group were suggested to be the result of it originating in terrestrial environments, where the pterosaur fossil record has traditionally been poor [1, 2], and its many features suggested to be adaptations to those environments [1, 2]. However, little evidence has been available to test this hypothesis, and it has not been supported by previous phylogenies or early pterodactyloid discoveries. We report here the earliest pterosaur with the diagnostic elongate metacarpus of the Pterodactyloidea, Kryptodrakon progenitor, gen. et sp. nov., from the terrestrial Middle-Upper Jurassic boundary of Northwest China. Phylogenetic analysis confirms this species as the basalmost pterodactyloid and reconstructs a terrestrial origin and a predominantly terrestrial history for the Pterodactyloidea. Phylogenetic comparative methods support this reconstruction by means of a significant correlation between wing shape and environment also found in modern flying vertebrates, indicating that pterosaurs lived in or were at least adapted to the environments in which they were preserved. Results and Discussion The Pterodactyloidea are a clade of elongate-metacarpus, short-tailed pterosaurs [3] that appeared in the middle Late Jurassic (Kimmeridgian) and quickly became the most diverse pterosaur group, replacing all others [1]. The reorganization of their body plan was so extensive that a consensus has not been reached on its basal relationships, or how to define its name based on its relationships [4, 5]. An apomorphy-based definition was therefore proposed based on the elongation of the metacarpus that is found only in the pterodactyloid pterosaurs [6] and is the only original character [3] still diagnostic for the clade. A new pterosaur species, Kryptodrakon progenitor, gen. et sp. nov., from Northwest China is the earliest and most basal pterosaur to bear this diagnostic apomorphy, extending the fossil record of the Pterodactyloidea by at least 5 megaannum (Ma) to the Middle-Upper Jurassic boundary. *Correspondence: brianandres@usf.edu In previous phylogenetic analyses, the Pterodactyloidea have traditionally had the longest branch in terms of apomorphy number [5] and temporal duration of gaps without fossils implied by phylogeny [4]. This long branch was hypothesized to be the result of the pterodactyloids originating in terrestrial environments [1, 2], where the pterosaur fossil record has historically been undersampled [1]. Due in part to the fragile construction of their skeletons, pterosaurs have been predominantly preserved in quiet marine environments [7], to such an extent that they were hypothesized to have been an almost exclusively marine group [8, 9]. A number of terrestrial deposits from China have yielded new pterosaurs in recent years [10, 11], but when preservation in terrestrial or marine environments is reconstructed for previous pterosaur phylogenies, they still support a predominantly marine history for the pterosaurs and a marine origin of the Pterodactyloidea (see Table S1 available online). Here we show that comprehensive phylogenetic and comparative analysis of the Pterosauria and Kryptodrakon supports a terrestrial origin and a predominantly terrestrial history for the Pterodactyloidea. In support of this reconstruction, we provide a novel test of the quality of the fossil record and demonstrate that pterosaurs were adapted to the environments in which they were preserved to a significant degree. The material described here is assigned to Pterosauria Owen 1842, Pterodactyloidea Plieninger 1901 sensu Padian 2004, Kryptodrakon progenitor, gen. et sp. nov. Etymology Kryptodrakon progenitor, from the Greek krypto (hidden) and drakon (serpent), referring to the movie Crouching Tiger, Hidden Dragon, filmed near where the species was discovered, and the Latin progenitor (ancestral or firstborn), referring to its status as the earliest pterodactyloid. Material A single specimen (holotype) consisting of an incomplete postcranium: IVPP V18184 (Figures 1, 2B, and S1), collected from a 30 cm 2 area separated from all other fossils by at least 10 m distance. The specimen is housed at the Institute of Vertebrate Paleontology and Paleoanthropology, Beijing, China. Locality and Stratigraphy The fossil was collected from a mudstone in the alluvial facies of the lower part of the Shishugou Formation at Wucaiwan, Junggar Basin, Xinjiang, China. The specimen is from 35 m below the T-1 marker tuff [13], which had been dated as Ma [14]. Recalibration of the age of the Fish Canyon sanidine monitor mineral [15] suggests that the age of the T-1 marker tuff should be adjusted 0.6% older. Thus, we regard its age as Ma. In the context of an average sediment accumulation rate of 7.2 cm per 1,000 years for this part of the basin [16], the fossil specimen would have an age of approximately Ma. Following the Geologic Time Scale adjustments of Gradstein et al. (2012) [17], this age is the very earliest Oxfordian of the Upper Jurassic but within the error margin of the end of the Middle Jurassic [17],

2 Current Biology Vol 24 No Figure 1. Preserved Remains of Kryptodrakon progenitor, gen. et sp. nov., IVPP V18184 (A) Partial sacrum. (B) Ventral ramus of left coracoid. (C) Anterior end of right scapulocoracoid. (D) Proximal end and shaft of left humerus. (E) Distal end of right radius. (F) Right distal syncarpal. (G) Right preaxial carpal. (H) Right metacarpal. (I) Proximal end of right first wing phalanx. (J) Proximal end and shaft of left second wing phalanx. (K) Shaft of right third wing phalanx. (L) Proximal end of right fourth wing phalanx. The skeleton is Pterodactylus antiquus reprinted with permission from Wellnhofer (1991) [12], with preserved elements of Kryptodrakon progenitor infilled in black. Scale bar represents 50 mm. See also Figure S1. and so it may be conservatively considered as dating to the Middle-Upper Jurassic boundary. This timescale contains significant age adjustments from previous scales, which had dated this and other Wucaiwan specimens to the Middle Jurassic. Diagnosis Small pterodactyloid pterosaur differentiated from all other pterosaurs in having autapomorphic distal radius with distinct ventral flange and dorsally positioned anterior tubercle, autapomorphic preaxial carpal wider than long due to two expanded proximal flanges, and unique combination of elongate wing metacarpal with length more than eight times dorsoventral width at midpoint and anteroposteriorly compressed proximal end with large ventral expansion. Description The preserved remains of Kryptodrakon progenitor include a partial sacrum, ventral ramus of the left coracoid, anterior end of the right scapula, proximal end and shaft of the left humerus, distal end of the right radius, right distal syncarpal, right preaxial carpal, right wing metacarpal, proximal end of the right first wing phalanx, proximal end and shaft of the left second wing phalanx, shaft of the right third wing phalanx, proximal end of the right fourth wing phalanx, and a number of indeterminate fragments (Figures 1 and S1). This individual is considered an osteological adult based upon the fusion of the sacrum, scapulocoracoid, and distal syncarpal [18]. It is a small pterodactyloid with an estimated wingspan of 1.4 m (Supplemental Experimental Procedures). The characters that determine the phylogenetic relationships of Kryptodrakon are predominantly found in the most complete element, the wing metacarpal (Figures 1H, 2B, and S1). The metacarpus of pterosaurs consists of the robust wing metacarpal (metacarpal IV) and the slender metacarpals I III that lie along the anterior surface of the wing metacarpal to articulate with the claw-bearing digits. The right wing metacarpal is missing a small segment of the shaft, but the attenuation of its cross-section and the continuation of a posterior fossa onto the distal end indicate that only about 1 cm of length is absent. The combined minimum length of the preserved fragments is at least 84% of the humerus length estimate. This is more than the 80% necessary for referral to the Pterodactyloidea [4, 6]. The wing metacarpal in this species exhibits a predominantly nonpterodactyloid proximal end and a pterodactyloid distal end. Like the nonpterodactyloids, the proximal end of the wing metacarpal has an anteroposteriorly compressed cross-section, a proximal fossa on the posterior surface, a low proximal tuberculum, and a large ventral expansion (Figures 2A and 2B). By contrast, the proximal ends of the other pterodactyloid wing metacarpals are anteroposteriorly broadened into a rectangular cross-section, have a much thicker anteriorly curving flange that is the broadest part of the proximal end, have a taller proximal tuberculum, and have a much smaller ventral expansion (Figure 2C). Like other pterodactyloids, the distal end of the wing metacarpal in Kryptodrakon has an oval cross-section and an elongate shaft, bringing the total length to over nine times the mid-width (Figures 2B and 2C). By contrast, the distal end of nonpterodactyloid wing metacarpals have an anteroposteriorly compressed cross-section, a dorsal ridge on the posterior surface of the shaft, and a short shaft, bringing the total length to less than seven times the mid-width (Figure 2A). Two pterodactyloid features are present on the proximal surface of the wing metacarpal of Kryptodrakon, a dorsoventrally oriented ventral articular surface and a proximal tuberculum surrounded by a crescentic sulcus (Figures 2B and 2C), allowing greater rotation to absorb shocks and stresses during active flight [19]. Comprehensive phylogenetic analysis of Kryptodrakon and the relationships of the Pterosauria confirm this species as the basalmost member of the Pterodactyloidea (tree length = , consistency index = 0.357, retention index = 0.800) (Figures 3 and S2). This is the largest phylogenetic analysis of pterosaurs and is significantly larger than earlier versions of the analysis [11, 20, 21]. It incorporates the valid species (Table S2; Supplemental Experimental Procedures) and characters used in the previous 145 years of pterosaur phylogenetic study (112 species, 224 characters). Because many of the characters that determine the relationships of pterosaurs in general and Kryptodrakon specifically are continuous and

3 Earliest Pterodactyloid and Origin of the Group 1013 Figure 2. Comparison of Three-Dimensionally Preserved Pterosaur Wing Metacarpals Illustrating the Character States of Nonpterodactyloid and Pterodactyloid Pterosaurs in Anterior and Proximal Views (A) Nonpterodactyloid Comodactylus ostromi Galton 1981 (YPM 9150), courtesy of the Yale Peabody Museum of Natural History. (B) Basal pterodactyloid Kryptodrakon progenitor (IVPP V18184). (C) Derived pterodactyloid cf. Santanadactylus pricei Wellnhofer 1985 (AMNH 22552), courtesy of the American Museum of Natural History. Abbreviations: da, dorsal articular surface; dc, distal condyles; cs, crescentic sulcus that encircles proximal tuberculum; McI III, metacarpals I III; pt, proximal tuberculum; va, ventral articular surface; ve, ventral expansion. Scale bar represents 50 mm. have been subjectively coded differently in previous analyses, the continuous coding option of TNT [22] was implemented. The phylogenetic position of Kryptodrakon is supported by changes in two continuous and two discrete characters of the wing metacarpal, further reducing the long branch of the Pterodactyloidea (Table S3). The Pterodactyloidea are most parsimoniously reconstructed as originating in terrestrial environments as part of an extensive lineage that had been living there for at least 5 Ma(Figure 3). Whereas previous analyses and previously known pterodactyloid discoveries support a predominantly marine history for pterosaurs with repeated invasions of terrestrial environments (Table S1), our analysis implies a change in their ecology near the origin of the Pterodactyloidea wherein pterosaurs became predominantly terrestrial and repeatedly invaded marine environments (Figure 3). When subjected to a phylogenetic signal test [23], occurrence in terrestrial or marine preservational environments is very highly correlated with phylogeny in the comprehensive analysis, suggesting that segments of pterosaur evolution occurred in specific environments (one-tailed p = , n = 10,000,000). The pterosaur fossil record has been considered to be highly incomplete and biased, with the greatest bias being toward specimens found in marine environments [9, 12, 24], and so the preservational environments of pterosaurs may reflect more their contact with environments of exceptional preservation than their preferred environments. However, there is a substantial literature identifying the ecology of extinct organisms using modern species as models. When this had been previously applied to pterosaurs as a whole, size-independent measures of their wing shapes (wing aspect) were compared to those of modern birds and used to suggest a marine ecology for almost all pterosaurs (e.g., [25, 26]) [27]. Modern vertebrates flying in terrestrial environments typically have broader (low-aspect) wings than the narrower (high-aspect) wings of their marine counterparts [28, 29]. This correlation is attributed to more maneuverable flight utilized by terrestrial fliers for cluttered habitats [28], escape [29], greater landing frequencies [25], and higher takeoff angles [30]. However, pterosaurs as a whole have narrow wings in comparison with other flying vertebrates by virtue of their flight membranes being attached to an extremely elongate wing finger and a much shorter hindlimb. To test whether the same morphological difference between the wings of terrestrial and marine species exists in pterosaurs, the wing aspects of 19 pterosaur species from the literature [27] were correlated with terrestrial or marine occurrence using phylogenetic independent contrasts [31] to remove the effect of phylogeny (r = 0.563, two-tailed p = 0.012, n = 19) (Figure S3). Preservational environment is significantly correlated with wing shape in pterosaurs, such that terrestrial pterosaurs have lower wing aspects than their marine relatives and marine pterosaurs have higher wing aspects than their terrestrial relatives (sign test, two-tailed p = 0.008, n = 19), as in modern flyers. Pterosaurs have the appropriate wing shapes for the environments in which they were preserved. The physical demands of flight are extreme and vary with environment [29]. At the origin of the Pterodactyloidea, the greatest change in the flight apparatus was the change in the metacarpus, which transitions from being the shortest and least variable in length of the major wing elements in nonpterodactyloid pterosaurs to being the longest and most variable in length within the pterodactyloid pterosaurs. This increased variation in metacarpal length begins right at the transition of pterosaurs to terrestrial environments (Figure 3), with Sordes and the Anurognathidae having the relatively shortest

4 Current Biology Vol 24 No (legend on next page)

5 Earliest Pterodactyloid and Origin of the Group 1015 metacarpals of all pterosaurs. The subsequent elongation of the metacarpus is the diagnostic apomorphy of the Pterodactyloidea, but the increased variation in length producing this elongation is likely a key innovation [32] allowing new and more varied wing shapes for the pterosaurs to radiate into new and more varied environments. The sudden appearance and many diagnostic features of the pterodactyloid pterosaurs have been suggested to be the result of the group originating in terrestrial environments where the pterosaur fossil record has traditionally been poor, but this hypothesis was largely based on lack of evidence. Here we report the earliest and basalmost pterodactyloid and verify that it was found in a terrestrial environment. Phylogenetic comparative methods reconstruct both a terrestrial origin of the Pterodactyloidea and a predominantly terrestrial history for the group. This ecological history is supported by a significant correlation recovered between wing shape and environment that is also reported in modern flying vertebrates. This novel test of the fossil record demonstrates that pterosaurs lived in or were at least adapted to the environments in which they were preserved, and that ecology can be objectively and quantitatively reconstructed in extinct organisms. Experimental Procedures Comprehensive phylogenetic analysis was conducted with TNT v1.1 [22] using continuous and discrete partitions. Continuous characters were scaled so that their values ranged from 0 to 1. Ordered and unordered characters were used, and all were equally weighted. Inapplicable features were reductively coded, and polymorphic coding was used to denote variation within species. Four outgroups were used with Euparkeria capensis (Broom 1913) as the first outgroup. A basic tree-searching analysis with ratchet was implemented in TNT using 2,000 random addition sequence replicates. The ratchet proved to not be necessary. Ambiguous branch support was not used, zero-length branches were automatically collapsed, and resultant trees were filtered for best score. The analysis resulted in a single most parsimonious tree (Figures 3 and S1). Phylogenetic comparative analyses were conducted using Mesquite v2.75 [33]. These analyses consist of an ancestral state reconstruction, a phylogenetic signal test [23], and phylogenetic independent contrast tests [31]. Ancestral state reconstruction was conducted with maximum parsimony, resulting in eight equally parsimonious optimizations for 18 preservational environment changes, differing only within the Ctenochasmatidae and Pteranodontia. The ACCTRAN character optimization of preservational environment is illustrated in Figure 3. The phylogenetic signal test was conducted by randomly reshuffling the terminal taxa and optimizing the preservational environment character 10 million times to create a null distribution of the number character changes. Only one random optimization equaled the 18 preservational environment changes from the ancestral state reconstruction. The phylogenetic independent contrast tests were conducted using PDAP:PDTREE module v1.15 [34] on a pruned subset of the comprehensive phylogenetic analysis containing the 19 pterosaur species that have calculated wing aspects (Figure S3). Wing aspect and preservational environment contrasts were successfully standardized using the subset phylogenetic analysis. When subjected to Pearson product-moment correlation, wing aspect as Y contrast was positively significantly correlated with preservational environment as X contrast. A more conservative sign test confirmed that the actual predicted direction of changes in wing aspect and paleoenvironment are significantly correlated. Accession Numbers Color illustrations of individual elements and additional analysis figures are available at under project number 860. Supplemental Information Supplemental Information includes three figures, three tables, Supplemental Experimental Procedures, and two Supplemental Data Sets and can be found with this article online at Acknowledgments Chris Sloan collected the fossil. Hai-Jun Wang assisted with fieldwork and prepared the specimen. The fieldwork was supported by the National Natural Science Foundation of China, the National Science Foundation Division of Earth Sciences, the Chinese Academy of Sciences, the National Geographic Society, the Jurassic Foundation, the Hilmar Sallee bequest, and George Washington University. Study of the specimen was supported by the Chinese Academy of Sciences, the National Science Foundation Division of Earth Sciences, and the National Natural Science Foundation of China. Received: October 20, 2013 Revised: January 17, 2014 Accepted: March 11, 2014 Published: April 24, 2014 References 1. Unwin, D.M. (2006). Pterosaurs from Deep Time (New York: Pi Press). 2. Lü, J., Unwin, D.M., Jin, X., Liu, Y., and Ji, Q. (2010). Evidence for modular evolution in a long-tailed pterosaur with a pterodactyloid skull. Proc. Biol. Sci. 277, Plieninger, F. (1901). Beiträge zur Kenntnis der Flugsaurier. Palaeontographica 48, Unwin, D.M. (2003). On the phylogeny and evolutionary history of pterosaurs. In Evolution and Palaeobiology of Pterosaurs, Volume 217, Buffetaut, E., and Mazin, J.-M., eds. (London: Geological Society), pp Kellner, A.W.A. (2003). Pterosaur phylogeny and comments on the evolutionary history of the group. In Evolution and Palaeobiology of Pterosaurs, Volume 217, Buffetaut, E., and Mazin, J.-M., eds. (London: Geological Society), pp Padian, K. (2004). The nomenclature of Pterosauria (Reptilia, Archosauria). In First International Phylogenetic Nomenclature Meeting, M. Laurin, ed. (Paris: Muséum National d Histoire Naturelle), p Padian, K., and Rayner, J.M.V. (1993). The wings of pterosaurs. Am. J. Sci. 293, Wellnhofer, P. (1978). Pterosauria (Stuttgart: Gustav Fischer Verlag). 9. McGowan, C. (1991). Dinosaurs, Spitfires, and Sea Dragons (Cambridge: Harvard University Press). 10. Wang, X., Kellner, A.W.A., Zhou, Z., and Campos, Dde.A. (2005). Pterosaur diversity and faunal turnover in Cretaceous terrestrial ecosystems in China. Nature 437, Andres, B., Clark, J., and Xu, X. (2010). A new rhamphorhynchid pterosaur from the Upper Jurassic of Xinjiang, China, and the phylogenetic relationships of basal pterosaurs. J. Vertebr. Paleontol. 30, Wellnhofer, P. (1991). The Illustrated Encyclopedia of Prehistoric Flying Reptiles (London: Salamander Books). 13. Eberth, D.A., Xu, X., and Clark, J.M. (2010). Dinosaur death pits from the Jurassic of China. Palaios 25, Clark, J.M., Xu, X., Eberth, D.A., Forster, C.A., Malkus, M., Hemming, S., and Hernandez, R. (2006). The Middle- to Late Jurassic terrestrial Figure 3. Single Most Parsimonious Tree from the Comprehensive Phylogenetic Analysis of Kryptodrakon progenitor, gen. et sp. nov., and the Relationships of the Pterosauria Ranges for species denote the greatest temporal resolution of stratigraphic dating. Parsimony reconstruction using ACCTRAN optimization for the occurrence of pterosaur taxa preserved in terrestrial or marine environments is shown via yellow or blue lineages, respectively. Species used in the phylogenetic comparative analysis of paleoenvironment occurrence and wing aspect are denoted by asterisks and depicted as silhouettes reprinted with permission from Wellnhofer (1991) [12], drawn to scale. Outgroup relationships are not shown. Branch lengths and support measures are listed in Table S3. See also Figures S2 and S3 and Tables S1, S2, and S3.

6 Current Biology Vol 24 No transition: New discoveries from the Shishugou Formation, Xinjiang, China. In Ninth International Symposium on Mesozoic Terrestrial Ecosystems and Biota, P.M. Barrett and S.E. Evans, eds. (Manchester: Cambridge Publications), pp Kuiper, K.F., Deino, A., Hilgen, F.J., Krijgsman, W., Renne, P.R., and Wijbrans, J.R. (2008). Synchronizing rock clocks of Earth history. Science 320, Eberth, D.A., Brinkman, D.B., Chen, P.J., Yuan, F.T., Wu, S.Z., Li, G., and Cheng, X.S. (2001). Sequence stratigraphy, paleoclimate patterns, and vertebrate fossil preservation in Jurassic Cretaceous strata of the Junggar Basin, Xinjiang Autonomous Region, People s Republic of China. Can. J. Earth Sci. 38, Gradstein, F., Ogg, J., Schmitz, M., and Ogg, G. (2012). The Geologic Time Scale 2012 (Burlington: Elsevier). 18. Bennett, S.C. (1993). The ontogeny of Pteranodon and other pterosaurs. Paleobiology 19, Padian, K. (1984). A large pterodactyloid pterosaur from the Two Medicine Formation (Campanian) of Montana. J. Vertebr. Paleontol. 4, Andres, B., and Ji, Q. (2008). A new pterosaur from the Liaoning Province of China, the phylogeny of the Pterodactyloidea, and convergence in their cervical vertebrae. Palaeontology 51, Andres, B., and Myers, T.S. (2013). Lone Star Pterosaurs. Earth Environ. Sci. Trans. R. Soc. Edinb. 103, Goloboff, P.A., Farris, J.S., and Nixon, K.C. (2008). TNT, a free program for phylogenetic analysis. Cladistics 24, Maddison, W.P., and Slatkin, M. (1991). Null Models for the Number of Evolutionary Steps in a Character on a Phylogenetic Tree. Evolution 45, Kellner, A.W.A. (1994). Remarks on pterosaur taphonomy and paleoecology. Acta Geol. Leopoldensia 39, Hazlehurst, G.A., and Rayner, J.M.V. (1992). Flight characteristics of Triassic and Jurassic Pterosauria: An appraisal based on wing shape. Paleobiology 18, Rayner, J.M.V. (1989). Mechanics and physiology of flight in fossil vertebrates. Trans. R. Soc. Edinb. Earth Sci. 80, Witton, M.P. (2008). A new approach to determining pterosaur body mass and its implications for pterosaur flight. Zitteliana B 28, Norberg, U.M., and Rayner, J.M.V. (1987). Ecological morphology and flight in bats (Mammalia; Chiroptera): Wing adaptations, flight performance, foraging strategy and echolocation. Philos. Trans. R. Soc. Lond. B Biol. Sci. 316, Rayner, J.M.V. (1988). Form and function in avian flight. Curr. Ornithol. 5, Pennycuick, C.J. (1983). Thermal soaring compared in three dissimilar tropical bird species, Fregata magnificens, Pelecanus occidentalis, and Cragyps atratus. J. Exp. Biol. 102, Felsenstein, J. (1985). Phylogenies and the comparative method. Am. Nat. 125, Simpson, G.G. (1953). The Major Features of Evolution (New York: Columbia University Press). 33. Maddison, W.P., and Maddison, D.R. (2008). Mesquite: A modular system for evolutionary analysis, v mesquite/mesquite.html. 34. Midford, P.E., Garland, T., and Maddison, W.P. (2005). PDAP package of Mesquite documentation, v mesquite/index.html.

7 Current Biology, Volume 24 Supplemental Information The Earliest Pterodactyloid and the Origin of the Group Brian Andres, James Clark, and Xing Xu

8 Supplemental Data

9 Figure S1. Preservation of Kryptodrakon progenitor gen. et sp. nov. (IVPP V18184), Related to Figure 1. Morphology shown in (i) anterior, (ii) dorsal, (iii) posterior, (iv) ventral, (v) right, (vi) left, (vii) proximal, and (viii) distal views. Scale equals 50 mm. (A), Partial sacrum. (B) Ventral ramus of left coracoid. (C) Anterior end of right scapulocoracoid. (D) Proximal end and shaft of left humerus. (E) Distal end of right radius. (F) Right distal syncarpal. (G) Right preaxial carpal. (H) Right metacarpal. (I) Proximal end of right first wing phalanx. (J) Proximal end and shaft of left second wing phalanx. (K) Shaft of right third wing phalanx. (L) Proximal end of right fourth wing phalanx.

10

11 Figure S2. Phylogeny of the Pterosauria and Kryptodrakon progenitor illustrating phylogenetic relationships and nomenclature erected, Related to Figure 3. Clades are defined in Supplemental Experimental Procedures and numbers refer to branches listed in Table S3.

12 Figure S3. Topology, branch lengths, species, and wing shapes used in comparative analysis, Related to Figure 3. Phylogenetic relationships, morphological branch lengths, size-independent measures of wing shape, and ancestral state reconstruction of paleoenvironment for 19 pterosaur species used in the comparative analysis of the pterosaur fossil record. The phylogenetic topology and branch lengths were created by pruning species without shape estimates from the comprehensive analysis. The size-independent measures of wing shape were taken from a principal components analysis of wing aspect ratio (Q3) in Witton (2008) [S1]. Terrestrial (yellow) and marine (blue) paleoenvironments were reconstructed using maximum parsimony. Branch lengths are listed below their respective branches,

13 Table S1. Previous Phylogenetic Studies and Reconstruction of Environmental Occurrence, Related to Figure 3. All previous phylogenetic studies from which data was incorporated into the comprehensive analysis and occurrence of pterosaur species in marine or terrestrial preservational environments. Percentage and ratio of terrestrial to marine pterosaur species, number of switches between marine and terrestrial environments implied by phylogeny, and in what environment the origin of pterodactyloids is most parsimoniously reconstructed are listed. For this last column, terrestrial and marine denote that environment of origin for the Pterodactyloidea, not present denotes analyses that do not include the Pterodactyloidea node either because they just consist of nonpterodactyloid pterosaurs or consist of a less inclusive clade within the Pterodactyloidea, terminal denotes a study in which the Pterodactyloidea is a supraspecific terminal taxon and so its internal relationships cannot help determine the ancestral state, ingroup denotes analysis in which the ingroup is the Pterodactyloidea and would not sample enough non-pterodactyloid species and clades to reliably reconstruct the origin of the Pterodactyloidea, and ambiguous denotes an analysis in which it is equally parsimonious for the Pterodactyloidea to have a terrestrial or marine origin or that some of the resultant most parsimonious trees result in different ancestral states for the same analysis. Phylogeny Percentage Terrestrial Pterosaurs Ratio of Terrestrial to Marine Pterosaurs Number of Environmental Switches Pterodactyloid Origin Environment Comprehensive Analysis 50.93% 56:52 18 terrestrial Wang et al. (2012) [S2] 59.38% 19:32 8 marine Vullo et al. (2012) [S3] 51.85% 14:13 8 ingroup Lü et al. (2012b) [S4] 44.07% 26:33 18 ambiguous Lü et al. (2012a) [S5] 36.36% 8:14 6 terminal Witton (2012) [S6] 60.00% 6:4 4 not present Pinheiro et al. (2011) [S7] 75.00% 12:4 2 not present Averianov (2010) [S8] 54.39% 31:26 11 ingroup Andres et al. (2010) [S9] 33.33% 6:12 3 terminal Lü et al. (2010) [S10] 38.36% 28:45 17 ambiguous Dalla Vecchia (2009b) [S11] 21.74% 5:18 4 terminal Wang et al. (2009) [S12] 35.85% 19:34 12 marine Lü (2009) [S13] 43.24% 16:21 10 ambiguous Dalla Vecchia (2009a) [S14] 20.83% 5:19 4 marine Lü et al. (2008b) [S15] 61.36% 27:17 10 ingroup Lü et al. (2008a) [S16] 48.84% 21:22 9 ingroup Andres and Ji (2008) [S17] 54.39% 31:26 11 ingroup Wang et al. (2008) [S18] 33.33% 15:30 9 marine Bennett (2007) [S19] 16.67% 1:5 1 terminal Lü and Ji (2006) [S20] 52.83% 28:25 12 ambiguous Martill and Naish (2006) 75.00% 6:2 2 not present [S21] Wang et al. (2005) [S22] 31.82% 14:30 8 marine Kellner (2004) [S23] 29.27% 12:29 6 marine Maisch et al. (2004) [S24] 50.00% 3:3 1 not present Unwin (2004) [S25] 27.78% 5:13 4 terminal Kellner (2003) [S26] 26.32% 10:28 6 marine Unwin (2003) [S27] 26.32% 5:14 5 marine Unwin (2002) [S28] 25.00% 2:6 2 not present

14 Viscardi et al. (1999) [S29] 23.26% 10:33 9 marine Unwin and Lü (1997) [S30] 20.00% 2:8 2 ingroup Peters (1997) [S31] 19.05% 4:17 3 marine Kellner (1996) [S32] 25.00% 8:24 5 marine Unwin (1995) [S33] 21.43% 3:11 3 marine Bennett (1994) [S34] 23.08% 6:20 5 ingroup Unwin (1991) [S35] 33.33% 6:12 4 marine Bennett (1991) [S36] 20.83% 5:19 4 ingroup Bennett (1989) [S37] 18.75% 3:13 3 ingroup Howse (1986) [S38] 20.00% 3:15 3 ingroup Wellnhofer (1978) [S39] 12.50% 2:14 2 marine Wild (1978) [S40] 0.00% 0:4 0 not present Wellnhofer (1975) [S41] 0.00% 0:7 0 not present Kuhn (1967) [S42] 22.22% 2:7 2 marine Young (1964) [S43] 7.69% 1:12 1 marine Haeckel (1895) [S44] 0.00% 0:4 0 marine Haeckel (1866) [S45] 0.00% 0:2 0 marine

15 Table S2. Species and Specimen Data for the Phylogenetic and Comparative Analyses, Related to Figure 3. Species terminal taxa and specimens used in the phylogenetic analysis of Kryptodrakon progenitor, gen. et sp. nov, and the relationships of the Pterosauria. Specimens used for coding species, average wing metacarpal (McIV) length divided by humerus (hu) length for each species, average wing metacarpal length divided by the wing metacarpal dorsoventral width at the midpoint for each species, and paleoenvironment identified from the literature are listed. Paleoenvironment was not considered for the nonvolant outgroups. The vast majority of specimens were coded first hand whenever possible and supplemented with information from casts and literature when not. Abbreviations: AMNH, American Museum of Natural History, New York, New York; BHI, Black Hills Institute of Geological Research, Hill City, South Dakota; BM, Beipei Museum, Zigong, Dashanpu, Sichuan, China; BPM, Beipiao Museum of Liaoning Province, Beipiao City, Liaoning Province, China; BPV, Beijing Natural Museum, Beijing, China; BSP, Bayerische Staatssamlung für Paläontologie, Munich, Germany; BXGM, Benxi Geological Museum, Liaoning Province, China; CAMSM, Sedgwick Museum, University of Cambridge, Cambridge, United Kingdom; CAD, Paleontology and Stratigraphy Center of the Jilin University, Changchun, Jilin Province, China; CCMGE, Museum of Central Scientific Research Institute for Geological Exploration, University of St. Petersburg, St. Petersburg, Russia; CMNH, Carnegie Museum, Pittsburgh, Pennsylvania; D, Dalian Natural History Museum, Dalian, China; DGM, Museum de Ciencia da Terra/Departmento Nacional Producao Mineral, Rio de Janeiro, Brazil; DMNH, Denver Museum of Natural History, Denver, Colorado; FHSM, Fort Hays State Museum, Fort Hays State University, Hays, Kansas; FMNH, Field Museum of Natural History, Chicago, Illinois; GIN, Institute of Geology, Mongolian Academy of Sciences, Ulan Bataar, Mongolia; GLGMV, Guilin Longshan Geological Museum, Gulin City, Gunagxi Zhuang Autonomous Region; GMN, Geological Museum of Nanjing, Nanjing, China; GPIB, Geologisch-Paläontologisches Institut der Universität Bonn, Bonn, Germany; GPIT, Institut und Museum für Geologie und Paläontologie der Universität Tübingen; GSM, Geological Survey Museum, Keyworth, Nottinghamshire, United Kingdom; Hauff, Urwelt-Museum Hauff, Holzmaden, Germany; HGM, Henan Geological Museum, Zhengzhou, China; IGO, Museo Mario Sánchez Roig, Instituto de Geología y Paleontología, La Habana, Cuba; IMCF, Iwaki Coal and Fossil Museum, Iwaki, Japan; IVPP, Institute of Vertebrate of Paleontology and Paleoanthropology, Beijing, China; JME, Jura Museum, Eichstätt, Bavaria, Germany; JPM, Henan Geological Museum, Henan Province; JZMP, Jinzhou Paleontological Museum, Liaoning Province; KJ, Private collection, Texas; KUVP, Museum of Natural History, University of Kansas, Kansas; LACM, Los Angeles County Museum of Natural History, Los Angeles, California; LPM, Liaoning Provincial Museum of Paleontology, Liaoning, China; MANCH, Manchester Museum, Manchester, United Kingdom; MB, Museum für Naturkunde, Humboldt Universität, Berlin, Germany; MCCM, Museo de las Ciencias de Castilla La Mancha, Cuenca, Spain; MCSNB, Museo Civico di Storia Naturale, Bergamo, Italy; MCT, Museu Ciencias da Terra, Setor de Paleontologia, Departmento Nacional da Producão Mineral, Rio de Janeiro, Brazil; MCZ, Museum of Comparative Zoology, Cambridge, Massachusetts; MFSN, Museo Friulano di Storia Naturale, Udine, Italy; MGCL, Musée Géologique Cantonal, Lausanne, France; MHIN-UNSL, Universidad Nacional de San Luis, Chacabuco y Pederna, San Luis, Argentina; MMNH, Mongolian Museum of Natural History, Ulan Bataar, Mongolia; MN,

16 Museu Nacional, Departmento de Geologia e Paleontologia, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil; MNHN, Museum of National d Histoire Naturelle, Paleontologie, Paris, France; MPUM, Museo Paleontologia Universitá di Milano, Italy; MSA, Museum am Solenhofer Aktienveriein, Maxberg bei Solnhofen, Germany; MTM, Természettudományi Múzeum, Budapest, Hungary; NAMAL, North American Museum of Ancient Life, Lehi, Utah; NGMC, National Geological Museum of China, Beijing, China; NHMUK, British Museum of Natural History, London, United Kingdom; NHMW, Naturhistorisches Museum Wien, Vienna, Germany; N M, Geological Institute, University of Amsterdam, Amsterdam, Netherlands; NMING, National Museum of Ireland, Dublin, Republic of Ireland; NRA, Natural Resources Authority, Amman, Jordan; NSM, Division of Vertebrate Paleontology, National Science Museum, Tokyo, Japan; PIN, Palaeontological Institute, Russian Academy of Sciences, Moscow, Russia; PSB, Banz Collection (Petrefaktensammlung Banz), Banz, Germany; PVL, Fundación- Insituto Miguel Lillo, Universidad Nacional, Tucuman, Argentina; Rohrbach, Fossil Museum of the Rohrbach Concrete Company, Dotternhausen, Germany; SAO, Collection Oberli, St. Gallen, Switzerland; SC, Museo Geologico della Carnia, Ampezzo, Italy; SM, Natur-museum und Forschungsinstitut Senckenberg, Frankfurt am Main; SMNK, Staatliches Museum für Naturkunde Karlsruhe, Karlsruhe, Germany; SMNS, Staatliches Museum für Naturkunde in Stuttgart, Zweigstelle Ludwigsburg; TMM, Texas Memorial Museum, University of Texas at Austin, Austin, Texas; UALVP, Geology Museum, University of Alberta, Edmonton, Canada; UERJ, Universidade Estadual do Rio de Janeiro, Rio de Janeiro, Brazil; UMZC, University Museum of Zoology, Cambridge, United Kingdom; UNSM, University of Nebraska State Museum, Lincoln, Nebraska; USNM, Smithsonian Museum of Natural History, Washington, DC; UUPM, Paläontologisches Institut der Universität Uppsala, Uppsala, Sweden; VF, Department of Geology of the University of Jordan at Amman, Amman, Jordan; YPM, Peabody Museum, Yale University, New Haven, Connecticut; and ZMNH, Zhejiang Museum of Natural History, Hangzhou, Zhejiang, China.

17 Table S3. Branch Lengths and Support Measures for the Phylogenetic Analysis of Kryptodrakon progenitor and the Pterosauria, Related to Figure 3. Minimum branch lengths, bootstrap scores, and jacknife scores from the comprehensive phylogenetic analysis of Kryptodrakon progenitor, gen. et sp. nov., and the relationships of the Pterosauria are listed for the branches numbered in Figure S2. Dashes denote the branch length for the root node or bootstrap and jacknife scores for terminal branches that are not calculable.

18 Supplemental Experimental Procedures Complete Morphological Description of Kryptodrakon progenitor The elements of the holotype and only specimen of Kryptodrakon progenitor (IVPP V18184), gen. et sp. nov. were collected from a 30 cm 2 area separated from all other fossils by at least 10 m distance. It is not possible for more than one individuals to be present in this discovery. The wingspan estimate of 1.47 m was measured according to the method of Bennett (2001) [S99], in which the lengths of the humeri, ulnae, wing metacarpals, and wing phalanges are summed to calculate the wingspan. The omission of the dimensions of the carpus, pectoral girdle, and torso are meant to offset the use of straight linear measurements without consideration of the flexures of the wing [S99]. Direct estimates of only the humerus and wing metacarpal lengths can be made, and so the other major wing bones lengths must be reconstructed. We used a phylogenetic method to reconstruct the missing elements. Fortunately, the continuous characters 25, 27, 30, 31, 32, and 33 code the ratios of the lengths of all these wing bones with respect to one another. The input of the length of one of these bones can be used with these ratios to successively calculate an estimated length of all others in succession. The codings for these characters were scaled back to their original values, imported into Mesquite version 2.74 [S100], and optimized on the tree from the phylogenetic analysis to determine their values for Kryptodrakon (Figures 3 and S2). Characters coding limb dimensions as ratios with other limb elements were combined with a metacarpal IV length of 82.2 mm to determine the estimated dimensions of these elements and the wingspan in Kryptodrakon. Kryptodrakon is described here in the pterosaur anatomical position of wings outstretched laterally as they would be in flight and on which wing shapes would be measured. Therefore, what might be termed the medial and lateral aspects of the more proximal wing bones are instead referred to as the ventral and dorsal aspects, respectively. The dorsoventral depth of forelimb elements is referred to as the width, and the dorsoventral depth at the midpoint of a forelimb element is referred to as the mid-width as in the main text. Three-dimensional preservation such as found in Kryptodrakon is exceedingly rare in pterosaur specimens. For many species, the more three-dimensional aspects of morphology can only be confirmed through fortuitous positioning and breaks of elements in specimens. Kryptodrakon was compared to all specimens listed in Table S2 through direct observation or published descriptions and illustrations to determine the distribution of its three-dimensional morphological features over phylogeny, including newly discovered basal monofenestrans. For simplicity of discussion, comparisons in the text were made with respect to Comodactylus ostromi Galton 1981 [S101] (YPM 9150), and cf. Santanadactylus pricei Wellnhofer 1985 [S102] (AMNH 22552), the most three-dimensionally preserved non-pterodactyloid and pterodactyloid specimens, respectively. These observations hold for all non-pterodactyloids and pterodactyloids for which the morphology can be assessed. Portions of the sacrum, left coracoid, right scapulocoracoid, left humerus, right radius, right distal syncarpal, right preaxial carpal, right wing metacarpal, right first wing phalanx, left second wing phalanx, right third wing phalanx, fourth wing phalanx, and a number of indeterminate fragments are preserved in the holotype. These elements are illustrated in Figures 1, 2, and S1. All elements are fused that fuse during the ontogeny of pterosaurs, indicating osteological maturity for this specimen [S103]. There are no preserved pneumatic foramina.

19 The axial skeleton of Kryptodrakon is represented by the posterior half of the second sacral vertebra fused to a small portion of the anterior cotyle of the third. The preserved length of this element is 7.4 mm, the preserved width is 7.2 mm, and the preserved height is 7.9 mm. The neural spine and transverse processes of the second sacral are broken off at their bases. These vertebrae are identified as sacrals based on the total fusion of their squat shapes, the merging of the neural arch with the centrum so that the transverse processes directly contact the centrum, the triangular cross-section of the transverse processes [S104], the neural canal constituting the majority of the middle cross-section, and the neural spine and transverse processes reaching almost to the posterior margin of the vertebra. Their identities as the second and third sacrals are based upon the lateral orientation of the transverse processes of the second sacral, and the second sacral having a greater excavation of the centrum. These are the two original archosaurian sacral vertebrae found in pterosaurs, and therefore an estimate of the total number of sacral vertebrae in Kryptodrakon in not currently possible. The preserved pectoral girdle of Kryptodrakon includes the thin ventral ramus of the left coracoid and the probable anterior end of the right scapulocoracoid. The left coracoid is broken off mid-shaft and is missing a portion of the left side of the sternal articulation. The preserved length of this element is 36.8 mm, its lateral width at its midsection is 5.4 mm, and its dorsoventral depth at its midsection is 6.6 mm. The remaining part of the articular surface is saddle-shaped and semi-circular in cross-section. A keel extends along the lateral surface, and a ventral keel begins just before the mid-shaft break. The anterior end of the right scapula is poorly preserved and unusual in overall shape. It is identified principally by its expanded triangular cross-section that flattens posteriorly, a thick lip overhanging a saddle-shaped glenoid fossa, a lateral edge that tapers into ventrally curving keel, a supraglenoid buttress, and a buttress posterior to the glenoid. It is 42 mm in preserved length with a lateral with of 7.7 mm and dorsoventral depth of 5.4 mm at its midsection. The scapula is remarkably straight preserving only the anterior-most extent of its curved region. The supraglenoid buttress is represented by a small tubercle dorsomedial to the glenoid lip. A small knob of bone attached to the ventral end of buttress posterior to the glenoid is identified as a fragment of the coracoid and indicates the fusion of the scapulocoracoid in this individual. The preserved portions of the left humerus of Kryptodrakon are the humeral head and the proximal end of the humeral shaft. The humeral head fragment has a preserved length of 13.2 mm, a dorsoventral width of 8.1 mm, and an anteroposterior breadth of 14.2 mm. The humeral shaft fragment has a preserved length of 30.2 mm, ad dorsoventral mid-width of 7.4 mm, and an anteroposterior breadth of 9.8 mm. These elements include the proximal and distal margins of the base of the deltopectoral crest. The ulnar (medial) crest is not preserved. The saddle-shaped articular surface of the humeral head is semi-circular in cross-section, unlike the more crescentic shape of the non-pterodactyloid pterosaurs [S9] and the more horseshoe shape of the azhdarchid pterosaurs [S35]. This surface tapers slightly in width anteriorly until it contacts a sharp vertical flange on the anterior surface. This flange reaches the proximal margin of the deltopectoral crest base. The humeral head contacts the anterior surface at a right angle forming a ridge that is continued onto the humeral shaft fragment until the distal margin of the deltopectoral crest. The shaft fragment preserves the majority of the base of a large, ventrally curving deltopectoral crest that expands proximally and has a straight distal margin. The preserved portion of this crest is most similar to the condition found in the archaeopterodactyloid pterosaurs. There is no constriction in the shaft distal to the deltopectoral crest.

20 Because the length of the humerus is relevant to the identification of Kryptodrakon as a pterodactyloid, it is necessary to estimate its length. More than half of humerus is preserved in this specimen, but consists of two pieces with broken and weathered ends so that it is not possible to fit them together. However, both pieces preserve part of the base of a proximally expanded deltopectoral crest with a straight distal margin. All pterosaurs studied in this analysis and listed in Table S3 with these features have a deltopectoral crest that terminates in a tip approximately the same size of the humeral head that is even with the proximal margin of the humerus. Extending these observations to this specimen indicates that only a couple of millimeters of bone are missing between these pieces, and is consistent with the similar shape between the preserved ends. This allows an estimate of the length from the proximal margin of the humerus to the extreme distal margin of the deltopectoral crest. A survey of this length in pterosaurs indicates that this measurement constitutes 35-40% of the total humeral length. The total humerus length in Kryptodrakon is therefore estimated to be between 75 and 86 mm. The minimum length of the wing metacarpal would then be between 84 and 96% the length of the humerus, more than the 80% required for referral to the Pterodactyloidea. A median humeral length of 80 mm is used for the phylogenetic analysis. The distal end of the right radius is preserved in Kryptodrakon. The preserved length of this fragment is 38.2 mm, its maximum dorsoventral width is 8.8 mm, and its maximum anteroposterior breadth is 4.2 mm. It is autapomorphic in having a large ventral flange and a second, dorsally positioned anterior tubercle. All pterosaur radii have a distal expansion, but the expansion in this specimen is more pronounced ventrally due to two distinct ventral tubercles with a flange extending between them and to the distal margin of the radius. The distal surface of pterosaur radii has dorsal and ventral articular surfaces for articulation with the proximal syncarpal. The ventral articular surface extends onto the anterior surface to create a raised structure termed the anterior tubercle [S104]. The dorsal articular surface extends onto the posterior surface, but in Kryptodrakon this surface is expanded anteriorly to produce a second anterior tubercle that is almost as pronounced as the more ventral anterior tubercle. A deep, wide sulcus divides these anterior tubercles. The preserved carpus of Kryptodrakon includes the right distal syncarpal and preaxial carpal (also termed the medial or lateral carpal in previous literature). The distal syncarpal has a proximodistal length of 10.8 mm, a dorsoventral width of 12.3 mm, and a anteroposterior breadth of 18.2 mm. It is rectangular in cross-section and has a semi-circular outline with a large anterior process for articulation with the preaxial carpal in dorsal and ventral views. The anterior aspect of this preaxial carpal process is inclined proximally. Similarly, the posterior surface has a proximally inclined rugose ridge. The proximal surface is dominated by two large, triangular wedges that are the intersyncarpal articulation surfaces for the proximal syncarpal. The ventral articulation surface is more elongate and reaches anteriorly to contact the preaxial carpal process. The posterior half of the distal surface mirrors the proximal surface of the wing metacarpal with which it articulates: a small dorsal articular surface that is separated from a larger ventral articular surface by an hourglass-shaped median ridge. Anteriorly, the dorsal articular surface contacts a hooked process in the middle of the distal margin of the dorsal surface, and the hourglass-shaped ridge contacts a circular fovea in the center of the distal surface. The distal surface of the preaxial carpal process consists of three small, vertically arranged divots that likely articulated with the proximal ends of the metacarpals I-III. The preaxial carpal of Kryptodrakon is rhomboid in outline and proximodistally compressed in cross-section. It has a proximodistal length of 5.2 mm, a maximum dorsoventral