A Virtual Phytosaur Endocast and Its Implications for Sensory System Evolution in Archosaurs Waymon Holloway [email protected]

Marshall University Marshall Digital Scholar

Theses, Dissertations and Capstones

1-1-2011 A Virtual Phytosaur Endocast and its Implications for Sensory System Evolution in Archosaurs Waymon Holloway [email protected]

Follow this and additional works at: http://mds.marshall.edu/etd Part of the Terrestrial and Aquatic Ecology Commons

Recommended Citation Holloway, Waymon, "A Virtual Phytosaur Endocast and its Implications for Sensory System Evolution in Archosaurs" (2011). Theses, Dissertations and Capstones. Paper 83.

This Thesis is brought to you for free and open access by Marshall Digital Scholar. It has been accepted for inclusion in Theses, Dissertations and Capstones by an authorized administrator of Marshall Digital Scholar. For more information, please contact [email protected].

A VIRTUAL PHYTOSAUR ENDOCAST AND ITS IMPLICATIONS FOR SENSORY SYSTEM EVOLUTION IN ARCHOSAURS

A Thesis submitted to The Graduate College of Marshall University

In partial fulfillment of the requirements for the degree of Master of Science in Biological Sciences

by Waymon Holloway

Approved by

Dr. F. Robin O’Keefe, Committee Chairperson Dr. Suzanne G. Strait Dr. Paul J. Constantino

Marshall University June 2011 Acknowledgements

This thesis was greatly improved by the comments of my advisor Dr. F. R.

O’Keefe and committee members Dr. S. Strait and Dr. P. Constantino. Description of the internal cranial anatomy of the study specimen would not have been possible without the patient consideration of Dr. M. Carrano, who arranged a loan of the specimen from the

NMNH, and S. Jabo, who facilitated the transport of the material. This research was supported by a MURC student travel grant to W. Holloway. Additional thanks are due to

N. Gardner, Dr. L. Witmer, and R. Ridgely for their assistance with various computer hardware and software issues, C. Richards for her support, K. Macys for her support and invaluable role as a test audience for my research ideas, and my mother Patricia for her varied and meaningful support throughout my academic career.

i Table of Contents

Acknowledgements ……………………………………………………….....…….……...i

List of Figures ………………………………………..………………………...………..iii

Abstract …………………………………………………………………………...... …...iv

Chapter 1 - Background information and rationale ………….……………….…...1-21

Introduction………………..…………………………………………………....1-5

Background of phytosaurs………………….………..…………………..….…5-11

Background of CT study techniques……...……….....……...……..………...11-18

Rationale……...………...………………………………..……………..….…18-21

Chapter 2 - Materials, methods, and results ………………………..…………..…21-38

Materials………………..………………………………………………….…21-23

CT scanning and 3D visualization…..…………………………………….….23-29

Cranial endocast……………………………………………………….……..29-36

Endosseous labyrinth………………………………..…………….………….36-38

Chapter 3 - Discussion and Conclusions …………….………………………..…...38-52

Endocast Comparison…………………………………………………...……38-41

Comparison of archosaurian endocasts………………………....……...…….41-49

Future Work……………………...... ………………………….………..….…49-50

Conclusions………………………...………………………….…………..…50-52

Appendix ……………………………..……………………………….……….……….. 53

References ………………………..……………………………………...…………...54-62

ii Table of Figures

Figure 1.1: Crurotarsi phylogeny………………………………………………………….6 Figure 1.2: Phytosaur morphology………………………………………………….…...11 Figure 1.3: CT scanner components………………………………………………..……13 Figure 1.4: Tyrannosaurus rex cranial endocast and pneumaticity……………...………15 Figure 1.5: CT slice density data………….………………………………….………….17

Figure 2.1: Skull of USNM 17098…………………………………....…….……………22 Figure 2.2: CT visualization process…………………………………………….…...25-29 Figure 2.3: Cranial cavity of USNM 17098 with break…………………………….…....32 Figure 2.4: Cranial endocast of USNM 17098…………………………………………..33 Figure 2.5: Head posture of USNM 17098……………………………………..………..38

Figure 3.1: USNM 17098 and Crocodile endocasts……………………....……………..40 Figure 3.2: Crurotarsi endocasts and phylogeny………...………………….…………....43 Figure 3.3: Theropoda endocasts and phylogeny ……………………………………….46 Figure 3.4: Sauropodomorpha endocasts and phylogeny …………………….…………47

iii Abstract

A virtual phytosaur endocast and its implications for sensory system evolution in archosaurs

By Waymon Holloway

Due to the overall morphological similarities between the Triassic archosaurs of the order Phytosauria and extant crocodilians, most studies have assumed that the two shared similar lifestyles. Many studies involving phytosaurs have focused on the external cranial morphology of various taxa. Internal cranial anatomy has received relatively little attention. As a result, comparative morphology studies of the braincases interior, or endocast, of phytosaurs are an area of potential exploration. Just as modern medical X- ray computed tomography (CT) can be used to create three-dimensional images of internal structures of living subjects, such technology offers a non-invasive means of studying the internal anatomy of fossil specimens. CT scan data, used to virtually model and analyze the endocast of a phytosaur specimen, reveals an endocast morphology conserved among phytosaurs, crocodilians, and other Crurotarsi. This conservatism persists in taxa with numerous synapomorphies and distinctive ecologies, unlike the endocast specialization seen in other archosaurs like Dinosauria.

iv Chapter 1 - Background information and rationale

Introduction

Paleontology is sometimes believed to be a restricted field of study because of a common misconception that very little determination about the lifestyle of an animal can be made from studying only its skeletal system. Although it is true that the skeletal system alone can be a poor indicator of life habits beyond those dependent upon a certain body shape or presence of certain osteological features, a close examination of the bones of an animal, living or dead, can reveal a great deal of information about other systems of the body (e.g. Witmer, 2005). For example, certain bones often exhibit muscle scars that would indicate the point of attachment, size, and orientation of various muscles and muscle groups. These data can then be used to create a model for the specifics of the method of locomotion, possible feeding habits, and any number of other muscle- dependent activities utilized by the animal. Similarly, the absence of bone in certain places or the size and shape of cavities within a skeletal structure can be an excellent source of information about the soft tissues of an animal. One region of the body that exhibits a number of important cavities is the skull. In particular, within the skull is housed the brain. Through advancements in technology, it has become possible to study the structure of the space occupied by the brain in great detail. Similar to more traditional paleontological methods of inferring information about the soft tissues based on features seen in the osteology of a specimen, an X-ray computed tomography (CT) aided virtual reconstruction of the endocasts of fossil specimens can be created and analyzed in order to describe previously overlooked aspects of the cranial anatomy (e.g.

1 Conroy & Vannier, 1984; Carlson et al., 2003), adding to the behavioral models

associated with these specimens (e.g. Witmer & Ridgely, 2009), and enabling a better

understanding of the taxonomic relationships of these specimens (Franzosa & Rowe,

2005).

Understanding the functional anatomy and behavior of an extinct species can be

an important source of information about why extant species live the way that they do.

Although predicting precise aspects of organismal function based on the form or structure

of individual morphological features is unlikely to be successful, the overall morphology

of an animal and the behavior or ecology exhibited by that animal are linked closely

enough to warrant general models (Lauder, 1995). In the case of morphologically

convergent taxa (i.e. those that share a similar morphology that is not a result of a close

phylogenetic relationship), similar behaviors are often assumed for both taxa based on an

inference that similar morphologies are closely associated with similar behaviors. Such

assumed behavioral similarities can be further supported by similarities in other aspects

like habitat.

Evolutionary convergence can also be seen in the morphology of more specific

features of taxa that share overall similar lifestyles. The brain is one area of possible

convergence between lifestyle-convergent taxa, even when those taxa are not considered

to be convergent in terms of their overall morphologies. An example of this can be seen

between pterosaurs and the earliest known bird, Archaeopteryx . Both pterosaurs, as a group, and Archaeopteryx are generally considered to have been aerial carnivores (e.g.

Witmer, 2004), yet they each evolved from an entirely different evolutionary lineage

(Laurin & Gauthier, 1996) and are not convergent in terms of their overall morphologies,

2 most notably their wing forms (e.g. Gauthier & Padian, 1985). In the brain and inner-ear

canals of Archaeopteryx , an expansion and reorganization of areas that are neurologically important for flight can be seen, and although these features are more fully developed in modern birds, the development and reorganization present in Archaeopteryx would have been sufficient for a lifestyle involving full flight ability (Domínguez Alonso et al.,

2004). The brain and inner-ear of pterosaurs exhibit an expansion and reorganization very much like that seen in Archaeopteryx (Witmer et al., 2003). This independently evolved similarity is an excellent example of convergence that can be seen in terms of specific features, like that of neurological morphology, and suggests that just as a specific overall morphology can be beneficial, if not required, for a certain lifestyle, a specific morphology of certain key features, like the brain and inner-ear canals, may indeed be a fundamental requirement for the ability to perform certain behaviors (Witmer, 2004).

Moreover, Archaeopteryx seems to have evolved a neurological morphology conducive to flight in a piece-meal fashion, co-opting the advanced neurological apparati inherited from its theropod dinosaur ancestors and refining it as flight improved, an idea supported by the further development of the brains of modern birds over that of early forms like

Archaeopteryx (Domínguez Alonso et al., 2004; Milner & Walsh, 2009). This repurposing and refinement in birds is very different from pterosaurs that developed their flight-enabling neurological morphology from scratch (Witmer et al., 2003), suggesting that the evolutionary path taken toward reaching similar neurological morphologies does not seem to be as important for enabling similar behaviors as does the degree of morphological similarity of the end-products.

3 The study of cranial endocasts was once largely dependent upon specimens with exposed and intact natural endocasts formed from matrix infilling of the spaces within the cranium. These natural endocasts were often difficult to obtain and provided data with limited detail (e.g. Edinger, 1975). Artificial endocasts can also be created through conventional means such as pouring a latex mold of the endocranial space, extracting that mold through the foramen magnum, and then using it to make a plaster cast (e.g.

Radinsky, 1968). However, because the cranial cavity in many intact specimens is often filled with matrix, creating artificial endocasts in this way is not typically possible.

Alternatively, serial sectioning of the skull (e.g. Sollas, 1904) or physical removal of the surrounding bones of the braincase in order to reveal a natural endocast (e.g. Hofer &

Wilson, 1967) can be used but are not commonly appropriate, especially when studying rare or unique specimens. Because of issues like these, paleoneurological studies are relatively rare (Edinger, 1975). Fortunately, the same technology that allows for a non- invasive, detailed study of the neurological morphology of extant specimens has been increasingly utilized in recent years to become a useful tool in doing the same types of modeling with fossil specimens (e.g. Conroy & Vannier, 1984; Carlson et al., 2003).

Because the technology used to create a virtual model of the internal cranial anatomy of a fossil specimen is relatively new and expensive and the criteria that a fossil must meet, like the degree of preservation and presence of certain cranial bones, is restrictive, a very small percentage of extinct taxa have been studied in this way. So far, the bulk of such studies has involved the most popular taxa, like many dinosaurs, or species with a particularly high amount of perceived phylogenetic or evolutionary importance, like Archaeopteryx . One clade that is not particularly well known or

4 understood, even within the vertebrate paleontological community, yet is potentially very

phylogenetically important and whose internal cranial anatomy has yet to be studied

using newly developed technologies and techniques is the order Phytosauria.

Background of phytosaurs

In order to understand the importance of phytosaurs and why a new study of the

internal cranial anatomy of phytosaurs would be of great benefit, it is necessary to first

understand what is already known about this clade. Archosauria can be divided into the

two unranked groups of Avemetatarsalia, which includes the ornithodiran groups

Dinosauromorpha and Pterosauria as well as the outgroup Scleromochlus (Benton, 1999), and Crurotarsi (Sereno & Arcucci, 1990), with shared derived conditions of the ankles being the most important characters. Crurotarsi is one of two terms typically applied to the latter group with Pseudosuchia being the other. The primary difference between these two terms is that Pseudosuchia is a stem-based clade, generally referring to crocodiles and all archosaurs closer to crocodiles than to birds (Gauthier & Padian, 1985; Gauthier,

1986; Senter, 2005) whereas Crurotarsi is node-based and refers to archosaurs with a crurotarsal ankle morphology (Sereno & Arcucci, 1990). Crurotarsi has been the term more often applied over last two decades, perhaps owing to the meaning of the term

Pseudosuchia as “false-crocodiles” while including true crocodiles within that group.

Phytosaurs are generally considered to be the basal-most member of the

Crurotarsi (e.g. Parrish, 1993) (Fig. 1.1) and are a group of extinct, semi-aquatic archosaurs belonging to the class Reptilia, division Archosauria, order Phytosauria, and

5 Figure 1.1. Cladogram illustrating the phylogenetic position of Phytosauria within Crurotarsi and the relationship to the rest of Archosauria. Modified from: Archosauromorpha to Crurotarsi, Nesbitt & Norrell (2006) and Nesbitt (2007); Crurotarsi to Crocodylomorpha, Parrish (1993), Nesbitt & Norrell (2006), and Brusatte et al. (2010); Crocodylomorpha to Crocodyliformes, Sereno et al. (2003) and Larsson & Sues (2007); Crocodyliformes to Eusuchia, Sereno et al., (2003), Larsson & Sues (2007), and Martin & Benton (2008); and Eusuchia, Martin & Benton (2008).

family Phytosauridae. Although temporally restricted, phytosaurs had a nearly global distribution, occurring in Upper Triassic sediments of North America, Europe, and northern Africa (e.g. Marsh, 1896; Doyle & Sues, 1995; Lucas, 1998; Hungerbuhler &

Hunt, 2000), India (Chatterjee, 1978), Turkey (Buffetaut et al., 1988), Thailand

(Buffetaut & Ingavat, 1982), Madagascar, (Gregory, 1969; Burmeister, 2000), and Brazil

(Kischlat & Lucas, 2003). The first phytosaur taxon to be described, Phytosaurus cylindricodon , was mistakenly identified as an herbivore (Jaeger, 1828). The order

6 designation Phytosauria was coined by von Meyer (1861) on the basis of Jaeger’s original description. Another name for the order, Parasuchia, was given by Huxley in

1875 during his description of the Indian taxon Parasuchus hislopi (Chatterjee, 1978).

These two order designations are used somewhat interchangeably seemingly based upon user preference for either the provenance of the former or the more accurate description provided by the latter (“alongside crocodiles” as opposed to “plant-eating lizards”), although the term “phytosaur” is generally recognized as referring to a member of this group. The lack of nomial clarity and agreement is nearly as much a hallmark of phytosaur studies as the anatomical description of specimens. Phytosaur systematics is generally regarded as problematic and is rarely the subject of large-scale research efforts.

Although some (i.e. pseudopalatine) phytosaurs have been the subject of relatively recent taxonomic revisions (e.g. Hungerbuhler, 2002), the rest of the group has been fairly neglected since Long and Murry (1995). A recent study by Stocker (2010) has focused on these previously overlooked taxa, particularly in reevaluating ‘ Leptosuchus ’, a poorly understood genus that had become somewhat of a “wastebasket taxon.” ‘Leptosuchus ’ is now considered paraphyletic, and consists of specimens only found in the Dockum Group of Northern Texas, with Arizona specimens now reassigned to Pravusuchus and

Smilosuchus (Stocker, 2010). Although the genus Rutiodon has been used to describe specimens from the western United States (e.g. Gregory, 1962; Ballew, 1989), Stocker

(2010) shows that Rutiodon should be restricted to material from the eastern United

States only. Furthermore, three species of Smilosuchus are now considered valid, with S. gregorii being the type species and both S. adamanensis and S. lithodendrorum being new combinations (Stocker, 2010). Smilosuchus adamanensis material had previously

7 been referred to ‘Leptosuchus’ adamanensis (Long and Murry, 1995), Rutiodon

(Gregory, 1962), and ‘Machaeroprosopus’ (Camp, 1930).

The oldest known phytosaur, Paleorhinus , occurs in earlier Late Carnian deposits of the Late Triassic throughout Europe, North America, North Africa, and India (e.g.

Chatterjee, 1978; Hunt & Lucas, 1991). These early specimens are found in deposits of freshwater lacustrine and braided river systems (Zatón et al., 2005; Dzik & Sulej, 2007).

The depositional settings in which phytosaurs are found remain similar throughout their occurrence (e.g. Lucas et al., 1994; Lehman & Chatterjee, 2005) and are generally similar to the freshwater lacustrine, stream, and swamp habitats of extant crocodilians. A somewhat more derived and larger form, Angistorhinus , appears at the end of or soon after Paleorhinus occurrences in western North America. Later in the Late Carnian, both of these taxa were replaced by more derived forms like Rutiodon , Leptosuchus , and

Smilosuchus in North America (e.g. Lucas, 1998; Heckert & Lucas, 2002; Lehman &

Chatterjee, 2005; Stocker, 2010). The Carnian-Norian extinction resulted in a succession of these phytosaur forms by new genera in the Early Norian like the regionally limited

Nicrosaurus of Europe and Pseudopalatus of western North America, both of which belong to the most derived clade of phytosaurs, the subfamily Pseudopalatinae (Lucas,

1998). Middle Norian deposits in Germany, northern Italy, and Thailand yield the widespread and highly derived Mystriosuchus (Hungerbühler, 2002). Postcranial anatomy of this genus suggests that it was more adapted to aquatic life than were other phytosaurs (Hungerbühler, 2002), and a specimen from northern Italy suggests that at least some Mystriosuchus specimens were capable of living in a marine habitat (Gozzi &

Renesto, 2003). In southwestern North America, the largest and most derived form,

8 Redondasaurus , supplants Pseudopalatus and lasts into the Rhaetian (Hunt & Lucas,

1993). Phytosaur footprints of the ichnotaxon Apatopus are known from the Late

Rhaetian of eastern North America (Olsen et al., 2002), indicating that phytosaurs

continued until the very end of the Triassic Period, at which time they, and many other

large members of Crurotarsi like aetosaurs, rauisuchians, and so on, went extinct as a

result of the end-Triassic extinction event (e.g. Olsen et al., 2002; Tanner et al., 2004).

Ankle types within the Archosauriformes fall into four categories. The first of

these is the primitive mesotarsal ankle, in which the astragalus and calcaneum are fixed

to the tibia and fibula, respectively, and the joint bends around the contact between these

bones and the foot in a somewhat flexible and loose way (Cruickshank & Benton, 1985;

Sereno & Arcucci, 1990). The archosauriform outgroup to Archosauria, Euparkeria ,

possesses this ankle type. The second type is a derived condition called crurotarsal in

which the astragalus is fixed to the tibia by a suture and has a peg-like projection that fits

into a socket in the calcaneum, which usually develops a large backward heel, and serves

as the rotation point for the joint (Sereno, 1990; Sereno & Arcucci, 1990). The Crurotarsi

family Ornithosuchidae is the only group to exhibit the third ankle type called reversed

crurotarsal (Cruickshank & Benton, 1985). The reversed crurotarsal ankle type is very

much like the crurotarsal ankle type except that it lacks the typical calcaneum heel and

the pivot-point structure is reversed, with the calcaneum having the peg and the

astragalus having the socket. The fourth is another derived form called the advanced

mesotarsal ankle in which a simple hinge develops between the enlarged astragalus plus

the reduced calcaneum and the distal tarsals or metatarsals. This is the ankle type seen in

birds and other dinosaurs (Cruickshank & Benton, 1985; Sereno & Arcucci, 1990).

9 The Phytosauria were the most common group of archosaurs in western North

America during the Late Triassic Period, the most common aquatic predators, and had an

overall body morphology (Fig. 1.2A), and probably lifestyle, very similar to extant

crocodilians (e.g. Hunt, 1989). The skulls of phytosaurs differ from those of crocodilians in a number of key ways, however. Although the skulls of phytosaurs exhibit an elongated rostrum similar to that of a crocodilian, the external nares of phytosaurs are located posteriorly on the rostrum at or behind the mid-point of the skull and often just anterior to the orbits (Fig. 1.2B) (e.g. Case, 1929; Eaton, 1965). The earliest phytosaurs already possessed posteriorly placed external nares. Their location in the earliest form,

Paleorhinus , suggests that this trait was derived from an ancestor that would have had external nares on the tip of the snout (Fara & Hungerbühler, 2000). This view is supported by the snout primarily consisting of a shortened maxilla and elongated premaxilla, rather than the reverse as is seen in other Crurotarsi (Eaton, 1965).

10

Figure 1.2. ( A) Mounted skeleton of the phytosaur Redondasaurus bermani at the Carnegie Museum of Natural History. ( B) Digital reconstruction of the skull of Smilosuchus adamanensis . The external nares (indicated by the yellow arrow) are located just dorsal and anterior to the orbits (outlined in red).

Background of CT study techniques

Much of the information gathered about a given paleontological specimen is dependent upon ever-advancing technologies and techniques for studying them. These often include technologies developed for use in other fields that are then adapted to fit the

11 needs of paleontologists. One such technology that has only relatively recently become available for use in paleontological studies is X-ray computed tomography (CT) (Conroy

& Vannier, 1984; Carlson et al., 2003). CT is a medical imaging method that utilizes tomography created by computer processing. Tomography is imaging by sections, through the use of any type of penetrating wave – in this case X-rays. X-ray section, or slice, data are generated using an X-ray source that rotates as part of a ring-shaped gantry around the specimen being scanned which is placed on a bed (Fig. 1.3). Sensors consisting of scintillation systems based on photo diodes are positioned on the opposite side of the rotating ring, and the abilities of the various structures and materials present within the specimen to block the X-ray beam result in a greater or lesser amount of X- rays that pass through the specimen and are received by the sensors. Helical or spiral CT machines take many progressive scans and process continuously changing cross sections as the specimen, on the bed, slowly passes through the plane of the scanning gantry, recording cross-sectional data in three planes: coronal, sagittal, and transverse (Herman,

2009). The series of cross-sectional images is then saved as an archive of serial DICOM type computer files which must then be processed using a form of tomographic reconstruction.

In paleontological studies, tomographic reconstruction is typically performed manually with the aid of segmentation software that organizes and stacks the cross- sections along a single axis of rotation for each plane of segmentation. Stacking slices in series, scan data can be compiled into a single volumetric representation of the original material that was scanned. By doing so, it is possible to create a virtual three- dimensional model of internal features that can be manipulated in real-time, allowing

12 A

B

Figure 1.3. ( A) Example of a Phillips 64 slice “Brilliance” Scanner (Easton, 2005) The specimen is placed on the bed, which passes through the center of the gantry while the gantry rotates. ( B) CT scanner with the cover removed, demonstrating the principle of operation. The X-ray tube and the detectors are mounted on a ring shaped gantry. R, direction of rotation; T, X-ray tube; D, X-ray detectors; X, direction of X-ray transmission.

13 researchers to view the model just as they would the real structure (Fig. 1.4). Each section is composed of pixels that are displayed in terms of relative radiodensity with more dense materials displaying lighter pixel shades and less dense materials displaying darker pixel shades. The pixel itself is a two-dimensional unit, displayed according to the mean attenuation of the tissue to which it corresponds. When the CT slice thickness is also factored, essentially creating a three-dimensional unit of density representation, the unit is called a voxel. Because the attenuation of a voxel is uniform, some segmentation software allows users to resample and reduce voxel dimensions, effectively reducing the volume represented by each voxel and changing what was once a single voxel of uniform attenuation into a computer calculated volume of many smaller voxels, each with their own uniform attenuation displays (Witmer & Ridgely, personal communication,

December 10, 2010). This process has the effect of turning one area of a section that was displayed as a single density by a single voxel into an area that is displayed as a more fine-scale gradation of densities by many smaller voxels. Manual segmentation helps to compensate for artifacts and other anomalies common to CT scans of paleontological specimens. One of these, the partial volume effect, occurs when the detector cannot differentiate between different materials. An example of this effect in a recent specimen would be when a large amount of cartilage and thin layer of compact bone cause the same attenuation in a voxel as would hyper-dense cartilage alone (Herman, 2009). In a fossil specimen, a similar problem could occur when a large amount of less dense matrix and a thin layer of fossil material occupy the same voxel. By resampling and resizing the voxels and using anatomical knowledge, a researcher would be able to discern between these two materials in a way that an automatic reconstruction by the CT computer could

14

Figure 1.4. Example of a three-dimensional, virtual reconstruction of the endocast and cranial pneumaticity of Tyrannosaurus rex from CT segmentation data. From Witmer & Ridgely (2008a).

15 not. Another potential problem is the occurrence of artifacts, caused by abrupt transitions between low density and high density materials and resulting in data values that exceed the dynamic range of the processing electronics (Herman, 2009). In particular, very dense materials like iron can completely extinguish the X-ray and result in streak artifacts

(Fig. 1.5).

As CT technology has advanced in recent years, it has become an increasingly viable method of paleontological study. Histroically, research into the cavities within the skulls of paleontological specimens has relied on the presence of and access to natural matrix infilling of those cavities (Edinger, 1975), resulting in an area of study reliant upon rarely occurring specimens and with a relatively low degree of detail in the data available. In recent specimens, using CT techniques instead of traditional dissection methods can allow researchers to study various aspects and features of neuroanatomy without altering the surrounding tissue. This type of study can provide a great deal of new information, particularly for delicate structures like blood vessels that may become damaged or dissociated from other features during more invasive studies (Holliday et al.,

2006). Although the internal structures of the brain of fossil specimens may remain impossible to determine, even with the use of CT, detailed surface features may still be seen by virtually filling in the empty space within the braincase of the specimen (e.g.

Conroy & Vannier, 1984; Carlson et al., 2003; Witmer et al., 2008). Scans of a skull specimen can be used to create a virtual three-dimensional model to illustrate the relative sizes, locations, orientations, and connections of various neural structures (e.g. Witmer et al., 2003; Domínguez Alonso et al., 2004; Franzosa & Rowe, 2005; Witmer et al., 2008).

These virtual endocasts can serve as an excellent source of information that can be added

16 Figure 1.5. Example of a transverse CT slice. Lighter regions, like that marked by the red arrow, indicate greater density and darker regions, like that marked by the blue arrow, indicate lower density. The yellow circle outlines a region of material dense enough to cause artifacts, in this case the light bands radiating outward.

to previous knowledge about a specimen in order to create a more complete and accurate picture of the lifestyle of the studied animal (Rogers, 2005). The orientation of the semicircular canals and other structures of the inner ear, for example, can be used to determine the head posture of the animal when it was alive (Witmer et al., 2003;

Domínguez Alonso et al., 2004). Similarly, a great deal of information about how the animal collected and analyzed sensory information can be obtained and used to make further inferences about its lifestyle (e.g. Rogers, 2005; Witmer & Ridgely, 2009; Witmer et al., 2010). For example, an enlarged area of the brain that deals with visual or auditory stimuli could indicate an increased importance of one of these senses to the animal’s

17 lifestyle, whereas a decrease in the size and development of such a feature can indicate a reduction in the degree to which the animal relied on the related sense (Witmer et al.,

2010). Such inferences are based on the “principle of proper mass” which states that the mass of the neural tissue of a particular region of the brain is correlated with the amount of information processing involved in performing the function associated with that region of the brain (Jerison, 1973). An inferred importance of sight to the animal may indicate that it was an active predator or lived in some other way that made the accurate judgment of distances a requirement for its continued survival (Witmer & Ridgely, 2009).

Likewise, a severely heightened sense of smell could suggest that the animal lived as a scavenger, a hunter that needed to track prey long before actually seeing it, or in some equally scent-dependent way (Witmer & Ridgely, 2009). In this way, endocast data collected and analyzed through CT segmentation techniques have the potential to impact a host of functional, behavioral, and systematic hypotheses.

Rationale

Most of the studies conducted on phytosaurs have involved morphological descriptions of cranial skeleton structures in order to make phylogenetic and taxonomic comparisons. Many of the early studies focused on external cranial morphology as a way to determine taxonomic status with little attention given to the varied functions served by the often distinctive features like rostral crests (Irmis, 2005; Parker & Irmis, 2006).

Some of these features have since been shown to be less useful in determining phylogenetic relationships between taxa than was originally believed. The presence or lack of a rostral crest, for example, has at times been used as evidence for sexual

18 dimorphism and a tool for phylogenetic analysis (e.g. Abel, 1922). Similarly, co- occurrence of taxa of both robust and gracile snout morphotypes has led to suggestions that those taxa are sexual dimorphs (Zeigler et al., 2002; 2003). However, the lack of co- occurrence of some taxa argued to be sexually dimorphic based on rostral crests (Camp,

1930), co-occurrence of slender and broad-snouted extant crocodiles (Webb & Manolis,

1989), and independent evolution and/or loss of rostral crests in multiple phytosaur lineages (Irmis, 2005; Parker & Irmis, 2006) suggest that, rather than being a diagnostically valid character viable as a means of concluding phylogenetic relationships among taxa, something as basic and immediately apparent as rostrum shape can be a source of serious confusion and disagreement in phytosaur phylogeny. As a result, historical methods of classification tend to create at least as many questions as answers as evidenced by the dramatic lack of consensus among paleontologists as to the true status of phytosaur phylogeny and taxonomy.

Recent comparative research has dealt with more clearly useful features of the skull and even suggests a striking transitional nature of the evolution of classic phytosaur traits (Lucas et al., 2007; Woody & Parker, 2004; Irmis, 2005; Parker & Irmis, 2006).

However, the internal cranial anatomy of phytosaurs remains an area of study that has yet to receive a similar benefit of recent satisfactory study utilizing modern techniques.

Although partially due to the braincase not being preserved or described in detail for most phytosaur specimens (Parker & Irmis, 2006), this oversight is in striking contrast to the wealth of attention paid to the internal cranial morphology of other archosaurs such as pterosaurs and avian and non-avian dinosaurs. The contributions of such studies on other archosaur taxa have had an undeniable impact on the understanding of these taxa in terms

19 of morphological, functional, behavioral, and systematic knowledge. Because the application of CT technologies and techniques to the study of fossil specimens is relatively new and expensive, perhaps it was inevitable that the Avemetatarsalia would be among the earliest subjects for these types of studies, given the popularity of dinosaurs and pterosaurs within the general public and the high level of interest in the evolutionary origins of birds held by many paleontologists. Still, the morphological, behavioral, and habitat variation seen among crocodile-line archosaurs rivals, if not exceeds, that seen among dinosaurs. As a result, the internal cranial anatomy of the Crurotarsi promises to be an area of study with a great amount of potential for new understanding.

Many of the questions about neurological morphology and its relationship to behavior that are often asked about dinosaurs can just as easily be asked about Crurotarsi.

The answers to such questions have the potential to not only explain evolution throughout the Crurotarsi but also to illuminate the mechanisms of behavior and convergence seen across all clades within Archosauria. Furthermore, with extant crocodilians having already been the focus of a number of internal cranial morphology studies and due to the evolutionary convergence of phytosaurs and extant crocodilians, reasonable and appropriate comparative data exist for a successful study of phytosaur endocasts. Given the basal position of phytosaurs within the Crurotarsi, phytosaurs are also a reasonable clade to be used as a first step in a larger study to determine what, if any, neurological convergence or divergence may exist throughout the Crurotarsi or even between members of Crurotarsi and Avemetatarsalia. The objective of this study, then, is to apply established methods for creating and studying a virtual endocast of a phytosaur specimen for the purposes of comparison with those of extant crocodilians. It is hypothesized that

20 because of the high level of morphological and ecological convergence between

phytosaurs and extant crocodilians, the endocasts of these two groups will exhibit a great

deal of overall similarity and be more morphologically similar to one another than either

are to other archosaurs that do not share such convergence.

Chapter 2 - Materials, methods, and results.

Materials

The specimen used in this study was chosen because it is of a genus of neither the

most basal (i.e. Paleorhinus ) nor the most derived (i.e. Pseudopalatus and Redonasaurus )

North American forms of phytosaurs (Stocker, 2010) and therefore serves as a typical representative of the order in North America. Additionally, access to the study specimen was aided by the relatively close proximity of the collection in which it resides. This study is the first stage of a larger planned project on descriptive, functional, and comparative Crurotarsi endocast morphology, and it is intended that additional phytosaur specimens of all possible genera will be scanned and reported on in the future.

The taxonomic identification and provenance of the specimen is as follows:

Smilosuchus adamanensis (National Museum of Natural History, Smithsonian

Institution, Washington, D. C., USNM V 17098): USNM 17098 (Fig. 2.1) is a nearly complete skull (skull length 63.94cm) of a moderately-sized, robust-snouted, and high- crested individual. The skull is complete, with the notable exceptions of the lower jaw,

tip of the snout, and all teeth. There is a fair amount of both plastic and brittle

21 deformation evident from the external portion of the skull, yet the large amount of matrix in-filling and lack of apparent bonding agents suggests that little, if any, reconstructive

A

B

Figure 2.1. Skull of Smilosuchus adamanensis (USNM 17098) in left lateral view ( A), showing examples of the numerous brittle deformation breaks throughout the specimen, and dorsoanterior view ( B), illustrating torsion effects of plastic deformation on the rostrum and right-hand side of the specimen.

work was performed on this specimen. At the time of loan, this specimen was listed in the NMNH Department of Paleobiology Collections catalogue as Machaeroprosopus

22 zunii . This specimen is of a size and overall skull morphology, particularly with regards

to the snout, consistent with descriptions of M. zunii by Camp (1930) and Colbert (1947).

However, based on systematic work by Long and Murry (1995) and, more recently,

Stocker (2010), this specimen is here referred to Smilosuchus adamanensis (see Chapter

1). The specimen is listed in the museum collection catalogue as having been collected by G. Hazen et al. in September, 1946 from the Late Triassic Chinle Formation, Arizona.

CT Scanning and 3D Visualization

The specimen was CT scanned at Cabell Huntington Hospital, Huntington, West

Virginia, using a General Electric (GE) Healthcare LightSpeed Xtra with the Extended

Hounsfield option, which greatly improves resolvability of detail from dense objects such as fossils by extending the dynamic range of images as much as 16-fold, and a bow-tie filter, which decreases beam-hardening artifacts. The specimen was scanned helically at a slice thickness of 0.625 mm, 120 kV, and 200 mA. The raw scan data were reconstructed using a bone algorithm and exported from the scanner computer in DICOM format. These files were then imported into Amira 4.0.1 (Mercury-TGS, Chelmsford,

MA) for viewing, analysis, and visualization. The scan data were analyzed on a 32-bit

PC workstation with 2 GB of RAM, an Intel Core 2 Duo E6850 (4 MB Cache, 3.0 GHz) processor, and nVidia Quadro FX 3000 video card installed with Microsoft Windows XP

Professional operating system. The data set was cropped to include only the area of the scan field that included data for the fossil, cutting out data representing empty space, the scanning bed, etc., and resampled to a voxel size of 0.075 x 0.075 x 0.075 mm in order to create a more detailed model of the skull. The raw data were then treated again in the

23 same way, but with a cropping area including only anatomical features of interest to this study (i.e. cranial endocast, endosseous labyrinth, etc.) and a voxel size of 0.035 x 0.035 x 0.035 mm. The regions of space occupied by the anatomical features of interest were then highlighted and digitally extracted through a series of sections along the transverse, coronal, and sagittal planes and compiling those sections into a three-dimensional virtual model using Amira’s segmentation tools for quantification and visualization (Fig. 2.2).

24

Figure 2.2-A. Screenshot of Amira 4.0.1 (Mercury-TGS, Chelmsford, MA) CT data visualization software in use. The windows within this screenshot are: the model in progress (upper left) and the specimen in sagittal (upper right), coronal (bottom right), and transverse (bottom left) sections. This screen shows the appearance of the data before any features have been digitally extracted. Areas of lower density appear as a darker shade of grey and generally represent space filled with matrix while areas of higher density appear lighter and generally represent space filled with fossilized bone.

25

Figure 2.2-B. Screenshot of Amira 4.0.1 (Mercury-TGS, Chelmsford, MA) CT data visualization software in use. The windows within this screenshot are: the model in progress (upper left) and the specimen in sagittal (upper right), coronal (bottom right), and transverse (bottom left) sections. This screen shows a portion of the data, in transverse section, representing the endocranial cavity that has been highlighted (red) and added to the model in progress. When a portion of the data is highlighted in one section, it is also highlighted in the other two sections, at illustrated by the red line shown in the sagittal and coronal views.

26

Figure 2.2-C. Screenshot of Amira 4.0.1 (Mercury-TGS, Chelmsford, MA) CT data visualization software in use. The windows within this screenshot are: the model in progress (upper left) and the specimen in sagittal (upper right), coronal (bottom right), and transverse (bottom left) sections. This screen shows the appearance of the model in progress once a series of sections of the endocranial cavity have been highlighted (red) in all of the three planes.

27

Figure 2.2-D. Screenshot of Amira 4.0.1 (Mercury-TGS, Chelmsford, MA) CT data visualization software in use. The windows within this screenshot are: the model in progress (upper left) and the specimen in sagittal (upper right), coronal (bottom right), and transverse (bottom left) sections. This screen shows the model in progress once the highlighted regions illustrated in Fig. 5-C have been saved as a material type (blue). Additional portions of the data, in transverse section, representing the optic nerve (CN II) have been highlighted (red) and added to the model in progress.

28

Figure 2.2-E. Screenshot of Amira 4.0.1 (Mercury-TGS, Chelmsford, MA) CT data visualization software in use. The windows within this screenshot are: the model in progress (upper left) and the specimen in sagittal (upper right), coronal (bottom right), and transverse (bottom left) sections. This screen shows the appearance of the model in progress once a series of sections of the optic nerve (CN II) have been highlighted (red) in all of the three planes. Once these regions have been saved to the model as a new material type, additional regions representing other features of anatomical interest will be treated in the same way.

Cranial Endocast

The digital endocast of the Smilosuchus adamanensis study specimen (USNM

17098) shows a fair amount of detail and allows for a description of the space once occupied by the brain, brainstem, endosseous labyrinth, cranial nerves, and craniocerebral vascular elements. As previously noted by numerous studies (e.g., Jerison, 1973;

Hopson, 1979; Witmer et al., 2008; Evans et al., 2009), the brains of most archosaurs,

29 with the exception of birds, fill only a relatively small portion of the endocranial cavity, unlike the condition in mammals and birds. In crocodilians, in particular, less than fifty percent of the volume of the endocranial cavity is filled by the brain with the remainder of the space consisting of the dura mater and venous sinuses within the dura (Hopson,

1979). This endocranial composition means that the morphology of deeper endocranial structures like the cerebellum are commonly overlain and obscured by relatively superficial structures within the endocranial cavity such as the dural venous sinuses, resulting in an endocranial endocast that more closely represents the contours of the external surface of the dura mater and other superficial structures where it was in contact with the internal surface of the braincase, rather than the actual morphology of the brain and brainstem. The specific morphology of certain structures, such as the position, orientation, size, and general organization of the cranial nerve trunks relative to elements of the braincase, can be determined, however. Additionally, because of the close relationship between the deeper structures of the brain and the overlying dura mater, many of the deeper features impart some information, like their position and relative size, on the internal surface of the braincase via the dura mater (e.g., Kley et al., 2010). Many features of the dura mater itself, such as many of the dural venous sinuses, constitute some of the most easily and accurately reconstructed elements of endocranial endocasts

(Kley et al., 2010).

The skull of this specimen exhibits a fair amount of both plastic and brittle deformation. This deformation is most evident in the torsion effects seen in the snout and the numerous examples of breakage throughout the skull, some of which pass entirely through the skull along a given plane (Fig. 2.1). A notable example of breakage involves

30 the cranial nerve roots, most of which are truncated or entirely obscured by a rather large

break that traverses the dorsal portions of the basisphenoid and basioccipital and the

ventral portion of the endocast along a horizontal plane that is largely the same as that of

most of the cranial nerves (Fig. 2.3). This same break is also near enough to the ventral

portion of the inner ear to preclude description and analysis of inner ear structures such as

the vestibule and cochlea. However, a great deal of the braincase is intact, at least to the

extent that a feature missing from one lateral view of the endocast is typically present in

the opposite view (Fig. 2.4B).

The general shape of the endocast (Fig. 2.4) is largely similar to those of extant

crocodilians like Gavialis gangeticus (Wharton, 2000), Alligator mississippiensis

(Witmer & Ridgely, 2008a), and Crocodylus johnstoni (Witmer et al., 2008). The olfactory apparatus is well defined, as constrained by the internal surface of the frontal and the dorsal surface of the interorbital septum. The olfactory tracts and bulbs are of a form and relative size similar to those of extant crocodilians (e.g. Witmer et al., 2008).

Additionally, the olfactory bulbs are exposed to the olfactory region of the nasal cavity along the ventral margin of their anterior-most terminus. The cerebral hemispheres are broad, relative to the width of the rest of the endocast, and taper both anteriorly and posteriorly (Fig. 2.4A, B). Postcerebral portions of the endocast exhibit an indistinct neural morphology.

A recent study of the cranial endocast of another phytosaur, Pseudopalatus pristinus , by Smith et al. (2010), showed no indication of an enlarged epiphysis (pineal body) and concluded that the referred structure in Smilosuchus gregorii probably corresponds to the longitudinal sinus. An enlarged, dorsally projecting prominence of

31

Figure 2.3. Reconstruction of the cranial cavity of Smilosuchus adamanensis (USNM 17098) in left lateral view. The orange line follows the path of a large break that transects the ventral section of the braincase interior. This break obscures most of the cranial nerve and ventral labyrinth data and truncates the ventral portion of the trigeminal nerve (CN V).

dura mater marking the junction of the fore- and midbrain is present in S. adamanensis

(Fig. 2.4) and is one of the more clearly defined features seen in the scan data. The

position and morphology of this prominence is consistent with that of the pineal body in

members of Avemetatarsalia (e.g. Rogers, 1999; Larsson, 2001; Sampson & Witmer,

2007) and epiphysis of ‘ Machaeroprospus ’ gregorii (Camp, 1930). This feature in S.

adamanensis is also well outside the expected size range of any dural venous sinus and

inconsistent with the typical morphology thereof. Therefore, the prominence seen in S.

adamanensis is here referred to the pineal body.

Strangely, the hypophyseal fossa is not apparent. Rather than being obscured by

breakage, as is the case with the entirety of most of the cranial nerve roots and the ventral

portion of the trigeminal nerve (CN V) (Fig. 2.3), the reason that the hypophyseal fossa is

not readily identifiable is unclear. One possible explanation is that this portion of the

endocranial cavity experienced such a high level of deformation that the space once

occupied by the hypophysis was essentially filled by the surrounding bones compacting

32

Figure 2.4. Reconstruction of the cranial cavity of Smilosuchus adamanensis (USNM 17098). Digital endocast in dorsal ( A), ventral (B), left rostrodorsolateral ( C), left caudodorsolateral ( D), and left lateral ( E) views generated from CT scans. Selected cranial nerve trunks, the endosseous labyrinth, and cranial vasculature are illustrated in yellow, green, and dark blue, respectively. See Appendix 1 for anatomical abbreviations.

33 toward each other. However, the lack of a noticeable boundary indicating a point of

contact between inward collapsed bones in any of the cross-sectional views and the

relatively localized nature of this hypothesized deformation make this an unlikely cause

for the inability to locate this feature. Instead, the most likely reason is that the matrix

infilling this portion of the cavity is of a density so similar to that of the surrounding

fossil material that attempts to differentiate the two were unsuccessful.

The most distinctive features of the cranial endocast are found along its dorsal

surface. In at least some extant crocodilians, such as Crocodylus johnstoni (Witmer et

al., 2008), the dorsal contour of the endocast is generally curvilinear. The dorsal contour

of the endocast of the Smilosuchus adamanensis (Fig. 2.4E) can, instead, be described in

terms of three distinct linear surfaces. The anterior segment consists of the olfactory

apparatus and is nearly linear along the horizontal plane. The middle segment is

represented by the dorsal longitudinal dural venous sinus, cerebrum and a large,

distinctive pineal body. The contour of the posterior segment is curvilinear but is

dorsoventrally concave, unlike the convex contour in C. johnstoni (Witmer et al., 2008), and is represented by the the occipital dural venous sinus complex, overlying the dorsal aspect of the medulla. Smilosuchus adamanensis is further differentiated from C. johnstoni in having a cerebrum that is expanded more dorsally than any other feature of the endocast, save the pineal body (Fig. 2.4E). Kley et al. (2010) noted a similar three- sectioned dorsal contour of the notosuchian Simosuchus clarki . However, the constituencies of each of the segments is different from those seen in S. adamanensis .

The position of the pineal body is clearly defined by the dorsally projecting prominence

34 of dura mater that marks the junction of the fore- and midbrain, just posterior to the cerebrum.

Much of the dural venous sinus system is illustrated in the reconstruction (Fig.

2.4). The dorsal longitudinal venous sinus overlays the olfactory apparatus, including the olfactory lobe and the anteriormost process of the cerebrum. Proceeding posteriorly, the dorsal longitudinal venous sinus then becomes undifferentiated within the dorsal expansions of the cerebrum and pineal body before being evident once again just posterior to the pineal body, where it presumably overlays tectal and cerebellar structures.

The occipital dural venous sinus overlies, and therefore obscures details of, the tectum, cerebellum, and medulla. This is consistent with the typical condition for archosaurs

(Sedlmayr, 2002; Witmer et al., 2008). The sphenoparietal dural venous sinus is transversely oriented, occupying the postcerebral constriction between the cerebrum and tectum, and would have connected the dorsal longitudinal sinus to the cavernous sinus lateral to the hypophysis, as well as other ventrally positioned sinuses. The transverse dural venous sinus occupies the tectal-otic sulcus between the optic tectum and otic capsule (see Witmer et al., 2008; Kley et al., 2010). The ventral longitudinal dural venous sinus is identifiable at the posteriormost portion of the ventral surface of the endocast where it overlies and obscures the medulla, ventrally.

Most features of the cerebellum are obscured in the cranial endocast reconstruction of Smilosuchus adamanensis , much as they are in other non-avian archosaurs (Sampson & Witmer, 2007; Kley et al., 2010), due to a well-developed occipital dural venous sinus system. Again, as is the exception exhibited by most other non-avian archosaurs (Kley et al., 2010), the flocculus is distinguishable just anterior to

35 the anterior semicircular canal (Fig. 2.4E). The relative size of the flocculus has been

used to model cervicocephalic, whole-body, and reflexive ocular-stabilization kinematics

in a variety of extinct archosaurs (Witmer et al., 2003; Sampson & Witmer, 2007).

Although the relatively small size of the flocculus in S. adamanensis may suggest a somewhat limited capacity in those areas, additional comparative data and a specimen with much greater detail of that feature would be required to reach any reliable conclusions about those kinematics in phytosaurs.

Endosseous Labyrinth

Although incomplete due to breakage that truncates the ventral margin of the lateral semicircular canal and makes a reconstruction of the more ventral components of the vestibular apparatus impossible, the features of the inner ear of Smilosuchus adamanensis that can be digitally reconstructed are similar to the condition in extant crocodilians and, indeed, a variety of other non-avian archosaurs (e.g., Witmer et al.,

2008; Evans et al., 2009; Kley et al., 2010). The three semicircular canals are oriented, as they are in all tetrapods, at approximately right angles to one another in the three planes of space. The anterior semicircular canal is slightly longer and more arched than the posterior semicircular canal, and the lateral semicircular canal is the shortest of the three.

The lateral semicircular canal has been used to infer the alert head posture of various extinct archosaurs by orienting the head such that the lateral semicircular canal is parallel to Earth horizontal (e.g., Witmer et al., 2003; Sereno et al., 2007; Witmer &

Ridgely, 2009). Recent reviews (Hullar, 2006; Taylor et al., 2009) have noted that experiments on extant animals have shown the resting but alert head posture to be such

36 that lateral semicircular canals are inclined slightly anterodorsally, relative to Earth

horizontal (e.g., Vidal et al., 1986; Erichsen et al., 1989; Graf et al., 1995). This position

has also been successfully used to support head posture estimates made by using other

evidence in at least one fossil taxon (Kley et al., 2010). By using the head posture typical

of Crocodylus johnstoni (Witmer et al., 2008) as a model, the skull of Smilosuchus adamanensis was oriented in an alert posture (Fig. 2.5). In this position, the lateral semicircular canals are anterodorsally inclined by ~ 9.5˚, consistent with the 5-15˚ range seen in Rock Pigeons ( Columbia livia ) in a resting, alert position (Erichsen et al., 1989) and used by Kley et al. (2010) to substantiate inferred head posture in Simosuchus clarki .

37

Figure 2.5. Three-dimensional reconstruction of the skull of of Smilosuchus adamanensis (USNM 17098), generated from CT scans. The skull, in left lateral view, is oriented according to the inferred resting, alert head posture, with the orientation of the lateral semicircular canal (solid line) relative to Earth horizontal (dashed line) indicated.

Chapter 3 - Discussion and Conclusions

Endocast comparison

The digitally reconstructed endocast of Smilosuchus adamanensis (USNM 17098) offers an excellent opportunity for comparison between the endocasts of the basalmost and crown group members of Crurotarsi (Fig 1.1). Furthermore, it allows for comparison of the cranial endomorphologies of two animals considered to have had similar ecologies due to their convergent overall morphologies and discovery in similar depositional environments. Based on previously noted convergence in cranial endocast morphology between other archosaurs with similar ecologies (Domínguez Alonso et al., 2004;

Witmer, 2004), it seems likely that phytosaurs and extant crocodilians would share convergent endocast morphologies. This does appear to be the case, as the general shape

38 of the endocast of the phytosaur Smilosuchus adamanensis is largely similar to those of

extant crocodilians like Gavialis gangeticus (Wharton, 2000), Alligator mississippiensis

(Witmer & Ridgely, 2008a), and Crocodylus johnstoni (Witmer et al., 2008).

The olfactory bulbs of Smilosuchus adamanensis are similar in size and shape to those of an extant crocodilian (Fig. 3.1). The olfactory tract is also of a similar size and length although the angle at which it projects from the rest of the endocast in S. adamanensis is considerably more horizontal than the somewhat more downward- oriented olfactory tracts of Alligator mississippiensis (Witmer & Ridgely, 2008a), and

Crocodylus johnstoni (Witmer et al., 2008). This difference in orientation is likely the result of differences in the planar organization of major skull elements (Kley et al., 2010) than significant functional differences. In S. adamanensis , the organization and morphology of many of the elements associated with the tectum, cerebellum, and medulla are obscured by various dural venous sinuses, just as is the case in extant crocodilians

(Sedlmayr, 2002; Witmer et al., 2008). Moreover, because each of these dural venous sinuses exhibits an organization similar to what is seen in extant crocodilians (Witmer et al., 2008) it is reasonable to project a similar organization of the endocranial elements overlain by and closely associated with these sinuses. This inference is supported by the identification of elements demarcated by their lateral surfaces, like the cerebrum and flocculus, the locations of which are consistent with those seen in extant crocodilians

(Witmer et al., 2008). The portions of the endosseous labyrinth of S. adamanensis for which reconstruction was possible also exhibit a similar morphology but slightly greater size, relative to the overall size of the endocast, to that of extant crocodilians (Witmer et

39 Figure 3.1. Endocasts of Smilosuchus adamanensis (A) and Crocodylus johnstoni (D). Gross organization of the two endocasts is similar, although the smaller cerebrum and larger pineal body of S. adamanensis are consistent with the primitive condition. C. johnstoni modified from Witmer et al. (2008).

al., 2008). The large size of the endosseous labyrinth is consistent with previous

descriptions of phytosaurs (e.g. Camp, 1930).

Although the overall morphologies, as well as the organization of many specific

features, of the endocrania of Smilosuchus adamanensis and extant crocodilians are

largely similar, a few differences can be discerned (Fig. 3.1). The cerebrum of S.

adamanensis , for example, exhibits a similar morphology to that of an extant crocodilian,

yet cerebrum size, relative to the size of the overall endocast, and the extent of lateral

expansion of the cerebrum are both smaller in S. adamanensis . This description is consistent with the morphology of the cerebrum of the more derived phytosaur

40 Pseudopalatus pristinus (Smith et al., 2010). A small relative size of the flocculus in S. adamanensis can also be seen, although deformation in this region of the specimen is significant enough to make precise relative size determinations difficult. One feature that has a significantly greater relative size in S. adamanensis compared to extant crocodilians is the dorsally located dural expansion that denotes the position of the pineal body (e.g.

Rogers, 1999; Larsson, 2001; Sampson & Witmer, 2007). A small cerebrum and flocculus and a large pineal body are all consistent with descriptions of primitive conditions in more basal reptiles (e.g. Hopson, 1979).

Excluding some minor differences, like those of dorsal surface contours and orientation angles of olfactory tracts that are largely attributable to slight differences in skull morphologies, the cranial endocast morphologies of extant crocodilians are very similar (Witmer et al., 2008; Kley et al., 2010). Such similarity would be expected among animals that not only share a similar ecology but are also relatively phylogenetically close to one another. Because of the overall similarities between the endocasts of the Smilosuchus adamanensis and those of extant crocodilians, it appears that not only is the commonly reconstructed ecology of phytosaurs as being similar to that of extant crocodilians supported, but that the similarity of ecologies seems to have a strong relationship with endocranial morphological convergence.

Comparison of archosaurian endocasts

The overall similarity of the endocranial morphologies of Smilosuchus adamanensis and extant crocodilians is significant and appears to support the hypothesis that overall morphological and ecological convergence tracks along with endocranial

41 morphological convergence. However, this statement has the caveat that so few

Crurotarsi have been the subjects of digital cranial endocast reconstructions and the phylogenetic positions of phytosaurs and crocodilians are so disparate, being basalmost and crown group members, respectively (Fig. 1.1), that there is little additional data to serve as comparison or control group. Fortunately, a very recent study by Kley et al.

(2010) does provide a description of the digitally reconstructed endocranium of the notosuchian Simosuchus clarki . Simosuchus clarki exhibits both a very distinct ecology, as a fully terrestrial herbivore, and morphology, being a relatively small (length of

~0.75m), short tailed notosuchian (Georgi & Krause, 2010) with osteoderm-covered limbs (Hill, 2010), a pug-nose, and clover-shaped teeth (Kley et al., 2010). Indeed, the numerous distinctive features and synapomorphies of S. clarki allow it to be used as an excellent control group in this study.

Interestingly, the overall morphology of the cranial endocast of Simosuchus clarki is very similar to those of extant crocodilians (Fig. 3.2), as noted by Kley et al. (2010).

This similiarity is surprising given the large number of differences seen in the overall morphologies and ecologies of these two taxa and suggests that the similarities seen between the endocasts of Smilosuchus adamanensis and extant crocodilians are not necessarily linked to overall morphological or ecological convergence. However, given the relatively close phylogenetic positions of notosuchians and crocodilians (Fig. 1.1), the similarities seen may be due to morphological conservatism of the endocranium among phylogenetically close taxa. Unfortunately, in order to add additional Crurotarsi taxa to the body of comparative data, endocranial reconstructions obtained by means other than

42 Figure 3.2. A series of endocasts of the Crurotarsi and the phylogenetic relationships of the taxa they represent. A comparison of the endocasts of: Smilosuchus adamanensis (A), Shuvosaurus inexpectatus (B) modified from Lehane (2005), Simosuchus clarki (C) modified from Kley et al. (2010) and Crocodylus johnstoni (D) modified from Witmer et al. (2008). Red arrows indicate the peak of the dural expansion, when present, indicating the position of the pineal body. Cladogram modified from: Archosauromorpha to Crurotarsi, Nesbitt & Norrell (2006) and Nesbitt (2007); Crurotarsi to Crocodylomorpha, Parrish (1993), Nesbitt & Norrell (2006), and Brusatte et al. (2010); Crocodylomorpha to Crocodyliformes, Sereno et al. (2003) and Larsson & Sues (2007); Crocodyliformes to Eusuchia, Sereno et al., (2003), Larsson & Sues (2007), and Martin & Benton (2008); and Eusuchia, Martin & Benton (2008).

digital reconstruction must be considered, and even so, very few additional descriptions exist. Still, a thesis by Lehane (2005) does include a drawn reconstruction of the cranial endocast of the poposaur Shuvosaurus inexpectatus , incorrectly interpreted at the time as an ornithomimid dinosaur (Chatterjee, 1993; Nesbitt, 2007). As is the case with

Simosuchus clarki , Shuvosaurus inexpectatus exhibits a number of distinctive overall morphological and ecological features that allow for its use as an effective control in this

43 study. Specifically, S. inexpectatus was a fully terrestrial, bipedal herbivore with a toothless beak (Chatterjee, 1993; Nesbitt, 2007). Given these features and a number of other synapomorphies, along with the increase in phylogenetic distance between poposaurs and crocodilians over that of notosuchians and crocodilians, a similarity between the overall morphologies of the cranial endocasts would seem unlikely.

Despite disparate morphologies, ecologies, and phylogenetic positions, a similarity between the overall endocranial morphologies of Shuvosaurus inexpectatus and extant crocodilians is exactly what is exhibited (Fig. 3.2). This discrepancy strongly suggests that the endocranial morphologies of the Crurotarsi are highly conserved, exhibiting only relatively minor changes attributable to variations in the planar organization of major elements of the skull or to the expansion or reduction in size of features associated with evolutionary trends from the basalmost to crown group condition. The most notable and readily apparent contrast between the endocrania of

Smilosuchus adamanensis and other members of Crurotarsi is the presence of an enlarged, dorsally positioned dural expansion that indicates the position of the pineal body. Whereas this dural expansion is quite large in the digitally reconstructed endocranium of S. adamanensis , it is entirely absent in the reconstructions of both extant crocodilians and Simosuchus clarki (Fig. 3.2). The endocranium reconstruction of

Shuvosaurus inexpectatus includes a feature that is consistent with the position of this dural expansion but of a size greatly reduced from that in Smilosuchus adamanensis. A figure included with the description of an aetosaur braincase by Sulej (2010) shows a projection of the endocranial cavity consistent with this type of dural expansion, although this feature is not specifically noted or identified in the description and a reconstruction

44 of the endocast was not included. Moreover, the morphology of the interior of the braincase of this aetosaur, Stagonolepis olenkae , suggests an overall endocranial morphology consistent with those seen in other Crurotarsi. The relative size of the potential dural expansion in Stagonolepis olenkae is greater than that seen in Shuvosaurus inexpectatus and considerably less than what is present in Smilosuchus adamanensis .

This pattern suggests that endocranial morphologies are highly conserved throughout the

Crurotarsi, apparently regardless of differences in overall body morphology or ecology, but that the pineal body, which is at its greatest relative size in S. adamanensis , is greatly reduced in size only a short phylogenetic distance closer to the crown group before becoming so reduced that it no longer appears in the cranial endocasts of taxa near the crown group.

In order to determine whether the conservatism seen in the cranial endomorphologies of the Crurotarsi is a unique pattern or one typical in Archosauria, the endocranial lineages of additional archosaur clades were added to the comparative data.

Specifically, series of endocasts presented for theropod dinosaurs in Witmer & Ridgely

(2009) and sauropod dinosaurs in Sereno et al. (2007) were aligned with a cladogram for each of those groups (Fig. 3.3; Fig. 3.4). The subsequent plots of similar and disparate features for these two dinosaur groups were then compared to the Crurotarsi endocast series presented here (Fig. 3.2). The first result of this comparison is that the endocasts of theropods and sauropods are generally discernable from one another, consistent with analysis by other studies (e.g. Witmer et al., 2008), despite the two clades being phylogenetically closer to each other than are phytosaurs and extant crocodilians. Both dinosaur groups are also discernable from the Crurotarsi. Furthermore, although there are

45

Figure 3.3. A series of endocasts of the Theropoda and the phylogenetic relationships of the taxa they represent. A comparison of the endocasts of: Majungasaurus crenatissimus (A), Allosaurus fragilis (B), Tyrannosaurus rex (C), Struthiomimus altus (D), Deinonychus antirrhopus (E), and Archaeopteryx lithographica (F), all modified from Witmer & Ridgely (2009). Red arrows indicate the peak of the dural expansion, when present, indicating the position of the pineal body. Cladogram modified from Holtz et al. (2004).

general trends seen in the endocranial morphologies of theropods and sauropods that make them discernable from one another, the differences seen within each of these clades are greater than the differences seen between the Crurotarsi. Of the endocasts represented for both theropods and sauropods, the basalmost members of each clade are the most morphologically similar to each other and, indeed, to the basalmost Crurotarsi.

This similarity suggests that the Crurotarsi-style cranial endocast is the most similar to

46 Figure 3.4. A series of endocasts of the Sauropodomorpha and the phylogenetic relationships of the taxa they represent. A comparison of the endocasts of: Massospondylus carinatus (A) from Sereno et al. (2007), Camarasaurus lentus (B) from Witmer et al. (2008), Diplodocus longus (C) from Witmer et al. (2008), and Nigersaurus taqueti (D) from Sereno et al. (2007), all from Sereno et al. (2007). Red arrows indicate the peak of the dural expansion, when present, indicating the position of the pineal body. Cladogram mdofied from: Sauropodomorpha to Sauropoda, Pol et al. (2010); Sauropoda through Macronaria, Wilson (2002); and Diplodocoidea, Sereno et al. (2007).

the primitive condition, and while various dinosaur clades exhibit derived endocranial morphologies, interpreted to have been related to derived and differentiated ecologies

(e.g. Witmer et al. 2008; Witmer & Ridgely, 2009), the Crurotarsi maintain a highly conservative and largely primitive endocranial morphology throughout derived and differentiated ecologies. Additionally, the trend of relative pineal body size decrease

47 from the large, primitive condition to the greatly reduced crown group condition is seen in both the theropod and sauropod lines, just as it is through the Crurotarsi.

The functional implications of this trend of pineal body reduction are difficult to ascertain. It has been noted that the pineal gland produces melatonin, which has sleep- inducing and temperature-lowering effects, in at least some animals (Macchi & Bruce,

2004; Arendt & Skene, 2005), related to its function of regulating circadian rhythm and seasonal cycles (Hopson, 1979). In rats, exposure to constant light has been shown to decrease pineal weight, and both a pinealectomy and exposure to constant light have the same effect of increasing and accelerating gonad growth (e.g. Aleandri et al., 1996). In at least some birds, the pineal gland affects gonad production activity associated with seasonal breeding (Ramachandran et al., 1996). The pineal gland has also been linked to migration patterns in newts, and similarly, metabolism and hibernation functions in at least some animals (Hopson, 1979). However, the functional benefits of pineal reduction are not readily apparent and are not agreed upon (Hopson, 1979). Furthermore, the exact function(s) of a primitively large pineal gland in phytosaurs may be difficult to determine as it seems unlikely that at least some of the noted pineal functions, such as migration and hibernation, are applicable for a semiaquatic predator. One reasonable hypothesis for the large phytosaur pineal body is that it is simply a transitional step between the relatively large pineal body of more basal diapsid reptiles like lizards and the extremely reduced pineal body of crown group crocodilians. Exposed at the skull surface in the form of a photoreceptive patch, the pineal gland, or pineal eye, of lizards serves as a receptor of light/dark sensory information (e.g. Hopson, 1979). As the aperture through which the pineal body was exposed to direct external light was evolutionarily lost, the relative size

48 of the pineal body seems to have been reduced because it was still able to function adequately at a reduced size and no longer needed to be large enough to contact the skull surface, saving the metabolic expense of developing and maintaining the gland at a large size. However, it may be the case that each of the three compared phylogenetic lineages exhibits the same pineal reduction trend for the same reason, entirely different and unique reasons, or any number or combinations of unique and shared reasons.

Future Work

The comparative discussion presented here highlights the need for a large body of data for comparative analysis and from which to draw substantial conclusions. At this time, the limited availability of digitally reconstructed endocasts for archosaurs, in general, and Crurotarsi, in particular, serves as an obstacle to making large-scale statements about evolutionary trends in endocast morphology. For that reason, this study is the first in a series of planned studies to digitally reconstruct, describe, and compare the virtual endocasts of representatives of numerous clades within Crurotarsi. As the basalmost members of Crurotarsi, phytosaurs are a natural starting point for a large-scale study of this type. The convergent evolution of phytosaurs and extant crocodilians also makes a comparative study of the endocasts of these two groups a reasonable venture and an excellent baseline for comparisons between other members of Crurotarsi. Additional clades of particular interest include aetosaurs, poposaurs and thalattosuchians, not only for the uniqueness of their morphologies and ecologies within Crurotarsi but also because of their overall convergence in these areas with ankylosaur and ornithomimid dinosaurs and ichtyosaurs, respectively. Of course, such a proposal means that enough other

49 archosaurs, like the aforementioned dinosaur groups, will need to undergo the same type of study in order for convergent-based comparisons to be possible. Also, dinosaur clades from which the endocast of a single taxon has been described will need to be sampled throughout their lineages in order to further investigate the trends of derivation from the primitive endocast form and reduction in relative pineal body size discussed here. The true nature of the primitive endocast form and pineal body size will also be tested by sampling Euparkeria and Erythrosuchus , as close outgroup members to Archosauria. All of these proposals create a picture of a very long-term, in-depth study that will require the large amount of time and resources necessary for such a survey.

Conclusions

A comparison of the endocasts of the phytosaur Smilosuchus adamanensis and an extant crocodilian supports the hypothesis that the two share a similar overall morphology. However, this support comes with a major caveat in that the rationale put forth here for a shared overall morphology does not seem to be supported. In other archosaurs, for instance pterosaurs and early birds, a shared overall endocast morphology is viewed as convergence due to a similar ecology of the two groups (Domínguez Alonso et al., 2004; Witmer et al., 2003) whereas in the case of phytosaurs and extant crocodilians, the similarity seems to be more of a case of extreme conservatism rather than convergence. This conclusion is born out by the addition of data from taxa with phylogenetic positions between, and body morphologies and ecologies that diverge from, those of phytosaurs and extant crocodilians, yet exhibit similar endocast morphologies.

This conservatism suggests a very different neurological adaptive strategy from the one

50 seen throughout the various clades of dinosaurs. Although dinosaurs exhibit a variety of

endocast morphologies, sometimes showing a great deal of variation between closely

related taxa, the Crurotarsi seem to diverge into a wide range of overall body

morphologies and ecologies without requiring a similar amount of deviation from the

plesiomorphic endocast morphology.

The use of CT data to digitally reconstruct and analyze the internal cranial

anatomy of fossil taxa is a relatively new method of study of study, having become a

regular part of the literature within the last decade or so. Still, only a handful of

paleontology labs are equipped with the tools and knowledge necessary for digitally

reconstructing endocasts, and only a small number of descriptions of new taxa within the

last decade have included such endocast reconstructions. Previously described taxa have

similarly been rare beneficiaries of digital endocast reconstructions. Although the vast

majority of archosaurian genera have yet to be treated in this way, most of the major

clades of dinosaurs, including theropods (e.g. Domínguez Alonso et al., 2004; Franzosa

& Rowe, 2005; Witmer et al., 2008a; and Witmer & Ridgely, 2009; Witmer et al., 2010),

sauropods (Sereno et al., 2007), ankylosaurs (Witmer & Ridgely, 2008a), ceratopsians

(Witmer & Ridgely, 2008b), and lambeosaurs (Evans et al., 2009), have already been

sampled, and some, like theropods and sauropods, benefit from a series of endocasts

throughout their phylogenetic lineages. The Crurotarsi are therefore the largest, and most

diverse, group of archosaurs to have the greatest potential for exploration of this type.

Indeed, excluding extant taxa, only one other specimen, Simosuchus clarki (Kley et al.,

2010), in one other clade, Notosuchia, of Crurotarsi has been the subject of a publication focusing on this topic. Although virtual endocasts of two other notosuchians,

51 Anatosuchus minor and Araripesuchus wegeneri (Sereno & Larsson, 2009), have been published, this was not the focus of the study, and the detail of the virtual endocasts is not comparable to those listed above.

Comparisons between endocasts of the Crurotarsi, and indeed between those of all archosaurs, are dependent upon a sample size large enough to form the basis of clear and valid conclusions. At this time, the sample size of crurotarsan endocasts is so small that any comparative analysis is preliminary in nature. Given the wide temporal (Early

Triassic to modern) and geographic (global) distribution of members of this group, a great range of studies could result from a larger number of high-quality virtual endocasts.

The Crurotarsi exhibit an incredible amount of morphotypic and ecological diversity, including semi-aquatic predators like phytosaurs and extant crocodilians; heavily armor- plated, terrestrial herbivores like aetosaurs; large, pillar-erect, terrestrial predators like rauisuchians; bipedal, toothless-beaked herbivores like poposaurs; small, long-limbed and gracile sphenosuchians; fully marine thalattosuchians; and very diverse and highly derived herbivorous notosuchians. In order to fully appreciate the similarities due to phylogenetic relationships and differences due to divergences in overall morphologies and ecologies of the endocasts currently available for comparison, the virtual endocasts of more of these very diverse forms must be made available.

52 APPENDIX 1. List of anatomical abbreviations.

asc , anterior semicircular canal (or endocast thereof) cer , cerebrum CN II , optic nerve CN V , trigeminal nerve CN IX + X + XI , glossopharyngeal, vagus, and accessory nerves dls , dorsal longitudinal dural venous sinus fl , flocculus ijv , interior jugular vein lsc , lateral semicircular canal (or endocast thereof) ob , olfactory bulb ocs , occipital dural venous sinus ot , olfactory tract pin , pineal body psc , posterior semicircular canal (or endocast thereof) sps , sphenoparietal dural venous sinus ts , transverse dural venous sinus vls , ventral longitudinal dural venous sinus

53 A Virtual Phytosaur Endocast and Its Implications for Sensory System Evolution in Archosaurs

References:

Abel, O. 1922. Die Schnauzenverletzungen der Parasuchier und ihre biologische Bedeutung. Palaontologische Zeitschrift : 26-57.

Ballew, K.L. 1989. A phylogenetic analysis of Phytosauria from the Late Triassic of the western United States. In: Lucas, S.G. & Hunt, A.P. (eds.). Dawn of the Age of Dinosaurs in the American Southwest. New Mexico Museum of Natural History, Albuquerque, NM: p. 309-339.

Benton, M.J. 1999. Scleromochlus taylori and the origin of dinosaurs and pterosaurs. Philosophical Transactions of the Royal Society of London 354 : 1423-1446.

Brusatte, S.L.; Benton, M.J.; Desojo, J.B.; Langer, M.C. 2010. The higher-level phylogeny of Archosauria (Tetrapoda: Diapsida). Journal of Systematic Palaeontology 8 (1): 3–47.

Buffetaut, E. & Ingavat, R. 1982. Phytosaur remains (Reptilia, Thecodontia) from the Upper Triassic of north-eastern Thailand. Geobios 15 (1): 7-17.

Buffetaut, E.; Martin, M.; Monod, O. 1988. Phytosaur remains from the Cenger Formation of the Lycian Taurus (Western Turkey): Stratigraphical implications. Geobios 21 : 237-243.

Burmeister, K.C. 2000. Paleogeographic and biostratigraphic implications of new early Mesozoic terrestrial vertebrate fossils from the Poamay site; central Morondava Basin, Madagascar. Unpublished Masters thesis, University of California, Santa Barbara: 109 p.

Camp, C.L. 1930. A study of the phytosaurs with description of new material from western North America. Memoirs of the University of California 10 : 1-174.

Carlson, W.D.; Rowe, T.; Ketcham, R.A.; Colbert, M.W. 2003. Geological applications of high resolution x-ray computed tomography in petrology, meteoritics and palaeontology. In: Mees, F.; Swennen, R.; Van Geet, M.; Jacobs, P. (eds). Applications of X-ray Computed Tomography in the Geosciences. Geological Society, London, UK: p. 7-22. Case, E.C. 1929. Description of the skull of a new form of phytosaur with notes on the characters of described North American phytosaurs. Michigan State University Paleontological Memoirs 2: 1-56.

54 Chatterjee, S. 1978. A primitive parasuchid (phytosaur) reptile from the Upper Triassic Maleri Formation of India, Palaeontology 21 : 83-127.

Colbert, E.H. 1947. Studies of the phytosaurs Machaeroprosopus and Rutiodon [Arizona, North Carolina]. Bulletin of the American Museum of Natural History 88 (2): 53- 96. Conroy, G.C.; Vannier, M.W. 1984. Noninvasive three-dimensional computer imaging of matrix filled fossil skulls by high-resolution computed tomography. Science 226 : 456-458.

Cruickshank, R.I. & Benton, M.J. 1985. Archosaur ankles and the relationships of the thecodontian and dinosaurian reptiles. Nature 317 : 715-717.

Domínguez Alonso, P.; Milner, A.C.; Ketcham, R.A.; Cookson, M. J.; Rowe, T.B. 2004. The avian nature of the brain and inner ear of Archaeopteryx . Nature 430 : 666- 669.

Doyle, K.D. & Sues, H.-D. 1995. Phytosaurs (Reptilia: Archosauria) from the upper Triassic New Oxford formation of York county, Pennsylvania. Journal of Vertebrate Paleontology 15 (3): 545-553.

Dzik, J. & Sulej, T. 2007. A review of the early Late Triassic Krasiejów biota from Silesia, Poland. Palaeontologia Polonica 64 : 3-27.

Easton, J. New generation of multi-channel CT scanners brings diagnostic power and speed. The University of Chicago Medical Center. April 1, 2005. Web. November 26, 2010.

Eaton, T.H, Jr. 1965. A new Wyoming phytosaur. The University of Kansas Paleontological Contributions 2: 1-6

Edinger, T. 1975. Paleoneurology 1804 –1966: An annotated bibliography. Advances in Anatomy, Embryology, and Cell Biology 49 :1-258.

Erichsen, J.T.; Hodos, W.; Evinger, C.; Bessette, B.B.; Phillips, S.J. 1989. Head orientation in pigeons; postural, locomotor, and visual determinants. Brain, Behavior, and Evolution 33 : 268-278.

Evans, D.; Witmer, L.M.; Ridgely, R.C. 2009. Endocranial anatomy of lambeosaurine dinosaurs: a sensorineural perspective on cranial crest function. The Anatomical Record 292:1315–1337. Fara, E., and Hungerbühler, A. 2000. Paleorhinus magnoculus from the Upper Triassic of Morocco: a juvenile primitive phytosaur (Archosauria). Comptes Rendus de l’Académie des Sciences, Paris, Sciences de la Terre et des planétes 331 : 831- 836.

55 Franzosa, J. & Rowe, T. 2005. Cranial endocast of the Cretaceous theropod dinosaur Acrocanthosaurus atokensis . Journal of Vertebrate Paleontology 25 (4): 859-864.

Gauthier, J.A. 1986. Saurischian monophyly and the origin of birds. In: Padian, K. (ed.). The Origin of Birds and the Evolution of Flight. Memoirs of the California Academy of Sciences 8. California Academy of Sciences, San Francisco: 1-55.

Gauthier, J. & Padian, K. 1985. Phylogenetic, functional, and aerodynamic analyses of the origin of birds and their flight. In: Hecht, M.K.; Ostrom, J.H.; Viohl, G. and Wellnhofer, P. (eds.). The Beginnings of Birds: proceedings of the international Archaeopteryx conference. 1984. Eichstätt: Freunde des Jura-Museums, Eichstätt, Germany: p. 185-197.

Gozzi, E. & Renesto, S.A. 2003. Complete specimen of Mystriosuchus (Reptilia, Phytosauria) from the Norian (Late Triassic) of Lombardy (Northern Italy). Rivista Italiana Di Paleontologia e Stratigrafia 109 (3): 475-498.

Graf, W.; de Waele, C.; Vidal, P.P. 1995. Functional anatomy of the head-neck movement system of quadropedal and bipedal mammals. Journal of Anatomy 186 : 55-74.

Gregory, J.T. 1962. The genera of the phytosaurs. American Journal of Science 260 : 652–690.

Gregory, J.T. 1969. Evolution und interkontinentale Beziehungen der Phytosauria (Reptilia). Paläontologische Zeitschrift 43 : 37-51.

Heckert, A.B. & Lucas, S.G. 2002. Revised Upper Triassic stratigraphy of the Petrified Forest National Park. In: Heckert, A.B. & Lucas, S.G. (eds.). Upper Triassic Stratigraphy and Paleontology. Bulletin of the New Mexico Museum of Natural History and Science 21 , Albuquerque, NM: 1-36.

Herman, G.T. 2010. Fundamentals of Computerized Tomography: Image Reconstruction From Projection, 2nd ed. Springer, New York, NY: 300 p.

Hofer, H.O. & Wilson, J.A. 1967. An endocranial cast of an early Oligocene primate. Folia Primatologica 5: 148-152.

Holliday, C.M.; Ridgely, R.C.; Balanoff, A.M.; Witmer, L.M. 2006. Cephalic vascular anatomy in flamingos ( Phoenicopterus ruber ) based on novel vascular injection and computed tomographic imaging analyses. The Anatomical Record; Part A 288A : 1031-1041.

Holtz, T.R., Jr.; Molnar, R.E.; Currie, P.J. 2004. Basal Tetanurae. In: Weishampel, D.B.; Dodson, P.; Osmólska, H. (eds.). The Dinosauria, Second Edition. University of California Press, Berkeley, CA: p. 71-110.

56 Hopson, J.A. 1979. Paleoneurology. In: Gans, C.; Northcutt, R.G.; Ulinki, P.S. (eds.). Biology of the Reptilia. Neurology A , Vol. 9. Academic Press, New York, NY: p. 39-146.

Hullar, T.E. 2006. Semicircular canal geometry, afferent sensitivity, and animal behavior. The Anatomical Record 288 : 466-472.

Hungerbühler, A. & Hunt, A.P. 2000. Two new phytosaur species (Archosauria, Crurotarsi) from the Upper Triassic of Southwest Germany. Neues Jahrbuch Fuer Geologie Und Palaontologie-Monatshefte 2000 (8): 467-484.

Hungerbühler, A. 2002. The Late Triassic phytosaur Mystriosuchus Westphali, with a revision of the genus. Palaeontology 45 (2): 377-418.

Hunt, A.P. 1989. Cranial morphology and ecology among phytosaurs. In: Lucas, S.G. & Hunt, A.P. (eds.). Dawn of the Age of Dinosaurs in the American Southwest. New Mexico Museum of Natural History, Albuquerque, NM: p. 349-354.

Hunt, A.P. & Lucas, S.G. 1991. The Paleorhinus biochron and the correlation of the non- marine Upper Triassic of Pangaea. Palaeontology 34 (2): 487-501.

Hunt, A.P. & Lucas, S.G. 1993. A new phytosaur (Reptilia: Archosauria) genus from the uppermost Triassic of the western United States and its biochronological significance. Bulletin of the New Mexico Museum of Natural History and Science 3:193-196.

Irmis, R.B. 2005. The vertebrate fauna of the Upper Triassic Chinle Formation in northern Arizona. In: Nesbitt, S.J.; Parker, W.G.; and Irmis, R.B. (eds.). Guidebook to the Triassic formations of the Colorado Plateau in northern Arizona: Geology, Paleontology, and History. Mesa Southwest Museum Bulletin , 9: 63-88.

Jaeger, G.F. 1828. Über die fossilen Reptilien, welche in Würtemberg aufgefunden worden sind. Metzler, Stuttgart, 48 pp., 6 pls.

Jerison, H.J. 1973. Evolution of the brain and intelligence. New York: Academic Press.

Kischlat, E.-E. & Lucas, S.G. 2003. A phytosaur from the Upper Triassic of Brazil. Journal of Vertebrate Paleontology , 23:464-467.

Kley, N.J.; Sertich, J.J.W.; Turner, A.H.; Krause, D.W.; O'Connor, P.M.; Georgi, J.A. 2010. Craniofacial morphology of Simosuchus clarki (Crocodyliformes: Notosuchia) from the Late Cretaceous of Madagascar. Journal of Vertebrate Paleontology 30 (6, Supplement): 13–98.

57 Larsson, H.C. E. & Sues, H.-D. 2007. Cranial osteology and phylogenetic relationships of Hamadasuchus rebouli (Crocodyliformes: Mesoeucrocodylia) from the Cretaceous of Morocco. Zoological Journal of the Linnean Society 149 : 533-567.

Lauder, G.V. 2005. On the inference of function from structure. In: Thomason, J.J. (ed.). Functional Morphology in Vertebrate Paleontology. Cambridge University Press, Cambridge, UK: p. 1-18.

Laurin, M. & Gauthier, J.A. 1996. "Amniota: mammals, reptiles (turtles, lizards, Sphenodon, crocodiles, birds) and their extinct relatives." Version 1, January, 1996. http://tolweb.org/Amniota/14990/1996.01.01 in The Tree of Life Web Project, http://tolweb.org/

Lehane, J. 2005. Anatomy and relationships of Shuvosaurus , a basal theropod from the Triassic of Texas. M.S. thesis, Texas Tech University, 92 p.

Lehman, T. & Chatterjee, S. 2005. The depositional setting and vertebrate biostratigraphy of the Triassic Dockum Group of Texas. Indian Journal of Earth System Sciences 114 (3): 325-351.

Long, R.A. and Murry, P.A. 1995. Late Triassic (Carnian and Norian) tetrapods from the southwestern United States. Bulletin of the New Mexico Museum of Natural History and Science 4: 1–254.

Lucas, S.G. 1998. Global Triassic tetrapod biostratigraphy and biochronology. Palaeogeography, Palaeoclimatology, Palaeoecology 143 : 347-384.

Lucas, S.G.: Anderson, O.J.; Hunt, A.P. 1994. Triassic stratigraphy and correlations, southern High Plains of New Mexico-Texas. New Mexico Bureau of Mines and Mineral Resources Bulletin 150 : 105-126.

Lucas, S.G.; Heckert, A.B.; Rinehart, L. 2007. A giant skull, ontogenetic variation and taxonomic validity of the Late Triassic phytosaur Parasuchus . Bulletin of the New Mexico Museum of Natural History and Science 41 : 222-228.

Marsh, O.C. 1896. A new belodont reptile (Stegomus) from the Connecticut River Sandstone. American Journal of Science 4 (2): 59-62.

Martin, J.E. & Benton, M.J. 2008. Crown Clades in vertebrate nomenclature: correcting the definition of crocodylia. Systematic Biology 57 : 1, 173-181.

Meyer, H.v. 1861. Reptilien aus dem Stubensandstein des oberen Keupers. Palaeontographica 7: 253-346.

58 Milner, A.C. & Walsh, S.A. 2009. Avian brain evolution: new data from Paleogene birds(Lower Eocene) from England. Zoological Journal of the Linnean Society 155 : 198-219.

Nesbitt, S.J. 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.

Nesbitt, S.J.; and Norrell, M.A. 2006. Extreme convergence in the body plans of an early suchian (Archosauria) and ornithomimid dinosaurs (Theropoda). Proceedings of the Royal Society B 273 (1590): 1045–1048.

Olsen, P.E.; Kent, D.V.; Sues, H.-D.; Koeberl, C.; Huber, H.; Montanari, A.; Rainforth, E.C.; Fowell, S.J.; Szajna, M.J.; Hartline, B.W. 2002. Ascent of dinosaurs linked to iridium anomaly at the Triassic-Jurassic boundary. Science 296 : 1305-1307.

Parker, W.G. & Irmis, R.B. 2006. A new species of the Late Triassic phytosaur Pseudopalatus (Archosauria: Pseudosuchia) from the Petrified Forest National Park, Arizona. Bulletin of the Museum of Northern Arizona 62 : 126-143.

Parrish, J.M. 1993. Phylogeny of the Crocodylotarsi, with reference to archosaurian and crurotarsan monophyly. Journal of Vertebrate Paleontology 13 : 287-308.

Pol, D.; Garrido, A.; Cerda, I.A. 2011. A new sauropodomorph dinosaur from the Early Jurassic of Patagonia and the origin and evolution of the Sauropod-type sacrum. Public Library of Science ONE 6 (1): e14572.

Radinsky, L. 1968. A new approach to mammalian cranial analysis, illustrated by examples of prosimian primates. Journal of Morphology 124 : 167-180.

Rogers; S.W. 2005. Reconstructing the behaviors of extinct species: an excursion into comparative paleoneurology. American Journal of Medical Genetics 134A : 349- 356.

Sampson, S.D. & Witmer, L.M. 2007. Craniofacial anatomy of Majungasaurus crenatissimus (Theropoda: Abelisauridae) from the Late Cretaceous of Madagascar. Society of Vertebrate Paleontology Memoir 8 27 (2, Supplement): 32–102.

Sedlmayr, J.C. 2002. Anatomy, evolution, and functional significance of cephalic vasculature in Archosauria. Ph.D. dissertation, Ohio University, 398 p.

Senter, P. 2005. Phylogenetic taxonomy and the names of the major archosaurian (Reptilia) clades. Paleobios 25 (2): 1-7.

59 Sereno, P.C. 1990. "Thecodonts", "ornithosuchians" and ankle "types": A review of the evidence and an alternative archosaur phylogeny. Journal of Vertebrate Paleontology , Supplement 10 : 41-42A.

Sereno, P.C., and A.B. Arcucci. 1990. The monophyly of crurotarsal archosaurs and the origin of bird and crocodile ankle joints. Neues Jahrbuch für Geologie und Paläontologie Abhandlungen 180 : 21-52.

Sereno, P.C.; Sidor, C.A.; Larsson, H.C.E.; Gado, B. 2003. A new notosuchian from the Early Cretaceous of Niger. Journal of Vertebrate Paleontology 23 (2): 477-482.

Sereno, P.C.; Wilson, J.A.; Witmer, L.M.; Whitlock, J.A.; Maga, A.; Ide, O.; Rowe, T.A. 2007. Structural extremes in a Cretaceous dinosaur. Public Library of Science ONE 2: e1230.

Smith, D.; Sanders, R.; Parker, W.; Cavanaugh, J. 2010. The endocranium, inner ear, and pneumatic structure of the Upper Triassic phytosaur Pseudopalatus pristinus [abstract]. In: Program and Abstracts of the 70th Anniversary Meeting of the Society of Vertebrate Paleontology; 2010 Oct 10-13; Pittsburgh, PA, USA; SVP. p. 167A.

Sollas, W.J. 1904. A method for the investigation of fossils by serial sections. Philosophical Transactions of the Royal Society of London B 196 : 259-265.

Stocker, M.R. 2010. A new taxon of phytosaur (Archosauria: Pseudosuchia) from the Late Triassic (Norian) Sonsela Member (Chinle Formation) in Arizona, and a critical reevaluation of Leptosuchus , Case, 1922. Palaeontology 53 (5): 997-1022.

Tanner, L.H.; Lucas, S.G.; Chapman, M.G. 2004. Assessing the record and causes of Late Triassic extinctions. Earth-Science Reviews 65 : 103-139.

Taylor, M.P.; Wedel, M.J.; Naish, D. 2009. Head and neck posture in sauropod dinosaurs inferred from extant animals. Acta Palaeontologica Polonica 54 : 213-220.

Vidal, P.P.; Graf, W.; Berthoz, A. 1986. The orientation of the cervical vertebral column in unrestrained awake animals. I. Resting position. Experimental Brain Research 61 : 549-559.

Webb, G. & Manolis, C. 1989. Crocodiles of Australia. Reed, Australia: 160 p.

Wharton, D.S. 2000. An enlarged endocranial venous system in Steneosaurus pictaviensis (Crocoylia: Thalattosuchia) from the Upper Jurassic of Les Lourdines, France. Comptes Rendus de l’Académie des Sciences, Série IIA 331 : 221-226.

60 Wilson, J.A. 2002. Sauropod dinosaur phylogeny: critique and cladistic analysis. Zoological Journal of the Linnean Society 136 : 217-276.

Witmer, L.M. 2004. Inside the oldest bird brain. Nature 430 : 619-620.

Witmer, L.M. 2005. The Extant Phylogenetic Bracket and the importance of reconstructing soft tissues in fossils. In: Thomason, J.J. (ed.). Functional Morphology in Vertebrate Paleontology. Cambridge University Press, Cambridge, UK: p. 19-33.

Witmer, L. M. & Ridgely, R.C. 2008a. The Paranasal Air Sinuses of Predatory and Armored Dinosaurs (Archosauria: Theropoda and Ankylosauria) and Their Contribution to Cephalic Structure. The Anatomical Record 291 : 1362-1388.

Witmer, L.M. & Ridgely, R.C. 2008b. Structure of the brain cavity and inner ear of the centrosaurine ceratopsid Pachyrhinosaurus based on CT scanning and 3D visualization. In: Currie, P.D. (ed.). A New Horned Dinosaur From an Upper Cretaceous Bone Bed in Alberta. National Research Council Research Press, Ottawa, Canada: p. 117-144.

Witmer, L.M. & Ridgely, R.C. 2009. New insights into the brain, braincase, and ear region of tyrannosaurs, with implications for sensory organization and behavior. The Anatomical Record 292 : 1266-1296.

Witmer, L. M.; Chatterjee, S.; Franzosa, J.; Rowe, T. 2003. Neuroanatomy of flying reptiles and implications for flight, posture and behaviour. Nature 425 : 950–953.

Witmer, L.M.; Ridgely, R.C.; Dufeau, D.L.; Semones, M.C. 2008. Using CT to peer into the past: 3D visualization of the brain and ear regions of birds, crocodiles, and nonavian dinosaurs. In: Endo, H. & Frey, R. (eds.), Anatomical Imaging: Towards a New Morphology. Springer-Verlag, Tokyo, Japan: p. 67-88.

Witmer, L.M.; Ridgely, R.C.; James, H.; Olson, S.; Iwaniuk, A. 2010. Neuroanatomy, skull morphology, and behavioral implications for the remarkable, recently extinct "platypus-duck" Talpanas lippa (Aves: Anseriformes) from Kauai, Hawaii Journal of Vertebrate Paleontology . (abstract)

Woody, D.T. & Parker, W.G. 2004. Evidence for a transitional fauna within the Sonsela Member of the Chinle Formation, Petrified Forest National Park, Arizona. Journal of Vertebrate Paleontology 24 (3 Supplement):132A.

Zatón, M.; Piechota, A.; Sienkiewicz, E. 2005. Late Triassic charophytes around the bone bearing bed at Krasiejów (SW Poland): palaeoecological and environmental remarks. Acta Geologica Polonica 55 (3): 283-293.

61 Zeigler, K.E.; Lucas, S.G.; Heckert, A.B. 2002. The Late Triassic Canjilon quarry (Upper Chinle Group, New Mexico) phytosaur skulls: evidence of sexual dimorphism in phytosaurs. Bulletin of the New Mexico Museum of Natural History and Science 21 : 179-188.

Zeigler, K.E.; Lucas, S.G.; Heckert, A.B. 2003. Variation in the Late Triassic Canjilon quarry (Upper Chinle Group, New Mexico) phytosaur skulls: evidence of sexual dimorphism. Paläontologische Zeitschrift 77 : 341-351.

62