A geometric morphometric analysis of Crocodylus Niloticus: evidence for a cryptic species complex

Transcription

1 University of Iowa Iowa Research Online Theses and Dissertations Summer 2012 A geometric morphometric analysis of Crocodylus Niloticus: evidence for a cryptic species complex Jennifer Halin Nestler University of Iowa Copyright 2012 Jennifer Halin Nestler This thesis is available at Iowa Research Online: Recommended Citation Nestler, Jennifer Halin. "A geometric morphometric analysis of Crocodylus Niloticus: evidence for a cryptic species complex." MS (Master of Science) thesis, University of Iowa, Follow this and additional works at: Part of the Geology Commons

2 A GEOMETRIC MORPHOMETRIC ANALYSIS OF CROCODYLUS NILOTICUS: EVIDENCE FOR A CRYPTIC SPECIES COMPLEX by Jennifer Halin Nestler A thesis submitted in partial fulfillment of the requirements for the Master of Science degree in Geoscience in the Graduate College of The University of Iowa July 2012 Thesis Supervisor: Associate Professor Christopher A. Brochu

3 Copyright by JENNIFER HALIN NESTLER 2012 All Rights Reserved

4 Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL MASTER'S THESIS This is to certify that the Master's thesis of Jennifer Halin Nestler has been approved by the Examining Committee for the thesis requirement for the Master of Science degree in Geoscience at the July 2012 graduation. Thesis Committee: Christopher A. Brochu, Thesis Supervisor Ann F. Budd Llewellyn D. Densmore III

5 We know now what was unknown to all the preceding caravan of generations: that men are only fellow-voyagers with other creatures in the odyssey of evolution. This new knowledge should have given us, by this time, a sense of kinship with fellow-creatures; a wish to live and let live; a sense of wonder over the magnitude and duration of the biotic enterprise. Aldo Leopold A Sand County Almanac, and Sketches Here and There ii

6 ACKNOWLEDGMENTS Thanks to my advisor, C.A. Brochu, for his guidance and expertise in all things crocodylian, and committee members A.F. Budd and L.D. Densmore III. This project was funded by the Department of Geoscience at the University of Iowa and the International Union for the Conservation of Nature Crocodile Specialist Group. Specimen access was kindly provided by J.S. Steyer at the Museum national d Histoire naturelle, G. Lenglet at the Royal Belgian Institute for Natural Sciences, G. Cael at the Royal Museum for Central Africa, G. Köhler at the Senckenberg Museum, M. Rödel at the Museum für Naturkunde, and C. Mehling and D. Kizirian at the American Museum of Natural History. J.M. Patterson was instrumental and invaluable in his help with specimen photography, geographic information systems, and constructive feedback. E.W. Wilberg and the UI Paleo Group provided consistently helpful comments. Finally, thanks to Mom, Dad, Michelle, Tipper, Bumbledog, and Fen for their constant support even though I wouldn t go to law school. iii

7 ABSTRACT The Nile crocodile Crocodylus niloticus currently has an extensive range throughout the African continent and Madagascar, though fossil and subfossil remains show that its historic range was considerably larger and included parts of the Sahara Desert, Mediterranean coast, and Arabian Peninsula. Recent molecular studies have yielded genetically distinct populations of C. niloticus, leading to the possibility that C. niloticus is actually multiple cryptic species, while morphological variation remains unassessed. This study compares skulls of C. niloticus to other members of the genus Crocodylus in dorsal view using geometric morphometrics to evaluate intraspecific and interspecific variation. The morphometric analysis is coupled with a geographic analysis to determine if the species is morphologically variable by geographic region as well as a model-based cluster analysis to determine and morphological clusters irrespective of other factors. These analyses indicate that C. niloticus exhibits populational variation that exceeds almost every other species of Crocodylus, with differences between geographic regions statistically disctinct. These results support the presence of a cryptic species complex. Additionally, an osteological description of Crocodylus niloticus is provided. iv

8 TABLE OF CONTENTS LIST OF TABLES... vii LIST OF FIGURES... viii INTRODUCTION...1 MATERIALS AND METHODS...6 Geometric Morphometrics...6 Landmarks...6 Specimens...8 Specimen Photography and Image Processing...9 Analyses...10 RESULTS AND CONCLUSIONS...12 Interspecific Morphometric Analysis of Crocodylus and C. niloticus sensu lato...12 Intraspecific Morphometric Analysis of Crocodylus niloticus sensu lato...14 Intraspecific Analysis of Crocodylus niloticus sensu lato Grouped by River Basin...15 Intraspecific Analysis of Crocodylus niloticus sensu lato Grouped by Morphological Cluster...15 DISCUSSION...17 CRANIAL OSTEOLOGY OF CROCODYLUS NILOTICUS...20 Skull...20 Premaxilla...20 Nasal...21 Maxilla...21 Lacrimal...21 Prefrontal...22 Frontal...22 Postorbital...22 Parietal...23 Squamosal...23 Laterosphenoid...23 Basisphenoid...24 Jugal...24 Quadratojugal...24 Quadrate...24 Palatine...25 Ectopterygoid...25 Pterygoid...25 Exoccipital...26 Supraoccipital...26 Basioccipital...26 v

9 Mandible...26 Dentary...26 Angular...27 Articular...27 Surangular...27 Splenial...28 Coronoid...28 Foramina and fenestrae...28 Foramen for first dentary tooth...28 Incisive foramen...28 External naris...29 Orbit...29 Supratemporal fenestra...29 Infratemporal fenestra...29 Palatal fenestra...30 Secondary choana...30 Posttemporal fenestra...30 Cranio-quadrate passage...30 Foramen aerium...30 Foramina for cranial nerve V (vagus foramina)...31 Foramen for cranial nerve X (carotid foramen)...31 Foramina for cranial nerve XII...31 Foramen magnum...31 Mandibular foramina and fenestra...31 Additional cranial features...32 Dentition...32 Ossified secondary palate...32 Notes on pathology of FMNH Note on specimen maturity...33 WORKS CITED...34 APPENDIX A. TABLES...38 APPENDIX B. FIGURES...46 vi

10 LIST OF TABLES Table A1. List of non-crocodylus niloticus specimens A2. List of Crocodylus niloticus specimens A3. Variance explained by relative warps of the interspecific analysis of Crocodylus A4. p values of the one-way non-parametric multivariate analysis of variance of the interspecific analysis of Crocodylus A5. Variance explained by relative warps of the intraspecific analysis of Crocodylus niloticus sensu lato A6. p values of the one-way non-parametric multivariate analysis of variance of the intraspecific analysis of Crocodylus niloticus s.l. separated into groups by river basin A7. p values of the one-way non-parametric multivariate analysis of variance of the intraspecific analysis of Crocodylus niloticus s.l. separated into groups using a model-based cluster analysis vii

11 LIST OF FIGURES Figure B1. Landmarks used in the morphometric analysis...46 B2. Map of large-scale river basins on the African continent B3. Relative warps 1 and 2 of the interspecific analysis of Crocodylus B4. Relative warps 1 and 3 of the interspecific analysis of Crocodylus B5. Relative warps 1 and 4 of the interspecific analysis of Crocodylus B6. Relative warps 1 and 5 of the interspecific analysis of Crocodylus B7. Thin-plate spline of the consensus configuration of the interspecific analysis B8. Thin-plate splines of the first relative warp of the interspecific analysis showing a) minimum relative warp values and b) maximum relative warp values B9. Thin-plate splines of the second relative warp of the interspecific analysis showing a) minimum relative warp values and b) minimum relative warp values B10. Thin-plate splines of the third relative warp of the interspecific analysis showing a) maximum relative warp values and b) minimum relative warp values B11. Thin-plate splines of the fourth relative warp of the interspecific analysis showing a) maximum relative warp values and b) minimum relative warp values B12. Thin-plate splines of the fifth relative warp of the interspecific analysis showing a) maximum relative warp values and b) minimum relative warp values B13. Relative warps 1 and 2 of the intraspecific analysis of Crocodylus niloticus sensu lato classified by geographic region B14. Relative warps 1 and 3 of the intraspecific analysis of Crocodylus niloticus s.l. classified by geographic region B15. Relative warps 1 and 4 of the intraspecific analysis of Crocodylus niloticus s.l. classified by geographic region B16. Thin-plate spline of the consensus configuration of the intraspecific analysis of Crocodylus niloticus s.l viii

12 B17. Thin-plate splines of the first relative warp of the intraspecific analysis of Crocodylus niloticus showing a) minimum relative warp values and b) maximum relative warp values B18. Thin-plate splines of the second relative warp of the intraspecific analysis of Crocodylus niloticus showing a) minimum relative warp values and b) minimum relative warp values B19.. Thin-plate splines of the third relative warp of the intraspecific analysis of Crocodylus niloticus showing a) maximum relative warp values and b) minimum relative warp values B20. Thin-plate splines of the fourth relative warp of the intraspecific analysis of Crocodylus niloticus showing a) maximum relative warp values and b) minimum relative warp values B21. Results of the cluster analysis of Crocodylus niloticus s.l B22. Crocodylus niloticus (FMNH 17157) skull shown in dorsal view B23. Crocodylus niloticus (FMNH 17157) skull shown in palatal view B24. Crocodylus niloticus (FMNH 17157) skull shown in posterior view B25. Crocodylus niloticus (FMNH 17157) mandible shown in occlusal view ix

13 1 INTRODUCTION Crocodylus niloticus Laurenti, 1768 currently has an extensive range that spans most of the African continent and Madagascar, although fossil and subfossil remains show that its historic range was considerably larger, and included parts of the Sahara, Mediterranean coast, and Arabian Peninsula. The total population is currently estimated to be between 250,000 and 500,000 individuals, but these estimates are largely data deficient (Ross, 1998). Populations of Crocodylus niloticus are relatively stable in the eastern portion of its range and more threatened in the west, with a significant risk of extirpation in some areas due to hunting and conflict with humans. Considerable variation in morphology has been anecdotally reported, and up to 15 subspecies were once proposed, though none are formally recognized (Heinz, 1953; Hekkala, 2011). Recent molecular studies have yielded genetically distinct populations and indicate that C. niloticus is actually a cryptic species complex (Schmitz, 2004; Hekkala et al., 2010, 2011; Oaks, 2011). Crocodylus niloticus was recently split into two species, C. niloticus sensu stricto and C. suchus, the latter of which had been previously synonymized with C. niloticus. Crocodylus niloticus sensu lato is among the largest living crocodylians, with reported lengths up to 6 meters, though the average size varies considerably by population (Ross, 1998; Fuchs, 2006). Males are highly territorial, which may impact dispersal abilities. Including C. niloticus s.l., there are four crocodylian species in three genera recognized in Africa today: the slender-snouted crocodile Mecistops cataphractus (formerly Crocodylus cataphractus), and the dwarf crocodiles Osteolaemus tetraspis and Osteolaemus osborni (Eaton et al., 2009; Hekkala et al., 2010). Of these, C. niloticus s.l. is the largest and most widely distributed. Recent genetic work on Osteolaemus and

14 2 Mecistops has resulted in significant taxonomic revision: Osteolaemus was found to be at least two species, and more likely three, and Mecistops was found to fall outside the genus Crocodylus as a sister taxon to Osteolaemus (Brochu, 2000, 2007; Schmitz et al., 2003; McAliley et al., 2006; Eaton et al., 2009). Previously, a number of subspecies were named based on single individuals from assorted populations, though these seem to be based on observed behavior, size, and variations in skin color and scale pattern (de Smet, 1999; Fuchs, 2006) rather than skeletal morphology, and are not generally recognized in current literature. Laurenti (1768) described C. niloticus niloticus from Egypt as well as C. n. africanus from East Africa. The type locality is in Tanzania, although the type specimen appears to be based in part on an unidentified lizard (Stejneger, 1933). Additional subspecies include C. n. chamses from the Congo River (Bory, 1824), C. n. cowiei from South Africa (Smith and Hewett, 1937), and C. n. madagascariensis from Madagascar (Grandidier, 1872). Geoffroy (1807) described C. n. pauciscutatus from Lake Rudolph in Kenya and C. n. suchus from the Niger River. Crocodylus niloticus s.l. is an economically important species, valued primarily for its skin (Hekkala et al., 2010). This has led to it being listed under Appendices 1 and 2 of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES; Ross, 1998). Over 80,000 skins were exported in 1993 alone (Hekkala et al., 2010). Much of the pressure due to the skin trade has been alleviated by farming and ranching practices, although these are far more prevalent in Eastern Africa than Western Africa. The species is also subject to hunting for meat and due to conflict with human activities (de Smet, 1999). Despite being listed as least concern by the International Union for Conservation of Nature and Natural Resources (IUCN), the species is known to be locally rare or extirpated in areas. A reintroduction was proposed for a national park in Algeria, where the species had been extirpated, and was soundly rejected by locals (de Smet, 1999). Threats to the species vary by region and it is likely that, should C.

15 3 niloticus s.l. actually represent a species complex, one or more of those species is possibly endangered. Additionally, although farmed crocodiles are often considered to be a genetic repository of sorts available to bolster wild populations should the need arise a recent study of farmed crocodiles in South Africa found a very low genetic diversity (Flint et al, 2000). While reintroduction is an important tool for conservation, these genetic analyses suggest that any reintroduction would need to be thoroughly studied to insure the genetic integrity of distinct isolated populations is preserved. Molecular studies on Crocodylus niloticus s.l. have revealed high levels of genetic variation. Schmitz et al. (2003) conducted a preliminary mitochondrial DNA analysis on 14 C. niloticus individuals, and included the freshwater crocodile (C. johnstoni) and Cuvier s dwarf caiman (Paleosuchus palpebrosus) as outgroups. Although the gene was chosen for the analysis, mitochondrial 12S, is relatively conservative, both maximumlikelihood and Bayesian analyses recovered two distinct groups: monophyletic West and Central African populations, and paraphyletic East African and Malagasy populations, which paired more closely with C. johnstoni. The study was prompted by the rediscovery of a population of dwarf C. niloticus that do not exceed a length of 2.3 meters from an arid region of Mauritania (Shine et al., 2001). Such relict populations appear to have been geographically isolated by the expansion and contraction of the Sahara during the Pliocene and Pleistocene and are likely the remnants of the last humid phase of the region (Shine et al., 2001). Further resolution regarding the interrelationships of the distinct populations was not possible with the limited sample size available (Schmitz et al., 2003). A second, larger study by Hekkala et al. (2010) analyzed DNA from 90 C. niloticus s.l. individuals from Madagascar, Kenya, Malawi, South Africa, Tanzania, Zimbabwe, and Western Sahara. They used multilocus genotyping and were able to successfully assign individuals to a population with a 94% accuracy rate, indicating

16 4 significant divergence of populations of the species, more so than studies characterizing the American alligator Alligator mississippiensis (Glenn et al., 1998; Davis et al., 2001) or Morelet s crocodile C. moreletii (Dever et al., 2002; Ray et al., 2004). It also found population differentiation to be roughly divided by drainage basins on both a regional and sub-regional scale. However, there were too few samples from Western Africa (6 out of 98 total) to thoroughly assess population structure (Hekkala et al., 2010). A more recent analysis by Hekkala et al. (2011) resurrected the species Crocodylus suchus after a DNA analysis indicated that C. niloticus s.l. was in fact a paraphyletic species complex, with one more closely aligned with Indopacific crocodylians and one with New World crocodylians. The analysis included nuclear and mitochondrial DNA from eight additional crocodylids as well as from eight mummified museum specimens. It is not known if morphological variation is also present to support a specieslevel division of C. niloticus s.l. Despite known genetic variation, described subspecies, and qualitative observation of morphological variation, a comprehensive, quantitative study of the morphology of C. niloticus s.l. has not been performed to date. This study tests the hypothesis that morphological variation within C. niloticus s.l. is higher than that of other crocodylid species. Additionally, both a model-based cluster analysis and a geographic analysis are used to identify any morphotypes or distribution to the morphological variation. Moreover, the techniques employed in the course of this research will have broad applications for paleontology, and in particular for integrating paleontology and conservation. Paleontology has the ability to inform and improve our decisions about conservation, but it is currently extremely under-utilized. This allows us to assess the diversity of a species or species complex across time and space and potentially elucidate its reaction to ecological change over time.

17 5 Nevertheless, it is only recently that quantitative comparisons have been applied to crocodylian morphology (Brochu, 2001; Sadleir and Makovicky, 2008; Piras et al., 2009, 2010). One method involves geometric morphometrics, which applies various multivariate statistical analyses of coordinate data that correspond to anatomical shape space (Zelditch et al., 2004). Because the crocodylian skull is dorsoventrally flattened, a significant number of anatomical characters can be assessed in two dimensions with a minimal amount of distortion, making crocodylians an ideal organism for a twodimensional geometric morphometric analysis (Pierce et al., 2008).

18 6 MATERIALS AND METHODS Geometric Morphometrics I used two-dimensional geometric morphometrics to assess morphological variation in crocodylids. Modern crocodylian diversity is fairly conservative, especially in the context of the diverse fossil record. Skull morphology overlaps significantly in morphospace. The purpose of this study is not to distinguish species using morphometrics alone, but to use it as a tool to assess the breadth of inter- and intraspecific variation. Procrustes superimposition in particular is a widely implemented technique that scales, rotates, and translates objects to remove all information that is not related to differences in shape (Walker, 2000). This is performed by overlaying homologous landmarks and minimizing Procrustes distances between them (Rohlf and Slice, 1990; Goodall, 1991). Landmarks The landmark set consists of 30 landmarks coded unilaterally on the dorsal surface of the skull (Figure B1). Landmarks were digitized from photographs using ImageJ (Schneider et al., 2012). For this study, a novel set of landmarks was selected based largely on prior work by Sadleir (2009), whose approach includes landmarks along the length of the maxilla to quantify change in snout shape and eliminate the possibility of a Pinocchio effect, which occurs when variance loadings are artificially high on small number of landmarks due to a disproportionately low number of landmarks in an anatomically variable region such as

19 7 the snout (Walker, 2000; Sadleir, 2009). This can result in erroneous classifications of skull type. To prevent this, landmarks along the dorsal margin of the snout are coded by superimposing the image on the skull in dorsal view and placing a landmark at every other alveolus in the jaw. To insure that the superimposition was correct I aligned landmarks 1 and 4 in dorsal view with their corresponding points in ventral view. The lateral margin of each alveolus was coded instead of the center of the alveolus or the tooth itself for a number of reasons. First, and perhaps most intuitively, the number of teeth remaining in an individual skull varies widely, adding a source of imprecision in coding. Second, remaining teeth are often glued into the skull, which is done for aesthetic purposes and does not necessarily represent the position and orientation of teeth in life. Another issue regarding the use of alveoli as landmarks is their number varies both within and among species, or even within an individual jaw (Langston, 1973). This raises the question of whether landmarks on teeth are in fact homologous. Maxillary teeth at least are assumed to be added or lost at the back of the mouth (Grenard, 1991). Additional study will need to address the matter of homology in crocodylian alveoli. In this study, the rearmost maxillary alveoli were excluded to avoid this issue. Not all points of contact between posterior bones of the skull were included so as not to over-represent relatively minute changes occurring along sutures. These landmarks quantify changes along the skull table as well as in snout shape and tooth position. The landmarks in this study are all either Type I or Type II (Zelditch, 2004). In dorsal view, landmarks lying on the midline include: (1) the anterior margin of the medial symphysis, (5) the suture of the nasals at the junction of the posterior limit of the naris, (15) the anterior point of the frontal at the junction with the nasals, (22) the midpoint on frontal-parietal suture, and (30) the midpoint of the posterior margin on the parietal.

20 8 Landmarks off the midline consist of: (2) the third premaxillary alveolus, (3) the fifth premaxillary alveolus, (4) the lateral limit of the premaxilla-maxilla suture, (6) the second maxillary alveolus, (7) the fourth maxillary alveolus, (8) the sixth maxillary alveolus, (9) the eighth maxillary alveolus, (10) the tenth maxillary alveolus, (11) the twelfth maxillary alveolus, (12) the junction of the maxilla, premaxilla, and nasal bones, (13) the junction of the lachrymal, maxilla, and nasal bones, (14) the junction of the prefrontal, lachrymal, and nasal bones, (16) the junction of the jugal, maxilla, and lachrymal bones, (17) the junction of the jugal-lachrymal suture at the margin of the orbit, (18) the limit of the prefrontal-lachrymal suture at the margin of the orbit, (19) the junction of the frontal-prefrontal suture at the margin of the orbit, (20) the junction of the frontal-postorbital suture at the margin of the orbit, (21) the junction of the postorbitaljugal suture at the orbit, (23) the junction of the postorbital-squamosal suture at the margin of the supratemporal fenestra, (24) the junction of the postorbital-squamosal suture at the lateral margin, (25) the anterior limit of the jugal-quadratojugal suture at the margin of the lateral temporal fenestra, (26) the posterior limit of the jugal-quadratojugal suture, (27) the posterior limit of the quadrate-quadratojugal suture, (28) the ventrolateral margin of the squamosal, (29) the anterior limit of the squamosal-parietal suture on the posterior margin of the supratemporal fenestra. Specimens Skeletonized specimens from museum collections were used. An individual was included if at least half of the skull containing the complete set of landmarks was preserved. Additionally, some level of locality information had to be present for Crocodylus niloticus s.l. individuals. Specimens of all species that were obviously immature or captive were excluded.

21 9 The complete dataset included 124 skulls from 12 species of modern, fossil, and subfossil crocodylids (Tables A1 and A2). These species are Crocodylus acutus (7 individuals), C. intermedius (5 individuals), C. johnstoni (2 individuals), C. megarhinus (1 individual), C. moreletii (4 individuals), C. niloticus (69 individuals), C. novaguinae (3 individuals), C. palustris (12 individuals), C. porosus (46 individuals), C. rhombifer (6 individuals including fossil and sufossil remains), C. siamensis (2 individuals), and Voay robustus (1 individual). Specimen Photography and Image Processing Specimens were photographed in dorsal and palatal view. A scale was placed level with the center of focus. In dorsal view, the skulls were oriented so the suture between the nasal bones was horizontal. The camera focus was centered approximately on the juncture of the nasal bones and the frontal, allowing the depth of field to encompass any vertical relief in the skull. In palatal view, skulls were oriented so the suture between the maxillae was horizontal, and the center of focus was placed along that suture. Images were photographed using either a Canon 5D Mark II or Canon 50D digital camera, which generate high-resolution digital images. The camera was mounted on a tripod with the ability to hold the camera horizontally. Smaller skulls were photographed with a Canon 60mm EF-S f/2.8 USM macro lens. Larger skulls were photographed using a Sigma 50mm f/2.8 EX DG macro lens. The largest skulls were photographed using a Canon 17-40mm f/4l USM lens with as long a focal length as possible (i.e. as close to 40mm as possible) to minimize distortion. In rare instances, a Canon f/ EF- S lens was used.

22 10 Images were saved as RAW files, which are a lossless image format that preserve more detail than JPG format and as such are ideal for specimen photography. Images are catalogued in Adobe Lightroom 3, which allows users to tag photos with keywords and save them as metadata that can be embedded with the image file. Keywords for this project include family, genus, and species; locality and temporal information; specimen number; and any other relevant information on the specimen label. These keywords make images easily searchable, facilitate organization, and keep photos from losing relevant data about specimens. For example, in the event that a file name is changed, keywords can be used to confirm the identity of a specimen. This greatly reduces the potential for the misidentification of a photo. Images were exported from Adobe Lightroom as JPG files and landmarks placed using Adobe Photoshop CS5. Images were cropped, rotated, resized, and converted to grayscale for compatibility with ImageJ. Landmarks were placed along the left half of the skull. In instances where the left half was damaged, the image was flipped along the anteroposterior axis. Landmarks were placed multiple times for each specimen to minimize human error. Analyses Analyses were run twice: once on a dataset that included all species and once on a dataset that included only Crocodylus niloticus s.l. A generalized least squares Procrustes superimposition was performed using tpsrelw (version 1.49; Rohlf, 2010). Both partial warps and relative warps were calculated in tpsrelw. Relative warps are linear combinations of partial warps, and are akin to a principal components analysis (Hammer and Harper, 2006). However, they take large-scale deformation into account more so than

23 11 a principal components analysis. Morphological changes were visualized using thin-plate splines generated in tpssplin. The two datasets were analyzed using multiple methods. The first analysis was based on geographic locality information of C. niloticus. Due to the limited availability of locality data, which in some instances were referenced to ambiguous or antiquated place names, large-scale hydrologic basins were the most specific locality information that could be ascribed consistently to individuals. These hydrologic basins were determined using a digital elevation model in ArcGIS. They correspond with large river systems, and were named herein to reflect the largest rivers they contain (Figure B2). This method was chosen because rivers likely represent the primary mode of dispersal for individuals. Additionally, the relative warps comprising >5% of the variance were Crocodylus niloticus s.l. was subject to a model-based cluster analysis to determine how many, if any, distinct morphotypes are present. The cluster analysis was run in R (version 2.9.2) using MCLUST (R Development Core Team, 2008; Fraley and Raftery, 2006). A permutational multivariate analysis of variance (MANOVA) was then run on the relative warps using PAST v.2.07 to determine the statistical significance between designated groups (Hammer et al., 2001). Groups comprising a single specimen each were removed. The MANOVA included relative warps comprising at least 5% of the variance each. Separate analyses included crocodylid species and geographical groups, C.niloticus s.l. geographical groups, and C. niloticus s.l. clusters.

24 12 RESULTS AND CONCLUSIONS Interspecific Morphometric Analysis of Crocodylus and C. niloticus sensu lato The relative warp analysis required 23 relative warps to explain 95% of the variance (Table A3). Five relative warps explained more than 5% of the variance each. The first relative warp explained 32.24% of the variance, the second explained 18.32% of the variance, the third relative warp explained 8.50%, the fourth relative warp explained 6.53%, and the fifth relative warp explained 5.07%. While individual species do cluster in morphospace, there is both a range of variation within species and a significant amount of overlap between species (Figures B3- B6). Given the relatively conservative morphology of modern crocodylians, this is expected. A thin-plate spline is shown of the consensus skull configuration is shown in Figure B7. The first relative warp largely encompasses variation in the mediolateral broadness of the skull (Figure B8). Scores in the positive direction indicate skulls that are narrower overall, although they are not proportionally longer. The nasals in particular become significantly narrower, and the junction of the premaxilla, maxilla, and nasal moves posteriorly. The second maxillary alveolus shifts posteriorly and the twelfth maxillary alveolus shifts so it is no longer located anterior to the junction of the jugal, lachrymal, and orbit. The junction of the quadratojugal and quadrate shifts posteriorly. Scores in the negative direction represent broader, stouter, more laterally festooned skulls. The orbit is smaller, although this could reflect a more lateral projection of the orbits rather than a change in size. The lateral margin of the quadratojugal move

25 13 anteriorly. The junction of the jugal, lachrymal, and maxilla moves anteriorly. The snout is broadened by shifting the maxillary alveoli anteriorly. The first relative warp best differentiates Crocodylus acutus, C. intermedius, C. johnstoni, and C. rhombifer from other species of Crocodylus. Voay is also differentiated. The second relative warp also encompasses a broadening and narrowing of the snout (Figure B9). In the positive direction, the maxillary alveoli expand laterally to broaden the snout. The jugal, quadratojugal, and quadrate are also expanded laterally. Scores in the negative direction represent extremely stenorostrine skulls. The premaxilla is mediolaterally compressed but anteroposteriorly expanded. The maxillary alveoli shift posteriorly. The jugal is mediolaterally compressed anteriorly but expanded posteriorly. The articular surface of the quadrate is nearly parallel to the posterior margin of the skull table. The second relative warp does not differentiate any species. The third relative warp shows changes largely in the location of the anterior maxillary alveoli and position of the juncture of the jugal, lachrymal, and maxilla (Figure B10). The anterior maxillary alveoli and intersection of the jugal, lachrymal, and maxilla are shifted posteriorly in positive relative warp scores and anteriorly in negative relative warp scores. The third relative warp does not differentiate any species. The fourth relative warp shows only slight deformation in the positive direction, with the posterior maxillary alveoli shifting medially (Figure B11). Scores in the negative direction show the lachrymal and prefrontal expanding anteriorly significantly. Crocodylus rhombifer is differentiated from other species on the fourth relative warp. The fifth relative warp shows deformation of the lachrymal, prefrontal, and jugal (Figure B12). In the positive direction, the jugal is expanded anteriorly and laterally. The landmark at the juncture of the jugal, lachrymal, and maxilla nearly contacts the nasal. The intersection at the prefrontal, nasal, and maxilla shifts anteriorly. The MANOVA results are summarized in Table A4.

26 14 Intraspecific Morphometric Analysis of Crocodylus niloticus sensu lato The relative warp analysis required 21 relative warps to explain a cumulative 95% of the variance (Table A5). Four relative warps explained more than 5% of the variance each. The first relative warp explained 30.02% of the variance (Figure B13). The second relative warp explained 19.41% of the variance, the third explained 8.61% of the variance, and the fourth explained 6.09% of the variance (Figures B13-B15). A thin-plate spline of the consensus skull configuration is shown in Figure B16. The first relative warp shows deformation along the lateral margin of the skull (Figure B17). Relative warp scores in the positive direction indicate skulls in which the anterior maxillary alveoli have shifted forward. The twelfth maxillary alveolus is located almost directly lateral to the intersection of the maxilla, lachrymal, and jugal. The maxilla is broader. The jugal is expanded mediolaterally, and the intersection of the lachrymal, jugal, and orbit shifts posteriorly. The lateral condyle at the intersection of the quadrat and qudratojugal has shifted anteriorly and is nearly level with the back of the skull table. Scores in the negative direction indicate a narrowing of the skull and a posterior shift of the maxillary alveoli. The lachrymal is mediolaterally compressed. The lateral condyle at the intersection of the quadrat and qudratojugal has shifted posteriorly. The second relative warp explains the mediolateral expansion and compression of the snout (Figure B18). Positive scores reflect a narrowing of the snout. The lateral margin on the maxilla is nearly linear. Negative scores show a broadening of the snout as well as a mediolateral expansion at the back of the skull. The squamosal expands laterally and posteriorly.

27 15 The third relative warp largely shows changes along the back of the skull (Figure B19). Relative warp scores in the positive direction represent a laterally expanded jugal and anteroposterially compressed quadrate. The postorbital-squamosal suture shift anteriorly. There is a slight lateral expansion of the snout. Scores in the negative direction represent skulls in which the quadrate projects posteriorly to the squamosal. The posterolateral margin of the squamosal itself is shifted posteriorly. The snout is more festooned. The fourth relative warp also explains variance in the posterior half of the skull with little change in the snout Figure (B20). Skulls in the positive direction show a quadrate that is shifted posteriorly along with the lateral limit of the quadratojugalquadrate suture. Skulls in the negative direction Intraspecific Analysis of Crocodylus niloticus sensu lato Grouped by River Basin The MANOVA shows a statistically significant difference in morphology between the Congo River C. niloticus s.l. and East African C. niloticus s.l. (p=0.0002; Table A6). There is also a statistically significant difference between the Congo River C. niloticus and the Nile River C. niloticus s.l. (p=0.0068). South African C. niloticus s.l. was not statistically significantly different from any group, but the sample size for that region was small (n=2). Nile River C. niloticus s.l. and East African C. niloticus s.l. were not statistically significantly different. Intraspecific Analysis of Crocodylus niloticus sensu lato Grouped by Morphological Cluster

28 16 The cluster analysis identified two distinct groups (1 and 2; Table A2). The first group is more tightly constrained in morphospace and is contained entirely within the Congo River Basin. The second group contains individuals from other geographic regions. The MANOVA indicates that the clusters are highly significantly different from each other (Table A7). However, the groups overlap using two-dimensional graphs of relative warps (Figure B21).

29 17 DISCUSSION A significant amount of variation exists within Crocodylus niloticus s.l. when analyzed both interspecifically and intraspecifically. When all specimens are considered, C. niloticus s.l. is the most morphologically variable species present in this study. The higher level of variation in C. niloticus s.l. is in part an artifact of sample size. Despite this, C. porosus shows as much morphological variation as C. niloticus s.l. from the Congo River basin alone, making it impossible to discount the fact that C. niloticus s.l. varies at least as much intraspecifically as crocodylian species vary from each other. Crocodylus porosus was the next largest sample size included and also has the greatest geographic range all included species. The fact that that variation is comparable to Crocodylus from a single river basin in Africa is remarkable. Despite the extent of variation within the Congo River Basin C. niloticus s.l., the sample was statistically significantly different from other regions included in the study. The Congo River Basin was also the only geographic region that included a unique group as indicated by the cluster analysis. Although the cluster shows overlap when viewed in a two-dimensional comparison between relative warps, the distinction is highly statistically significantly different between clusters (p<0.0002). The fact that the cluster is so distinct and yet shows overlap in two-dimensional morphospace shows how complex these anatomical differences are. While geometric morphometrics is extremely effective at quantifying small- and large-scale anatomical variation, it is important to consider that shape space is multidimensional and exceedingly complex. Additionally, relative warps that explain less of the variance overall may still hold key anatomical differences. Although a morphometric analysis alone is not enough to distinguish a species, the results here are largely congruent with molecular studies of C. niloticus s.l. However,

30 18 the variation exhibited within single geographic regions is indicative that the sympatry between the species of C. niloticus s.l. was likely very extensive in the recent past. Additionally, the variation within the species equals or exceeds that of other crocodylians with similarly expansive geographic ranges. Some of these, such as C. acutus, have also exhibited morphological and genetic variation that merits a reevaluation of their species status (Weaver, 2012). There is sufficient evidence to indicate that C. suchus is morphologically as well as genetically distinct. However, further osteological investigation is necessary to determine the exact delineation of the species, as the morphological variation across the range of C. niloticus s.l. is extremely complex. The extent of variation within C. niloticus s.l. from the Congo River Basin indicates that morphological variation is more complex than populational segregation by river basin, and the two morphotypes present have been sympatric in the recent past. Hekkala et al. (2011) found a shift in genotypes in samples taken after 1975 that resulted in the geographic separation of the two newly-separate species C. niloticus s.s. and C. suchus. Given that the known collection dates for the C. niloticus s.l. in this study range from , the geographic ranges of the morphotypes here likely represent the prior populational distribution elucidated in genetic studies. Whether they are sympatric today in the Congo River Basin is unknown, as no genetic samples were from the region (Hekkala, 2011). The shift in population structure merits conservation consideration. Should the historical population structure of these two morphotypes, which may represent distinct species, be finely resolved using morphometrics, it may be worth considering whether any reintroduction efforts should restore that structure. It is also important to consider the cause of this shift. It is possible that population bottlenecks and subsequent recolonization by a small population of Crocodylus contributed to the shift, although it is also possible that this is the result in changes in habitat and climate played a role.

31 19 Though only five C. rhombifer skulls were included, there is a large amount of morphological variation within that species. The largest difference lies between fossil and subfossil specimens, potentially indicating that subfossil C. rhombifer is a separate subspecies (Weaver, 2012). Additionally, the subfossil C. rhombifer skulls are substantially larger than any known modern specimen.

32 20 CRANIAL OSTEOLOGY OF CROCODYLUS NILOTICUS FMNH is a cranial specimen of Crocodylus niloticus collected in 1930 from the Chobe River, which forms part of the border between Botswana and Namibia. The skull and mandible are largely complete, although both articulars are missing (Figures B22-B24). Skull Premaxilla The premaxilla is paired and makes up the anteriormost portion of the snout. In ventral view, the premaxillae meet at the midline anterior to the external naris. Posterior to the external naris they are separated by the nasals. There is a rugosity present lateral to the external naris that corresponds to the enlarged fourth premaxillary alveolus. This rugosity expands the premaxilla laterally into the external naris, making it subcircular in shape. The premaxilla contacts the nasal posteromedially and the maxilla posterolaterally, and comes to a point where the nasal, maxilla, and premaxilla intersect. The premaxilla intersects posteriorly with the maxilla in ventral view. The sutures between the premaxilla and maxilla are irregularly shaped. From the midline, the suture angles posteromedially. It comes to a point and angles anteriomedially to form the notch with which the fourth dentary tooth occludes The incisive foramen is made up of the medial margins of the premaxillae. It has a rounded edge and comes to a point anteriorly.

33 21 Nasal The nasal is paired. It contacts the premaxilla anteriorly and itself medially. Laterally, from anterior to posterior, it contacts the maxilla, lacrimal, and prefrontal. It contacts the frontal posteriorly. The nasal narrows anteriorly where it intersects with the premaxilla, coming to a point that projects anteriorly into the external naris. In FMNH 17157, the suture is obliterated along part of the shared medial margin of the nasals. Maxilla The paired maxillae are the largest bones in the snout. Each contacts the premaxilla anteriorly. In dorsal view, it contact the nasals along their medial margin. The maxilla is excluded from the external naris. The posterior margin contacts the lacrimal medially and the jugal laterally. There is a large tuberosity toward the anterior portion of each maxilla above the large 5 th maxillary tooth. In lateral view, each maxilla has two large scallops often referred to as festooning (Mook 1924). The anterior scallop contains the expanded 5 th maxillary alveolus. In palatal view, the maxillae meet at the midline anteromedially. They are separated by the palatines posteriorly. The maxilla Where the maxilla meets the palatine at the palatal fenestra the palatines overlay the maxilla. Lacrimal The paired lacrimals lie anterior and lateral to the prefrontals. In dorsal view, the medial surface of each lacrimal contacts the nasal anteriorly and the prefrontal posteriorly. It contacts the maxilla anterolaterally and the jugal posterolaterally. The posterior margin of the lacrimal makes up the anteriormost margin of the orbit. The

34 22 posterior surface of the lacrimal contains the lacrimal duct. The suture between the maxilla and the lacrimal is highly interdigitated. Prefrontal The paired prefrontals are shaped like irregular pentagons in dorsal view. The anterior portion of each prefrontal is wedge-shaped and projects anteriorly. It does not contact the maxilla. It contacts the frontal bone posteriorly and medially, and the lacrimal laterally. The posterolateral margin of the prefrontal makes up the anteromedial portion of the orbital margin. Ventrally, the prefrontals form a pillar that contacts the dorsal surfaces of the pterygoids and the palatines. This pillar supports the secondary ossified palate. The prefrontal pillars do not contact each other except for a small medial projection of bone approximately halfway down the pillar. Frontal There is a single frontal located between the orbits. In dorsal view, it is bordered anteriorly by the nasals, anteromedially by the prefrontals, posterolaterally by the postorbitals, and posteriorly by the parietal. It is widest where it contacts the orbits. The frontal contacts the posteriormost margin of the prefrontals, then narrows as it extends forward to contact the nasals. Postorbital The postorbitals contact the frontal anteromedially and form part of the orbital margin anterolaterally. There is a medial contact with the parietal that excludes the frontal from the supratemporal fenestrae. The postorbital has a descending processes that

35 23 contacts the ascending process of the jugal to form the postorbital bar. This descending process contacts the jugal in a lap joint that is angled ventromedially. Parietal The parietal is unpaired. In dorsal view, the parietal contacts the frontal anteriorly, the squamosals posterolaterally, and has a narrow contact with the postorbitals anterolaterally. This contact excludes the frontal from the supratemporal fenestrae. Most of the lateral margins of the parietal comprise the supratemporal fenestrae. In posterior view, the parietal bone contacts the squamosal laterally. The ventral margin of the parietal makes up a portion of the posttemporal fenestrae, which is ventrally-facing a greatly reduced in crocodylians. The medial margin of the posterior edge of the parietal, which lies between the posttemporal fenestrae, is obscured by the dorsal projection of the supraoccipital. Squamosal The squamosals are paired and make up the posterior and lateral portion of the skull table. Dorsally, each squamosal contacts the postorbital anteriorly, parietal laterally, and forms the posterior margin of the supratemporal fenestra anteromedially. The squamosal extends posteromedially to the skull table and overlies the exoccipital where the exoccipital forms the paraoccipital process. There are lateral tuberosities that project anterolaterally, although the area between them is level. Laterosphenoid The paired laterosphenoids lie ventral to the frontal and parietal. It is overlain in part by the postorbital along its lateral edge. The ventral projection of the laterosphenoid

36 24 contacts the pterygoid, quadrate, and basisphenoid, where it makes up a portion of the braincase. Basisphenoid The basisphenoid is almost entirely not visible externally. It projects dorsoventrally. Dorsally, it contacts the laterosphenoid. It is surrounded by the pterygoids anteriorly and exoccipital posteriorly. It is visible ventrally, just anterior to the secondary choana. Jugal The jugal is paired and makes up the lateral margin of the orbit. It also comprises a portion of the postorbital bar. In dorsal view, it is bordered anteriorly by the maxilla, anterolaterally by the lacrimal, and posteriorly by the quadratojugal. The suture between the lacrimal and the jugal is interdigitated. In palatal view, the jugal meets the ectopterygoid medially. Quadratojugal The qudratojugal is a paired, elongate bone sandwiched between the jugal anterolaterally and the quadrate posteromedially. It comprises a portion of the posterior margin of the infratemporal fenestra. It has sharp, pointed process that projects anteriorly into the infratemportal fenestra. Quadrate The quadrate is a paired, elongate bone that makes up the portion of the lateral condyle that articulates with the articular bone in the mandible. In dorsal view it contacts

37 25 the quadratojugal anterioromedially. The anterior portion of the quadrate is overlain by the squamosal and exoccipital medially and dorsally. The quadrate contains the foramen aerium, which pneumatizes the skull. In FMNH 17157, only the right foramen aerium is present while the left has been reduced to a groove. Palatine The palatine is paired and meets at the midline. Each palatine contacts the maxilla anteriolaterally, the pterygoid posteriorly. Its lateral margin makes up the medial edge of the palatal fenestra. Ectopterygoid The ectopterygoid is paired. It is a tripartite bone, with one potion projecting anteriorly, one posterodorsally, and one ventrally. It contacts the maxilla anteriorly and medially. The ventral projection of the ectopterygoid is overlain by the pterygoid. Pterygoid The pterygoid is paired. It is a large, broad bone that extends from the midline to form a lateral pterygoidal flange. In palatal view, it is bordered anteriorly by the palatine. It comprises the posteriormost margin of the palatal fenestra. It meets the ectopterygoid anterolaterally; the pterygoid also overlays the ectopterygoid. The pterygoids meet at the midline. Dorsally, the pterygoid extends anteriorly to overlay the palatine and contact the laterosphenoid. It completely encloses the secondary choana and meets the basisphenoid posteriorly. The pterygoid has a sulcus that is oriented posteromedially.

38 26 Exoccipital The exoccipital forms a significant portion of the back of the skull table. In choanal view, it is bordered dorsally by the squamosal and supraoccipital. It makes up the medial portion of the ventral border of the posttemporal fenestra. There is a large, slightly recurved paraoccipital process that extends laterally. This process is bordered dorsally and anteriorly by the squamosal and ventrally by the quadrate. The exoccipital meets the basioccipital ventrally, although the suture between these two bones is not visible. Supraoccipital The supraoccipital is located ventral to the parietal and dorsal to the exoccipital in choanal view. It makes up a portion of the posttemporal fenestra. In FMNH 17157, the sutures between the exoccipital and supraoccipital have largely been obliterated. Basioccipital The basioccipital makes up the ventralmost portion of the back of the skull table, including the occipital condyle. The suture with the exoccipital is visible laterally but not anteromedially (dorsal to the occipital condyle). It contacts the basisphenoid along its anterior margin. Mandible Dentary The paired dentary is the largest bone in the mandible. Each dentary contains fifteen alveoli. The fourth and eleventh dentary alveoli are enlarged. Anteriorly, the dentaries meet at the midline and form the mandibular symphysis. The symphysis

39 27 extends from the first dentary alveolus to the beginning of the fifth. Beginning at the seventh dentary alveolus, the dentary lies in contact with the splenial. Posterior to the last alveolus, the dentary is overlain by the surangular. The dentary extends posteriorly to form the anterior portion of the mandibular fenestra. Angular The paired angular is a cup-shaped bone that makes up the ventral portion of the mandible posterior to the dentary. Laterally, it forms part of the ventral margin of the mandibular fenestra. Posterior to the mandibular fenestra, the angular contacts the surangular along its dorsal (or occlusal) margin. The anteromedial portion of the angular forms the bulk of the The medial surface of the angular has a large, broad lip. The angular projects posteriorly and dorsally. It supports the articular along its medial surface and forms part of the retroarticular process. Articular The articulars are missing in FMNH Surangular The paired surangular is a dorsoventrally broad bone. It projects forward above the mandibular fenestra and rests between the dentary and splenial. It contacts the angular ventrally and the articularmedially. It narrows posteriorly to form the retroarticular process along with the angular and articular.

40 28 Splenial The paired splenial is a long, dorsoventrally broad bone that lies medial to the dentary. It contacts the dentary dorsally and ventrally along its medial margins. The splenial is lingual-buccally narrow. It tapers dorsoventrally towards the anterior of the mandible. It contacts the surangular posterodorsally. The dorsomedial margin of the splenial contacts the coronoid, and the ventromedial margin contacts the angular. The splenial makes up the dorsal and anterior margins of both mandibular foramina. Coronoid The coranoid is relatively small. It contacts the splenial anteriorly, the angular ventrally, and the surangular dorsomedially. It forms the posterior and ventral surface of the foramen intermandibularis medius. The posterior surface is sharply concave. Foramina and fenestrae Foramen for first dentary tooth The foramen for the first dentary tooth is ovoid in ventral view. It expands in palatal view, where it is roughly circular. Incisive foramen The incisive foramen comes to a point anteriorly and expands into each premaxilla ventrolaterally before meeting at the midline.

41 29 External naris The external naris opens dorsally. It is bounded almost entirely by the premaxillae, with the exception of where the nasals project anteriorly. Orbit The anteriormost portion of the orbit is made up of the lacrimal. The medial margin of the orbit consists of the prefrontal and frontal. Posterior to the frontal, the postorbital contributes to the orbit. Dorsomedially, the postorbital makes up a portion of the skull table. It forms the postorbital bar along with the jugal The lateral portion of the orbit consists of the jugal, which is oriented dorsoventrally at the margin of the orbit. There is a palpebral ridge along the medial portion of the orbit that expands dorsally. Supratemporal fenestra The parietal expands to make up most of the interior portion of the supratemporal fenestrae. This results in the lateral expansion of the parietal beneath the squamosal along the ventral portion of the anterior margin of the supratemporal fenestrae. Infratemporal fenestra The infratemporal fenestra is located directly posterior to the orbit. It is significantly smaller than the orbit. It is bordered anteriorly by the postorbital bar, which is made up of the jugal and postorbital bones. The lateral border of the infratemporal fenestra is comprised of the jugal. The posterior margin is comprised of the quadratojugal and quadrate, while the dorsomedial border is comprised of the squamosal.

42 30 Palatal fenestra The palatal fenestra is paired and is ovoid in shape. It is bounded anteriorly by the maxilla, medially by the palatine, and posterolaterally by the ectopterygoid. The posteriormost margin is bounded by the pterygoid. The maxilla projects posteriorly and overlays the ectopterygoid within the fenestra. Secondary choana The secondary choana is located at the posteromedial portion of the palate. It is enclosed entirely by the pterygoids. Posttemporal fenestra The postemporal fenestrae are paired and are greatly reduced. Each fenestra is located dorsomedially to the foramen magnum. It is comprised of the parietal dorsomedially, the squamosal dorsolaterally, the exoccipital ventrolateralally, and the supraoccipital ventromedially. Cranio-quadrate passage The cranio-quadrate passage opens laterally beneath the paroccipital process. Foramen aerium The foramen aerium is a bilateral structure that contributes to skull pneumatization. It is located on medial portion of the dorsal surface of the quadrate and at a corresponding opening on the articular. However, in this specimen the left foramen aerium has been reduced to a groove. Because the articulars are missing, the effect on the left articular is unknown.

43 31 Foramina for cranial nerve V (vagus foramina) The externally visible foramina for the vagus nerve lie in a depression lateral to the foramen for cranial nerve XII. Foramen for cranial nerve X (carotid foramen) The foramen for the carotid artery lies lateral and ventral to the foramen for cranial nerve XII, just dorsal to the basioccipital. Foramina for cranial nerve XII The foramen for cranial nerve XII lies directly lateral to the foramen magnum. They are ovoid in shape and open slightly ventrally. Foramen magnum The foramen magnum opens posteriorly. It lies above the occipital condyle. Mandibular foramina and fenestra The mandibular foramen is ovoid and lies on the lateral margin of the mandible. Its margins consist of the dentary, surangular, and angular. The foramen intermandibularis caudalis is relatively elongate. It lies on the medial margin of the jaw and is comprised of the splenial and angular. The foramen intermandibularis medius is small and is comprised of the splenial and the coranoid.

44 32 Additional cranial features Dentition Teeth in skeletonized skulls are rarely preserved as they were in life due to their position being dependent on ligamentous attachment. Without adhesive, very few will remain stable in an alveolus. Most teeth have been removed from the skull and mandible in this specimen. The teeth that remain are broken, but are subcircular are appear slightly compressed linguobuccally. There is a sharp carina along the mesial and distal surfaces of the tooth. Each premaxilla contains five alveoli. The first and second maxillary alveoli are separated by the foramen for the first dentary tooth. The second alveolus is the smallest in the premaxilla and is separated from the third premaxillary alveolus by only a thin wall of bone. The fourth alveolus is enlarged relative to the other premaxillary alveoli. There is a pit between the third and fourth alveoli that fits the second (? check this) dentary tooth when the jaws occlude. The maxillae contain differing numbers of alveoli in this specimen; the left maxilla contains 13 alveoli and the right maxilla contains 14 alveoli. Based on the location of the 14 th alveolus and symmetrical appearance of the preceding alveoli, it appears to have been added at the back of the jaw. The 5 th alveolus is greatly enlarged. There are pits on either side of the 7 th alveolus for occlusion with dentary teeth. Ossified secondary palate The secondary palate consists of the premaxillae, maxillae, and palatines. It extends from the external naris to the secondary choanae.

45 33 Notes on pathology of FMNH A small overgrowth of bone, likely a tumor, is present on left palatine. There is a healed fracture at the junction of right premaxilla, maxilla, and nasal. The maxilla expanded into cavity during healing. Note on specimen maturity The laterosphenoids meet at the midline, indicating that the specimen is an adult (Brochu and Gingerich 2000). This is present in mature individuals; however, it is unknown whether the specimen had reached maximum size. Based on the ornamentation of the dorsal surface of the skull, this seems unlikely.

46 34 WORKS CITED Brochu, C.A. (1995). Heterochrony in the Crocodylian Scapulocoracoid. Society for the Study of Amphibians and Reptiles, 28(9): Brochu, C.A. (2001). Crocodylian Snouts in Space and Time: Phylogenetic Approaches Toward Adaptive Radiation. Integrative and Comparative Biology, 41 (3): Brochu, C.A. (2003). Phylogenetic Approaches Toward Crocodylian History. Annual Review of Earth and Planetary Science, 31: Brochu, C.A. (2007). Morphology, relationships, and biogeographical significance of an extinct horned crocodile (Crocodylia, Crocodylidae) from the Quaternary of Madagascar. Zoological Journal of the Linnean Society, 150: de Smet, Klaas. (1999). Status of the Nile crocodile in the Sahara desert. Hydrobiologia, 391: Davis, L.M., Glenn, T.C., Elsey, R.M., Dessauer, H.C. & Sawyer, R.H. (2001) Multiple paternity and mating patterns in the American Alligator, Alligator mississippiensis. Molecular Ecology, 10 (4): Dever, J.A., Strauss, R.E.,, Rainwater, T.R., McMurry, S.T., & Densmore, L.D. (2002). Genetic diversity, population subdivision and gene flow in Morelet s Crocodile (Crocodylus moreletii) from Belize, Central America. Copeia 2000: Drake, N.A., Blench, R.M., Armitage, S.J., Barlow, C.S., White, K.H. (2011). Ancient watercourses and biogeography of the Sahara explain the peopling of the desert. Proceedings of the National Academy of Sciences. 108(2), Eaton, M.J., Martin, A., Thorbjarnarson, J., Amato, G. (2009). Species-level diversification of African dwarf crocodiles (Genus Osteolaemus): A geographic and phylogenetic perspective. Molecular Phylogenetics and Evolution, 50: Flint, N.S., van der Bank, F.H., & Grobler, J.P. (2000). A lack of genetic variation in commercially bred Nile crocodiles (Crocodylus niloticus) in the North-West Province of South Africa. South African Water Research Commission, 26(1): Fraley, C. and Raftery, A.E. (2006). MCLUST Version 3 for R: Normal Mixture Modeling and Model-based Clustering. University of Washington, Seattle.

47 35 Fraser, D.F., Bernatchez, L. (2001). Adaptive evolutionary conservation: toward a unified concept for defining conservation units. Molecular Ecology, 10: Fuchs, K. (2006). The Crocodile Skin: Important Characteristics in Identifying Crocodilian Species. Salamandra, 10: Glenn, T.C., Dessauer, H.C. & Braun, M.J. (1998) Characterization of microsatellite DNA loci in American Alligators. Copeia 1998 (3): Goodall, C. (1991). Procrustes Methods in the Statistical Analysis of Shape. Journal of the Royal Statistical Society, 53(2): Hammer, Ø., Harper, D.A.T., and Ryan, P.D. (2001). PAST: Paleontological statistics software package for education and data analysis. Paleontol. Electron. 4. Hammer, Ø., and Harper,D.A.T. (2006). Paleontological data analysis. Blackwell, Oxford. Hekkala, E.R., Amato, G., DeSalle, R., & Blum, M.J. (2010). Molecular assessment of population differentiation and individual assignment potential of Nile crocodile (Crocodylus niloticus) populations. Conservation Genetics, 11: Hekkala, E., Shirley, M.H., Amato, G., Austin, J.G., Charter, S.L., Thorbjarnarson, J., Vliet, K.A., Houck, M.L., Desalle, R., Blum, M.J., (2011). An ancient icon reveals new myseteries: mummy DNA resurrects a cryptic species within the Nile crocodile. Molecular Ecology, 20: Langston, W. (1973). The crocodilian skull in historical perspective, In C. Gans and T. Parsons (eds.), Biology of the Reptilia. Vol. 4, pp Academic Press, London. Laurenti, J.N. (1768). Specimen medicum, exhibens synopsin reptilium emendatam cum experimentis circa venena et antidota reptilium austracorum, quod authoritate et consensu. Thomaei, Vienna. McAliley, Willis, Ray, D.A., White, Brochu, C.A., and Densmore III, L.D. (2006). Are crocodiles really monophyletic? Evidence for subdivisions from sequence and morphological data. Molecular Phylogenetics and Evolution, 39: Meganathan, P.R., Dubey, B.,Batzer, M.A., Ray, D.A., and Haque, I. (2010). Molecular phylogenetic analyses of genus Crocodylus (Eusuchia, Crocodylia, Crocodylidae) and the taxonomic position of Crocodylus porosus. Molecular Phylogenetics and Evolution 57 (1): Mook, C.C. (1921). Skull characters of Recent Crocodilia with notes on the affinities of the Recent genera. Bulletin of the American Museum of Natural History, 44(13).

48 36 Pierce, S.E., Angielczyk, K.D., & Rayfield, E.J. (2008). Patterns of Morphospace Occupation and Mechanical Performance in Extant Crocodilian Skulls: A Combined Geometric Morphometric and Finite Element Modeling Approach. Journal of Morphology, 269: Piras, P., Colangelo, P., Adams, D.C., Buscalioni, A., Cubo, J., Kotsakis, T., Meloro, C., and Raia, P. (2010). The Gavialis-Tomistoma debate: the contribution of skull ontogenetic allometry and growth trajectories to the styudy of crocodilian relationships. Evolution and Development, 12(6): Piras, P., Teresi, L., Buscalioni, A.D., and Cubo, J. (2009). The shadow of forgotten ancestors differently constrains the fate of Alligatoroidea and Crocodyloidea. Global Ecology and Biogeography, 18: R Development Core Team (2008). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Ray, D.A., Dever, J.A., Platt, S.G., Rainwater, T.R., Finger, A.G., McMurry, S.T., Batzer, M.A., Barr, B., Stafford, P.J., McKnight, J., & Densmore, L.D. (2004). Low levels of nucleotide diversity in Crocodylus moreletii and evidence of hybridization with C. acutus. Conservation Genetics, 5: Rohlf, F.J. (2010). TPS Program Suite: Available at: Accessed December Ross, J.P. (1998). Crocodiles: An Action Plan for their Conservation. IUCN/SSC Crocodile Specialist Group Sadleir, R.W. (2009). A morphometric study of crocodilian ecomorphology through ontogeny and phylogeny (Doctoral dissertation). University of Chicago, Chicago Illinois. Sadleir, R.W., and Makovicky. (2008). Cranial shape and correlated characters in crocodilian evolution. Journal of Evolutionary Biology, 21: Schmitz, A. Mausfeld, P., Hekkala, E., Shine, T., Nickel, H., Amato, G, and Wolfgang. B. (2003). Molecular evidence for species level divergence in African Nile Crocodiles Crocodylus niloticus (Laurenti 1786). C.R. Paleovol, 2: Schneider, C.A., Rasband, W.S., Eliceiri, K.W. (2012). NIH Image to ImageJ: 25 years of image analysis. Nature Methods, 9:

49 37 Shine, T., Böhme,W., Nickel, H., Thies, D.F., & Wilms, T. (2001). Rediscovery of relict populations of the Nile crocodile Crocodylus niloticus in south-eastern Mauritania, with observations on their natural history. Oryx, 35(3): Smith, R.A. (1971). The effect of unequal group size on Tukey s HSD procedure. Psychometrika. 25(1): Stejneger, L. (1933). Crocodilian Nomenclature. Copeia, 3: Weaver, J.M. (2012).Geometric Morphometrics of Antillean Crocodiles (Master s Thesis). Texas Tech University, Lubbock, Texas. Wermuth, Heinz Systematik der rentzen Krokodile. Mitt. Zool. Mus. Berlin. 29: Zelditch, M.L., Swiderski, D.L., Sheets, H.D., Fink, W.L. (2004). Geometric Morphometrics for Biologists: A Primer. San Diego, California: Elsevier Academic Press.

50 38 APPENDIX A. TABLES Table A1. List of non-crocodylus niloticus specimens. Species Museum Specimen Species Museum Specimen Crocodylus acutus MNHN C. porosus MFN 2811 C. acutus MNHN A5310 C. porosus MFN C. acutus MNHN NFD C. porosus MFN C. acutus SMF C. porosus MFN C. acutus SMF C. porosus MFN C. acutus SMF C. porosus MFN C. acutus SMF C. porosus MFN C. intermedius MNHN C. porosus MNHN C. intermedius MNHN C. porosus MNHN C. intermedius SMF C. porosus MNHN C. intermedius SMF C. porosus MNHN A5316 C. intermedius SMF C. porosus MNHN A5317 C. johnstoni SMF C. porosus RBINS C. johnstoni SMF C. porosus RBINS * C. megarhinus YPM Fayum C. porosus RBINS 160 C. moreletii MFN C. porosus RBINS 161b C. moreletii MFN C. porosus RBINS 161c C. moreletii SMF C. porosus SMF C. novaguinae SMF C. porosus SMF C. novaguinae SMF C. porosus SMF C. novaguinae SMF *C. rhombifer AMNH 6179 C. palustris MNHN *C. rhombifer AMNH 6185 C. palustris SMF C. rhombifer Cuba 1 C. palustris SMF C. rhombifer Cuba 3 C. palustris SMF C. rhombifer Cuba 4 C. palustris SMF *Voay robustus AMNH 3101 C. palustris SMF C. siamensis MNHN C. porosus MFN C. siamensis SMF Note: Institutional abbreviations: MNHN (Museum national d Histoire naturelle), SMF (Senckenberg Museum), YPM (Yale Peabody Museum), MFN (Museum für Naturkunde), RBINS (Royal Belgian Institute for Natural Sciences), AMNH (American Museum of Natural History). Asterisks indicate fossil and subfossil specimens.

51 39 Table A2. List of Crocodylus niloticus specimens. Species Museum Identification Date of Collection River Basin Morphotype Cluster C. niloticus AMNH Nile 1 C. niloticus AMNH 7136 SAC 1 C. niloticus MFN EAC 1 C. niloticus MFN Congo 1 C. niloticus MFN EAC 1 C. niloticus OR 2? EAC 1 C. niloticus OR 274 EAC 1 C. niloticus OR 274 EAC 1 C. niloticus OR 275 EAC 1 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 2 C. niloticus RBINS 3310 Congo 1 C. niloticus RBINS 3313 Congo 1 C. niloticus RBINS 3632 Congo 1 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 1 Note: River basin abbreviations are as follows: Congo (Congo River Basin), EAC (East African Complex), Niger (Niger River Basin), Nile (Nile River Basin), SAC (South African Complex), and Zambezi (Zambezi River Basin).

52 40 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 1 C. niloticus RBINS Congo 2 C. niloticus RBINS Congo 2 C. niloticus RBINS 5859 Congo 1 C. niloticus RMCA R Congo 1 C. niloticus RMCA RG Congo 1 C. niloticus RMCA RG EAC 1 C. niloticus RMCA RG Zambezi 1 C. niloticus RMCA RG Congo 1 C. niloticus RMCA RG Congo 1 C. niloticus RMCA RG Congo 1 C. niloticus RMCA RG Congo 2 C. niloticus RMCA RG Congo 1 C. niloticus SMF Nile 1 C. niloticus SMF Nile 1 C. niloticus SMF Nile 1 C. niloticus SMF Nile 1 C. niloticus SMF EAC 1 C. niloticus SMF EAC 1 C. niloticus SMF EAC 1 C. niloticus SMF Niger 1 C. niloticus SMF EAC 1 C. niloticus SMF EAC 1 C. niloticus SMF SAC 1 Table A2 continued.

53 41 Table A3. Variance explained by relative warps of the interspecific analysis of Crocodylus. Relative Warp Percent Variance Explained % 32.24% % 50.56% % 59.16% % 65.16% % 70.77% % 73.54% % 76.05% % 78.50% % 80.59% % 82.61% % 84.43% % 86.02% % 87.50% % 88.74% % 89.81% % 90.71% % 91.44% % 92.15% % 92.84% % 93.44% % 94.01% % 94.53% % 95.00% Cumulative Variance Explained

54 42 Table A4. p values of the one-way non-parametric multivariate analysis of variance of the interspecific analysis of Crocodylus. C. acutus Congo EAC C. intermedius C. johnstoni C. moreletii Nile C. novaguinae C. palustris C. porosus C. rhombifer SAC C. acutus Congo E EAC C. intermedius C. johnsoni C. moreletii Nile C. novaguinae C. palustris C. porosus C. rhombifer SAC C. siamensis Note: Statistically significant values (p<0.05) are shown in bold. C. siamensis

55 43 Table A5. Variance explained by relative warps of the intraspecific analysis of Crocodylus niloticus sensu lato. Relative Warp Percent Variance Explained % 30.02% % 49.43% % 58.05% % 64.14% % 69.05% % 73.00% % 76.25% % 79.25% % 81.49% % 83.63% % 85.26% % 86.90% % 88.27% % 89.48% % 90.58% % 91.59% % 92.48% % 93.26% % 93.97% % 94.54% % 95.04% Cumulative Variance Explained

56 44 Table A6. p values of the one-way non-parametric multivariate analysis of variance of the intraspecific analysis of Crocodylus niloticus s.l. separated into groups by river basin. Congo EAC Nile SAC Congo EAC Nile SAC Note: Statistically significant values (p<0.05) are shown in bold.

57 45 Table A7. p values of the one-way non-parametric multivariate analysis of variance of the intraspecific analysis of Crocodylus niloticus s.l. separated into groups using a modelbased cluster analysis. Group 1 Group 2 Group Group

58 46 APPENDIX B. FIGURES Figure B1. Landmarks used in the morphometric analysis. Note: Alveoli are indicated by green landmarks, which were coded in palatal view and superimposed in dorsal view.

59 Figure B2. Map of large-scale river basins on the African continent. 47

60 48 Crocodylus acutus C. megarhinus C. intermedius Voay robustus C. johnstoni Congo River C. niloticus C. moreletii East African C. niloticus C. novaguinae Nile C. niloticus C. palustris South African C. niloticus C. porosus Niger C. niloticus C. rhombifer Zambezi C. niloticus C. siamensis Figure B3. Relative warps 1 and 2 of the interspecific analysis of Crocodylus.

61 49 Crocodylus acutus C. megarhinus C. intermedius Voay robustus C. johnstoni Congo River C. niloticus C. moreletii East African C. niloticus C. novaguinae Nile C. niloticus C. palustris South African C. niloticus C. porosus Niger C. niloticus C. rhombifer Zambezi C. niloticus C. siamensis Figure B4. Relative warps 1 and 3 of the interspecific analysis of Crocodylus.

62 50 Crocodylus acutus C. megarhinus C. intermedius Voay robustus C. johnstoni Congo River C. niloticus C. moreletii East African C. niloticus C. novaguinae Nile C. niloticus C. palustris South African C. niloticus C. porosus Niger C. niloticus C. rhombifer Zambezi C. niloticus C. siamensis Figure B5. Relative warps 1 and 4 of the interspecific analysis of Crocodylus.

63 51 Crocodylus acutus C. megarhinus C. intermedius Voay robustus C. johnstoni Congo River C. niloticus C. moreletii East African C. niloticus C. novaguinae Nile C. niloticus C. palustris South African C. niloticus C. porosus Niger C. niloticus C. rhombifer Zambezi C. niloticus C. siamensis Figure B6. Relative warps 1 and 5 of the interspecific analysis of Crocodylus.

64 Figure B7. Thin-plate spline of the consensus configuration of the interspecific analysis. Note: The orbit is shaded. 52

65 53 a b Figure B8. Thin-plate splines of the first relative warp of the interspecific analysis showing a) minimum relative warp values and b) maximum relative warp values. Note: The orbit is shaded.

66 54 a b Figure B9. Thin-plate splines of the second relative warp of the interspecific analysis showing a) minimum relative warp values and b) minimum relative warp values. Note: The orbit is shaded.

67 55 a b Figure B10. Thin-plate splines of the third relative warp of the interspecific analysis showing a) maximum relative warp values and b) minimum relative warp values. Note: The orbit is shaded.

68 56 a b Figure B11. Thin-plate splines of the fourth relative warp of the interspecific analysis showing a) maximum relative warp values and b) minimum relative warp values. Note: The orbit is shaded.

69 57 a b Figure B12. Thin-plate splines of the fifth relative warp of the interspecific analysis showing a) maximum relative warp values and b) minimum relative warp values. Note: The orbit is shaded.

70 58 Congo River C. niloticus East African C. niloticus Niger C. niloticus Nile C. niloticus South African C. niloticus Zambezi C. niloticus Figure B13. Relative warps 1 and 2 of the intraspecific analysis of Crocodylus niloticus sensu lato classified by geographic region.

71 59 Congo River C. niloticus East African C. niloticus Niger C. niloticus Nile C. niloticus South African C. niloticus Zambezi C. niloticus Figure B14. Relative warps 1 and 3 of the intraspecific analysis of Crocodylus niloticus s.l. classified by geographic region.

72 60 Congo River C. niloticus East African C. niloticus Niger C. niloticus Nile C. niloticus South African C. niloticus Zambezi C. niloticus Figure B15. Relative warps 1 and 4 of the intraspecific analysis of Crocodylus niloticus s.l. classified by geographic region.

73 Figure B16. Thin-plate spline of the consensus configuration of the intraspecific analysis of Crocodylus niloticus s.l. Note: The orbit is shaded. 61

74 62 a b Figure B17. Thin-plate splines of the first relative warp of the intraspecific analysis of Crocodylus niloticus showing a) minimum relative warp values values and b) maximum relative warp values. Note: The orbit is shaded.

75 63 a b Figure B18. Thin-plate splines of the second relative warp of the intraspecific analysis of Crocodylus niloticus showing a) minimum relative warp values and b) minimum relative warp values. Note: The orbit is shaded.

76 64 a b Figure B19. Thin-plate splines of the third relative warp of the intraspecific analysis of Crocodylus niloticus showing a) maximum relative warp values and b) minimum relative warp values. Note: The orbit is shaded.

77 65 a b Figure B20. Thin-plate splines of the fourth relative warp of the intraspecific analysis of Crocodylus niloticus showing a) maximum relative warp values and b) minimum relative warp values. Note: The orbit is shaded.

78 66 Cluster 1: Congo River C. niloticus Cluster 2: Congo River C. niloticus Cluster 2: East African C. niloticus Cluster 2: Niger C. niloticus Cluster 2: Nile C. niloticus Cluster 2: South African C. niloticus Cluster 2:Zambezi C. niloticus Figure B21. Results of the cluster analysis of Crocodylus niloticus s.l.

79 Figure B22. Crocodylus niloticus (FMNH 17157) skull shown in dorsal view. Note: Dashed lines indicate where sutures have fused but would be present. 67

80 Figure B23. Crocodylus niloticus (FMNH 17157) skull shown in palatal view. 68

81 Figure B24. Crocodylus niloticus (FMNH 17157) skull shown in posterior view. 69

82 Figure 25. Crocodylus niloticus (FMNH 17157) mandible shown in occlusal view. Note: the articulars are missing. 70