UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Cranial shape correlates with diet specialization in northeast Pacific killer whale (Orcinus orca) ecotypes. Fung, Charissa W. 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2016_november_fung_charissa.pdf [ 4.3MB ]
Metadata
JSON: 24-1.0314137.json
JSON-LD: 24-1.0314137-ld.json
RDF/XML (Pretty): 24-1.0314137-rdf.xml
RDF/JSON: 24-1.0314137-rdf.json
Turtle: 24-1.0314137-turtle.txt
N-Triples: 24-1.0314137-rdf-ntriples.txt
Original Record: 24-1.0314137-source.json
Full Text
24-1.0314137-fulltext.txt
Citation
24-1.0314137.ris

Full Text

1 CRANIAL SHAPE CORRELATES WITH DIET SPECIALIZATION IN NORTHEAST PACIFIC KILLER WHALE (ORCINUS ORCA) ECOTYPES. by  Charissa W. Fung  B.Sc., The University of Calgary, 1999  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Zoology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2016  © Charissa W. Fung, 2016   ii Abstract  Resident, transient (Bigg’s), and offshore killer whales (Orcinus orca) live in sympatric and parapatric ranges in the northeast Pacific Ocean. These ecotypes have different vocal repertoires (Ford and Fisher, 1982; Ford, 1991; Yurk, 2002), echolocation use (Barrett-Lennard et al., 1996), foraging strategies (Bigg et al., 1987; Ford et al., 1998; Baird et al., 1992; Deecke et al., 2002; Ford et al., 2011), and sociobiology (Ford and Fisher, 1982; Bigg et al., 1987; Deecke et al., 2000; Baird and Whitehead, 2000; Riesch et al., 2012). Genetic studies corroborate the behavioural evidence that the resident and transient (Bigg’s) populations are reproductively isolated despite the absence of any geographic or temporal barrier (Stevens et al., 1989; Barrett-Lennard, 2000; Hoelzel and Dover, 1991; Morin et al., 2010). The behavioural segregation between the sympatric ecotypes is apparently maintained by cultural mechanisms alone, which is extremely unusual among non-human mammalian species (Barrett-Lennard, 2000; Riesch et al., 2012). These ecotypes also exhibit dramatic resource polymorphisms: resident killer whales feed exclusively on fish, transient (Bigg’s) killer whales primarily hunt marine mammals (Bigg, 1982; Baird et al., 1992, Ford et al., 1998) and offshore killer whales are thought to feed on fishes including Pacific sleeper shark (Somniosus pacificus) (Ford et al., 2011). We do not know if cranial features related to capturing and processing prey have evolved to reflect the dramatic dietary specializations observed in these three ecotypes. The goal of this research was to determine whether there has been divergence of cranial morphology among the three ecotypes. To this end, I measured and compared cranial shape using traditional and geometric morphometrics techniques. I found that transient (Bigg’s) killer whales that bite and tear apart large mammals have more robust cranial skeletons than the piscivorous resident and offshore killer whales that handle smaller prey items. I found that resident and transient (Bigg’s) killer whales are distinguishable based on skull width, rostral width, and mandibular shape, and that offshore killer whales have a more variable morphology that precludes identification based on cranial shape alone.      iii Preface Charissa Fung developed this research project under the guidance of Dr. William K. Milsom and Dr. Lance Barrett-Lennard. All of the morphological data were collected and analyzed by C. Fung, with statistical advice from Dr. Douglas Altshuler and Dr. Roslyn Dakin, and R programming assistance from Shannon Obradovich. Dr. Lance Barrett-Lennard and Allyson Miscampbell provided genetic identification of skeletal material.    iv Table of contents  Abstract ................................................................................................................................................. ii Preface .................................................................................................................................................. iii Table of contents................................................................................................................................... iv List of figures ........................................................................................................................................ ix List of abbreviations: museums and institutions ................................................................................ xii List of abbreviations: data collection ................................................................................................. xiii List of abbreviations: data analysis .................................................................................................... xv Acknowledgements ............................................................................................................................. xvi Dedication ......................................................................................................................................... xviii Chapter 1: General introduction .......................................................................................................... 1 1.1 Killer whale (Orcinus orca) diversification and taxonomic uncertainty .................................. 1 1.2 Resident, transient (Bigg’s), and offshore killer whale ecotypes of the northeast Pacific Ocean .....................................................................................................................................................1 1.2.1 Distribution and movement patterns of northeast Pacific ecotypes................................... 2 1.2.2 Behavioural differences between ecotypes ...................................................................... 4 1.2.3 Field observations of morphological differences in ecotypes ........................................... 5 1.2.4 Dietary differences/trophic level differences among ecotypes ......................................... 6 1.2.4.1 Resident killer whales hunt salmonids ..................................................................... 6 1.2.4.2 Offshore killer whales feed on fishes (e.g., Pacific sleeper shark) ............................ 7 1.2.4.3 Transient (Bigg’s) killer whales hunt marine mammals............................................ 7 1.2.5 Dentition differences between ecotypes .......................................................................... 9 1.2.6 Inference of jaw function from simplified lever system and muscle attachment ............. 10 1.3 Summary................................................................................................................................. 10   v Chapter 2: Comparison of cranial shape among northeast Pacific killer whale ecotypes ................. 11 2.1 Introduction ......................................................................................................................... 11 2.2 Materials and methods ......................................................................................................... 12 2.2.1 Measuring cranial morphology: geometric morphometrics and traditional morphometrics ............................................................................................................................................12 2.2.2 Cranial skeleton anatomy.............................................................................................. 13 2.2.3 Materials ...................................................................................................................... 18 2.2.3.1 Specimens ............................................................................................................. 18 2.2.3.2 Ecotype assignment ............................................................................................... 19 2.2.4 Traditional morphometrics measurements ..................................................................... 22 2.2.4.1 Linear measurements............................................................................................. 22 2.2.4.2 Traditional morphometrics - analysis ..................................................................... 22 2.2.5 Geometric morphometrics ............................................................................................ 26 2.2.5.1 Standardized photography protocol ....................................................................... 26 2.2.5.2 Landmark digitization from standardized photographs and analyses of landmark coordinates ........................................................................................................................... 28 2.2.5.3 Geometric morphometrics - analysis ...................................................................... 28 2.3 Results................................................................................................................................. 30 2.3.1 Total body length and condylobasal length measurements............................................. 30 2.3.2 Results from traditional morphometrics of cranium and dentary bone ........................... 30 2.3.2.1 Overall skull width and height across the vault and occipital regions ..................... 33 2.3.2.2 Rostrum (facial bones and the palate) .................................................................... 33 2.3.2.3 Bones associated with jaw adductor muscles (temporalis) ...................................... 33 2.3.2.4 Length of tooth-bearing alveolar bones on maxilla and dentary .............................. 44 2.3.2.5 Dentary bone shape ............................................................................................... 44   vi 2.3.2.6 The foramen magnum, occipital condyles, and internal nares ................................. 44 2.3.3 Results from geometric morphometrics ......................................................................... 50 2.3.3.1 Cranium, dorsal view ............................................................................................ 50 2.3.3.1.1 PCA of Procrustes coordinates, dorsal view of cranium .................................. 50 2.3.3.1.2 CVA of Procrustes coordinates, dorsal view of cranium ................................. 54 2.3.3.2 Left dentary, labial view ........................................................................................ 56 2.3.3.2.1 PCA of Procrustes coordinates, dentary bone ................................................. 56 2.3.3.2.2 CVA of Procrustes coordinates, dentary bone................................................. 57 2.4 Discussion ........................................................................................................................... 61 2.4.1 Total body length and condylobasal length measurements............................................. 61 2.4.1.1 Functional inferences: do differences in overall body length reflect differences in dietary specialization ............................................................................................................ 61 2.4.2 Traditional morphometrics............................................................................................ 62 2.4.2.1 Possible morphological correlates to bite strength: skull width, and size of temporal fenestrae and fossae .............................................................................................................. 62 2.4.2.2 Resistance to torsional and bending forces: the rostrum and palate ......................... 64 2.4.2.3 Alveolar (tooth-bearing) bone................................................................................ 65 2.4.2.4 Dentary characteristics .......................................................................................... 66 2.4.2.5 Measurements of the foramen magnum, occipital condyles and nares .................... 66 2.4.3 Geometric morphometrics ............................................................................................ 66 2.4.3.1 PCA and CVA on skull landmarks ........................................................................ 67 2.4.3.2 PCA and CVA on dentary landmarks .................................................................... 67 2.4.4 The effect of diet on cranial morphology – plasticity, genetics, and age......................... 68 2.4.5 Sexual dimorphism and allometry ................................................................................. 69 2.4.6 Summary/conclusions................................................................................................... 70   vii Chapter 3: Conclusion ......................................................................................................................... 73 3.1 The trouble with offshores ................................................................................................... 73 3.2 False discovery rate control.................................................................................................. 74 3.3 Future directions .................................................................................................................. 75 3.4 Killer whale taxonomy......................................................................................................... 76 References ............................................................................................................................................ 77 Appendices ........................................................................................................................................... 86 Appendix A Specimen information ................................................................................................ 86 A.1 General location, collection year, ecotype, and ecotype evidence .............................. 86 A.2 Specimen source or collection locations.................................................................... 92 Appendix B Geometric morphometrics landmarks.......................................................................... 95 B.1 Landmarks on dorsal aspect of cranium .................................................................... 95 B.2 Landmarks on labial view of left dentary .................................................................. 96 Appendix C Two-stage Benjamini & Hochberg step-up false discovery rate - controlling procedure: raw p-values and adjusted p-values ............................................................................. 97 Appendix D Geometric morphometrics analysis for dorsal skull ..................................................... 98 D.1 Eigenvalues and percent variance (dorsal skull landmarks) ....................................... 98 D.2 Principal component coefficients (dorsal skull landmarks) ........................................ 99 D.3 Canonical coefficients for Procrustes coordinates of the dorsal skull. ...................... 107 Appendix E Geometric morphometrics analysis of left dentary ..................................................... 108 E.1 Eigenvalues and percent variance (left dentary landmarks) ..................................... 108 E.2 Principal component coefficients (left dentary landmarks) ...................................... 109 E.3 Canonical coefficients for Procrustes coordinates of the left dentary ....................... 113 Appendix F Sample photographs of resident, offshore and transient (Bigg's) left dentary bones and skulls (dorsal view). ................................................................................................. 114   viii F.1 Offshore left dentary (A) SIRS 0120, female; (B) NMML 0087, male; (C) NMML 0080, female. Scale bars are 30 cm rulers. .............................................................. 114 F.2 Resident left dentary (A) NMML 0090, female and (B) RBCM 16814, male. ......... 115 F.3 Transient left dentary (A) LACM 84291, female and (B) RBCM 10001, male. ....... 116 F.4 Offshore dorsal skull (A) SIRS 0120, female, and (B) NMML 0087, male. ............. 117 F.5 Resident dorsal skull (A) NMML 0090, female and (B) RBCM 16814, male. ......... 118 F.6 Transient dorsal skull (A) MVZ 129686, female and (B) RBCM 10001, male. ........ 119    ix List of figures Figure 1. Map of the approximate ranges of northeast Pacific resident and offshore killer whale populations .......... 3       Figure 2. Map of the approximate ranges of northeast Pacific transient (Bigg’s) killer whale populations shown relative to the resident and offshore killer whale ecotypes ........................................................................................... 4 Figure 3. Photograph and illustration of adult male resident killer whale (O. orca), left view of skull (RBCM8386) ...................................................................................................................................................................................... 14 Figure 4. Photograph and illustration of adult male resident killer whale (O. orca), dorsal view of skull (RBCM8386) .............................................................................................................................................................. 15 Figure 5. Photograph and illustration of adult male resident killer whale (O. orca), ventral view of skull (RBCM8386) .............................................................................................................................................................. 16 Figure 6. Photograph and illustration of adult male resident killer whale (O. orca), occipital view of skull (RBCM8386) .............................................................................................................................................................. 17 Figure 7. Photograph and illustration of adult male resident killer whale (O. orca), (A) labial view and (B) lingual view of left dentary bone (RBCM8386) ..................................................................................................................... 18 Figure 8. Collection locations of specimens from the northeast Pacific (Alaska, British Columbia, and Washington State) ........................................................................................................................................................................... 20 Figure 9. Collection locations of specimens from the northeast Pacific (California and Mexico) ............................ 21 Figure 10. Measurements from the dorsal (A), ventral (B), and left lateral (C) views of cranium ........................... 24 Figure 11. Measurements from the occipital view of the cranium ............................................................................. 25 Figure 12. Measurements from the labial view (A) and the lingual view (B) of the left dentary .............................. 25 Figure 13. The left postero-dorsal view of the skull showing the temporal fenestre through which the temporalis muscles pass ................................................................................................................................................................ 26 Figure 14. 21 landmarks digitized from standardized photographs of the skull (dorsal view) .................................. 29 Figure 15. 14 landmarks digitized from standardized photographs of the left dentary (labial view) ........................ 29 Figure 16. Boxplot showing the effect of ecotype on the total body length, TLF ..................................................... 31 Figure 17. Boxplot showing the effect of ecotype on condylobasal length, SKL ..................................................... 32    x Figure 18. Boxplot showing the effect of ecotype on the relative post-orbital width of the cranium, POS/SKL ..... 34 Figure 19. Boxplot of the relative width of the cranium as measured on the exoccipitals for each ecotype, OCW/SKL ................................................................................................................................................................... 35 Figure 20. Boxplot showing no difference among ecotypes for measurements of the relative height of the cranium, OCH/SKL .................................................................................................................................................................... 36 Figure 21. Boxplot of the relative distance between the anterior-most edges of the antorbital processes for each ecotype, WAP/SKL...................................................................................................................................................... 37 Figure 22. Boxplot showing the effect of ecotype on the greatest relative width of the rostrum, WRP/SKL ........... 38 Figure 23. Boxplot showing the effect of ecotype on the relative width of the rostrum at the antorbital notches WAN/SKL .................................................................................................................................................................. 39 Figure 24. Boxplot showing the effect of ecotype on the relative width of the palate, WAL/SKL ........................... 40 Figure 25. Boxplot showing the effect of ecotype on the (A) relative width, WPF/SKL, and (B) relative length, LPF/SKL of the temporal fossae ................................................................................................................................. 41 Figure 26. Boxplot showing the effect of ecotype on (A) the relative major diameter of the left temporal fenestre, MAD/SKL and (B) the relative minor diameter of left temporal fenestre, MID/SKL ............................................... 42 Figure 27. Boxplot showing the relationship between ecotype and (A) the relative length of the upper left tooth row, LMT/SKL, and (B) the relative length of the dentary tooth row, AMN/SKL ............................................................ 43 Figure 28. Boxplot showing the effect of ecotype on the relative length of the left dentary (LLR/SKL).................. 45 Figure 29. Boxplot showing the effect of ecotype on (A) the relative length of the mandibular symphysis, LSY/SKL and (B) the relative length of the mandibular fossa, LLF/SKL................................................................................... 46 Figure 30. Boxplot showing the effect of ecotype on (A) the relative foramen magnum width, FMW/SKL and (B) relative foramen magnum height, FMH/SKL.............................................................................................................. 47 Figure 31. Boxplot showing (A) the relative length of the left occipital condyle (LOC/SKL) and (B) the relative width across both occipital condyles (WOC/SKL) ..................................................................................................... 48 Figure 32. Boxplot showing the relative width of the internal nares, WIN/SKL ...................................................... 49 Figure 33. The percentage contribution to the variance in shape from each principal component, for the shape PCA of the dorsal view of the skull ..................................................................................................................................... 51   xi Figure 34. Scatterplot of PCA 1 and PCA 2 scores for Procrustes coordinates from the dorsal view of the skull. The wireframe graphs show the shape changes associated with PC 1 and PC 2................................................................ 52 Figure 35. Scatterplot of PCA 3 and PCA 4 scores for Procrustes coordinates from the dorsal view of the skull. The wireframe graphs show the shape changes associated with PC 3 and PC 4 ............................................................... 53 Figure 36. Scatterplot of CV 1 and CV 2 values for skulls of individuals from each ecotype. The wireframe graphs indicate the shape changes associated with CV1 and CV2 ......................................................................................... 55 Figure 37. The percentage contribution to the variance in shape from each principal component, for shape PCA of the left dentary ............................................................................................................................................................ 57 Figure 38. Scatterplot of PC1 and PC2 scores for the Procrustes coordinates of each dentary bone. The wireframe graphs show the shape changes that are associated with PC 1 and PC 2 .................................................................... 58 Figure 39. Scatterplot of PC3 and PC4 scores for the Procrustes coordinates of each dentary bone. The wireframe graphs show the shape changes that are associated with PC 3 and PC 4 .................................................................... 59 Figure 40. Scatter plot of CV 1 and CV 2 values for the dentary bones from individuals from each ecotype. The wireframe graphs for CV1 and CV2 indicate the shape changes associated with each canonical variate ................. 60 Figure 41. Left dentary of LACM84291, a 27-year-old female transient (Bigg’s) killer whale that was kept in captivity for 20 years ................................................................................................................................................... 69 Figure 42. (A) Left and (B) right dentaries (labial view) of RBCM16006 (southern resident L66, female, age 62 years) ........................................................................................................................................................................... 69        xii List of abbreviations: museums and institutions CAS   California Academy of Sciences, San Francisco, California COWAN Cowan Vertebrate Museum (Beatty Biodiversity Museum), University of British Columbia, Vancouver; also known as UBCBBM  JSKWICS Johnstone Strait Killer Whale Interpretive Centre Society, Telegraph Cove, British Columbia  KHS   Killarney High School, Vancouver, British Columbia LACM   Los Angeles County Museum, Los Angeles, California MVZ   Berkeley Vertebrate Museum of Zoology, Berkeley, California NMML   National Marine Mammal Laboratory, Northwest Fisheries Science Center, Seattle,     Washington  RBCM   Royal British Columbia Museum, Victoria, British Columbia ROM   Royal Ontario Museum, Toronto, Ontario SBMNH  Santa Barbara Museum of Natural History, Santa Barbara, California SIRS   Strawberry Isle Marine Research Society, Tofino, British Columbia SSSC   Sitka Sound Science Centre, Sitka, Alaska SWFSC   Southwest Fisheries Science Centre, La Jolla, California USNMA Smithsonian Institution, National Museum of Natural History, Washington, District of Columbia  VAMSC  Vancouver Aquarium Marine Science Centre, Vancouver, British Columbia      xiii List of abbreviations: data collection AMN length lower left tooth row, from hindmost margin of hindmost alveolus to tip of dentary (cm) ECO  ecotype f  female FMH  greatest height foramen magnum (cm) FMW  greatest width foramen magnum (cm) GEN  ecotype designation from genetics (haplotype) LAT  latitude LLF length of left mandibular fossa, measured to mesial rim of internal surface of mandibular condyle (measured to closest 0.1 cm) LLR  greatest length of ramus, left dentary (measured to closest 0.1cm) LMT  length of upper left tooth row (measured on maxilla only not premaxilla, cm) LOC  locality, specific LOG  locality, general LON  longitude LPF greatest length of left post temporal, to edge of frontal bone (cm) LSY  greatest length of mandibular symphysis, measured on left dentary (cm) m  male MAD  major diameter of left temporal fenestre (cm) MID  minor diameter of left temporal fenestre (cm) OCW  maximum width of occipital view of cranium (cm) OCH  total height of cranium, occipital view (cm) POS  greatest postorbital width (measured at squamosals) (cm) SEX  sex (male, female, or unknown) SKL  skull total length, or condylobasal length (cm) TLF  total body length as measured in the field or during necropsy (cm) unk  unknown   xiv WAL distance between last two alveoli in maxillae (from rear of tooth margin to opposite side) (cm) WAN  width of rostrum at antorbital notch (cm) WAP greatest width at antorbital process, suture between maxilla, frontal and lacrimal (cm) WIN  greatest width of internal nares, from ventral view (cm) WOC  width across occipital condyle (cm) WPF greatest width of post-temporal fossa at right angles to greatest length, (cm) WRP  greatest width of rostrum anterior to antorbital notch; from processes (cm) YR  collection year     xv List of abbreviations: data analysis ANOVA analysis of variance d.f.  degrees of freedom FDR  false discovery rate HSD  honestly significant difference CVA  canonical variate analysis n  sample size p  probability PAST  PAleontological STatistics PCA  principal components analysis TSBH  two-stage Benjamini & Hochberg step-up false discovery rate-controlling procedure    xvi Acknowledgements  There is no way to adequately thank those who have supported me through this endeavor. First I must profusely thank Dr. William Milsom for shepherding me through this research with such patience, wisdom and generosity. I could not have completed this project without him. Dr. Lance Barrett-Lennard showed great faith in me from the very beginning and I am so grateful for his talent, mentorship, and friendship. Dr. John Ford was also greatly supportive and helpful in launching this project. Dr. Robert Shadwick and Dr. Doug Altshuler served on my committee and helped me pull my ideas together, and I thank them for their time and insights. The wonderful Allyson Miscampbell identified many tricky bone samples at the Genetic Data Centre and I am so appreciative for her hard work and time. Dr. Altshuler and Dr. Roslyn Dakin provided helpful statistical advice. Shannon Obradovich was invaluable in my rookie R programming attempts. Doug Sandlilands kindly provided maps. Dr. William F. Perrin thoughtfully discussed killer whale skulls with me and introduced me to the treacherous waters of species delineation in cetaceans.   Rod Palm and Strawberry Isle Marine Research Society were the first to start measuring killer whale jaws in B.C. I thank them for their encouragement and for sharing their data.   This research was financially supported by the UBC Department of Zoology and the Killer Whale Adoption Program. I wish to acknowledge the contributors to KWAP whose generosity and passion are so tremendously valued by researchers and students.  Numerous people helped with data collection: Michael Etnier, Andrea Hunter, Jennifer Wang, Perry Poon, Natalie Brewster, Romney Porter, Scott Rogers, Nicola Dedeluk, Paul Malcolm, Lisa Cooper, Liliana Fajardo, Joanna Piercy, Vicky Lividitis, Susan Heaslip, Peter Schulze, Cheryl Talkington Altman, Andrea Rambeau, Melissa Webb, Mandy Wong, Amanda Bradford, Anne Mackie, and Lieneke Norman.   Charles Potter, James Mead, Dee Allen (USNMA), Bill Perrin (SWFSC), Moe Flannery, Douglas Long (CAS), Ruth Foster, Rod McVicar, Rex Kenner (UBC), Rod Palm (SIRS), Jim and Mary Borrowman, Jim Cosgrove, Nick Panter, Lesley Kennes, Gavin Hanke (RBCM), John Heyning, David Janiger (LACM), Paul Collins (SBMNH), Eileen Lacey (MVZ), Michael Etnier (NMML), Judith Eger, Susan Woodward (ROM), Alex Baillie (KHS), Perry Poon (VAMSC) and the Cetacean Research Lab all kindly provided access to killer whale material.    xvii  Science just cannot be done without a community and I was so fortunate to have ties at both UBC and the Vancouver Aquarium. In particular I thank the Cetacean Research Lab and the Milsom Lab members past and present, including but not exclusively: Valeria Vergara, Kathy Heise, Katie Kuker, Cara Lachmuth, Volker Deecke, Harald Yurk, Nancy Janzen, Michaela McDonald, Doug Sandilands, Judy McVeigh, Nicola Dedeluk, Nadine Pinnell, Nicola Brabyn, Caitlin Birdsall, Meghan McKillop-Moore, Stephen Raverty, Angie O'Neill, Lieneke Norman, Kim Borg, Colin Sanders, Glenn Tattersall, Beth Zimmer, Lisa Louth, Basia Gajda, Emily Coolidge, Joanna Piercy, Andrea Corcoran, Catalina Reyes, Stella Lee, Cosima Porteus, Graham Scott, Sarah Jenkin, Sabine Lague, Yvonne Dzal, Julia York, Angelina Fong, and Jessica Meir.   I am so lucky to have had the support of friends, classmates, and family including: Anne Dalziel, Heather Bears, Erin MacRae, Talitha Greenwood, Michael Brush, Gigi Lau, Jane Lee, Lili Yao, Shannon Obradovich, Katelyn Tovey, Erica Ross, Michelle Au, Emily Gallagher, Andrew Thompson, Dan Baker, Ken Savage, Angela Scott, Michael Sheriff, Lynn Norman, Marina Giacomin, Malcolm MacRae, Ernie Lin, Jen Wang, Rosy Wong, Sarane Poon, Kaneez Rehmanji, Manoji Wirasinghe, Jodi Pawluski, Meaghan MacNutt, Heather Young, Ashwin Usgaocar, the Harleys, the Mackies, the MacFarlanes, Jenny Hamilton Harding, the Weber-Hardings, the O'Neills, Greg Bole, Colin Stopper, Craig Anderson, Isabelle Mousseau, the Watson/Hodges family, Mary McAvoy and Michael Purvis, Bridget Milsom, the Woos, Chans, Chows, Tsangs, Laws, and Fungs. Thank you, all.   xviii Dedication       With love, I dedicate this to Mom and Dad, Thomas, Mildred, and Charlotte.1 Chapter 1: General introduction 1.1 Killer whale (Orcinus orca) diversification and taxonomic uncertainty The species Orcinus orca (Linnaeus, 1758) comprises many populations of killer whales in different ocean basins (Heyning and Dahlheim, 1988) that are behaviourally and ecologically diverse (e.g., Barrett-Lennard and Heise, 2006; LeDuc et al., 2008; Morin et al., 2010; Foote et al., 2011). The species is currently under taxonomic revision (Reeves et al., 2004; Committee on Taxonomy, 2016) in light of these field observations, as well as molecular evidence in support of multiple ecological divergences among populations from distant locations. For example, behavioural, acoustic and genetic evidence is accumulating to suggest that parallel patterns of ecological diversification have occurred in the north Atlantic (Deecke et al., 2011; Foote et al., 2016) and the Antarctic (Berzin and Vladimirov, 1983; LeDuc et al., 2008), which has raised considerable interest in formal recognition of new species or subspecies (Berzin and Vladimirov, 1983; Pitman and Ensor, 2003; Morin et al., 2010; Committee on Taxonomy, 2016). This thesis focuses on three northeast Pacific ecotypes.  1.2 Resident, transient (Bigg’s), and offshore killer whale ecotypes of the northeast Pacific Ocean Three northeast Pacific killer whale ecotypes comprise genetically distinct populations, which are identified by the strict dietary differences between them: the piscivorous resident, mammal-hunting transient (Bigg’s), and piscivorous shark-hunting offshore killer whales (Bigg et al., 1987; Saulitis et al., 2000; Ford and Fisher, 1982; Ford et al., 1998, 2011; Dahlheim et al., 2008).  Dr. Michael Bigg was the first to closely study killer whales on the coast of British Columbia, and meticulously document their natural history. Upon discovering and describing the remarkable behavioural differences between two groups he christened “resident” and “transient,” he surmised that due to their longevity and intelligence, it would be possible for behavioural and morphological differences to arise in the lineages even in sympatry, naming their social isolation as a reproductive barrier that would work just as effectively as geographic isolation (Ford, 2011). Even though foraging behaviours are culturally transmitted and maintained in populations, and are likely not genetic polymorphisms, these ecological specializations could serve to enforce the population 1   2 segregation and assortative mating patterns that are necessary conditions for ecological diversification and possible speciation (Barrett-Lennard and Heise, 2006; Foote et al., 2011). Molecular and genetic studies corroborate the behavioural evidence that the resident, transient (Bigg’s), and offshore killer whale populations are reproductively isolated despite the absence of any oceanographic or geographic barrier (Stevens et al., 1989; Barrett-Lennard, 2000; Hoelzel and Dover, 1991; Morin et al., 2010). Upon sequencing the entire mitochondrial genome, Morin et al. (2010) proposed that the genetic divergence between the sympatric resident and transient (Bigg’s) killer whales of the northeast Pacific was evidence for late stage speciation or even the completion of speciation. The evolutionary history of the offshore killer whales is less clear given conflicting phylogenies based on mitochondrial DNA versus microsatellite (nuclear) DNA (Barrett-Lennard, 2000), but one proposed scenario is that transient (Bigg’s) killer whales diverged early in the O. orca lineage, and much later on residents and offshore killer whales diverged. There was likely some backcrossing between offshores and transients (e.g., offshore females with transient males), but the level of gene flow between the groups is thought to be negligible (Barrett-Lennard, 2000; Morin et al., 2010). Regarding the nomenclature currently in use, the descriptors ‘resident’, ‘transient’, and ‘offshore,’ are gradually falling out of favour due to the impreciseness and potentially misleading terminology that alludes to geography and site fidelity but not the more informative ecotypic differences among the groups. However, the only group to receive a well-accepted revised colloquial name is the transient killer whales, which were recently recognized as the un-named subspecies “Eastern North Pacific transient killer whale/Bigg’s killer whale.” This un-named taxon is considered distinct from the “Eastern North Pacific resident killer whale” un-named subspecies. Formal work on the contentious species, subspecies, and population designations for the Orcinus taxon is ongoing (Krahn et al., 2004; Reeves et al., 2004; Foote et al., 2009; Foote, Morin, et al., 2011; Morin et al., 2010; Moura et al., 2014; Committee on Taxonomy, 2016).  1.2.1 Distribution and movement patterns of northeast Pacific ecotypes  In general the subpopulations of resident and transient (Bigg’s) killer whales occupy sympatric (spatially overlapping) ranges while the offshore killer whales are characterized as occupying parapatric (adjacent, but not necessarily overlapping) ranges relative to the other two groups (Ford and Ellis, 1999; Ford et al., 2000; Scheel et   3 al., 2001; Von Ziegesar et al., 1986; Matkin et al., 1997) (Figure 1 and Figure 2). Bigg’s original naming of the resident and transient populations was based upon initial observations of their apparent distributions, range fidelity, and timing (seasonality) of appearances in coastal British Columbia (Bigg et al., 1987). Later, offshore killer whales were encountered and named for their discovery near the continental shelf, off Haida Gwaii and the west coast of Vancouver Island, British Columbia (Dahlheim et al., 2008).   Figure 1. Map of the approximate ranges of northeast Pacific resident (orange, purple, pink and teal) and offshore (green) killer whale populations. Offshore killer whales are thought to occupy a mostly parapatric distribution relative to the more coastal populations of resident and transient (Bigg’s) killer whales (map provided by Doug Sandilands, BC Cetacean Sightings Network).   4  Figure 2. Map of the approximate ranges of northeast Pacific transient (Bigg’s) killer whale populations (black hash marks), shown relative to the resident (orange, purple, pink and teal) and offshore (green) killer whale ecotypes. Transient (Bigg’s) and resident killer whales are spatially sympatric over most of their known ranges (map provided by Doug Sandilands, BC Cetacean Sightings Network).  1.2.2 Behavioural differences between ecotypes Behavioural differences between the northeast Pacific killer whale ecotypes have been correlated with the marked divergence in hunting preferences and strategies (Baird et al., 1992; Baird and Whitehead, 2000; Riesch et al., 2012). These differences have been partially attributed to the behavioural and cognitive abilities of their respective prey because the killer whales must overcome vastly different anti-predator strategies to successfully feed (Baird et al., 1992; Barrett-Lennard et al., 1996; Deecke et al., 2002, 2005). For example, fish-hunting resident and offshore killer whales emit echolocation clicks to detect and capture their prey, whereas transient (Bigg’s) killer   5 whales tend to hunt quietly while emitting only sporadic, cryptic clicks to avoid alerting their acoustically sensitive and savvy mammalian prey (Barrett-Lennard et al., 1996; Deecke et al., 2002). Marine mammals, such as harbour seals, are able to detect and discern ecotype-specific repertoires, and are even capable of habituating to the sounds of harmless fish-eating resident killer whales while evading the calls of transient (Bigg’s) killer whales or other unfamiliar whales (Deecke et al., 2002). Also, group size and social structure in transient (Bigg’s) killer whales appear to be constrained by hunting strategy. While seeking mammalian prey, they tend to travel closer to shore in small, quiet, cryptic groups (Baird and Dill, 1995; Ford and Ellis, 1999; Baird and Whitehead, 2000). Pod composition in transient (Bigg’s) killer whales is therefore by necessity relatively fluid, and group fission is observed. It is not uncommon for transient (Bigg’s) killer whale offspring to disperse from their mothers, resulting in smaller hunting groups (Ford and Ellis, 1999; Baird and Whitehead, 2000). Also their social vocal activity is often reserved for feeding and post-feeding times when the mammalian prey is sufficiently incapacitated and their social noises will no longer affect hunting success by scaring away the prey (Deecke et al., 2005). In contrast, resident killer whales live in larger and more vocal groups: their pods are composed of strict matrilines where all male and female offspring associate closely with their mothers and maternal relatives for their entire lives (Bigg et al., 1987; Ford et al., 1998, 2000). Their frequent social vocal activity is important for maintaining pod cohesion (Ford, 1991), and salmon are unlikely to hear killer whale vocalization and echolocation activity (Au et al., 2004). Also in contrast to transient (Bigg’s) killer whales, offshore killer whales are found in very large, very vocal groups that call and echolocate frequently, an observation consistent with the finding that they specialize on fish prey that likely cannot detect them acoustically (Au et al., 2004; Ford et al., 2011).   1.2.3 Field observations of morphological differences in ecotypes So far, most morphological studies have been restricted to field observations from a distance, and a few observations of stranded animals. Small differences in dorsal fin shape and saddle-patch pigmentation have been described (Bigg et al., 1987; Baird and Stacey, 1988; Ford et al., 2000; Barrett-Lennard and Heise, 2006; Dahlheim et al., 2008). Only anecdotal field observations of differences in overall body size, or body shape have been noted between resident, transient (Bigg’s), and offshore killer whales. Transient (Bigg’s) killer whales are thought to be   6 larger on average than residents or offshores (e.g., Barrett-Lennard and Heise, 2006), and members of the offshore population are considered to be the smallest of the three (Ford et al., 2000; Dahlheim et al., 2008).   1.2.4 Dietary differences/trophic level differences among ecotypes There are dramatic diet differences among these populations; in fact, the behavioural and ecological differences associated with these dietary differences are so great that these ecotypes are described as occupying different trophic levels (Ylitalo et al., 2001; Krahn et al., 2007; Newsome et al., 2009) as quantified by stable isotope ratios of carbon and nitrogen, ratios of persistent organic pollutants, and fatty acid signatures/profiles (Herman et al., 2005; Krahn et al., 2007; Newsome et al., 2009). Many behavioural characteristics of each ecotype have been correlated with their diet differences that can be linked to the biomechanical and behavioural challenges of hunting, killing, and processing the different preferred prey types. In some vertebrates, the size and hardness of the diet correlates to skull size and shape (e.g., Adams and Rohlf, 2000; Santana et al., 2010).  1.2.4.1 Resident killer whales hunt salmonids The resident killer whale ecotype is known for specifically favoring salmon (Oncorhynchus spp.) among many possible marine fishes. Resident killer whales show a marked preference for Chinook salmon (Oncorhynchus tshawytscha) (Ford et al., 1998; Ford and Ellis, 2006; Hanson et al., 2010), which can attain masses greater than 60 kg (Healey, 1991; Quinn, 2005). Southern resident killer whales hunt mostly (relatively older and larger) four- to five-year-old Chinook salmon (Ford and Ellis, 2006). Resident killer whales echolocate to detect prey (Barrett-Lennard et al., 1996), chase individuals, and then ostensibly use snap-feeding to capture each fish, grasping them in their teeth and ripping them apart (Ford and Ellis, 2006). It is possible for killer whales to use suction feeding to swallow fish pieces (Young et al., 2012). Although salmon are captured individually, most resident killer whales usually divide each prey into smaller pieces to share with other pod members, even relatively small fishes, and so are unlikely to consume any fish whole. Most of the tissue is usually consumed, and all that usually remains after a feeding event are a few scales and an oily slick (Ford and Ellis, 2006). Dividing a salmon primarily involves biting skeletal muscle apart because the majority of the fish’s body mass is muscle and viscera that are bound within a thin layer of skin and thin bony   7 scales. The post-cranial skeleton is composed of relatively small and thin flexible bones, e.g., vertebrae, dorsal and ventral ribs, girdles, appendages and fin rays, which is in contrast to the anatomy of mammalian and shark prey that are hunted by the transient (Bigg’s) and offshore killer whales.   1.2.4.2 Offshore killer whales feed on fishes (e.g., Pacific sleeper shark) The offshore killer whales were not discovered until the late 1980s, but their frequent vocal activity and echolocation use suggest that they hunt animals that cannot discern these vocalizations (Dahlheim et al., 2008). Anecdotal observations and stomach contents support the hypothesis that offshore killer whales are piscivorous: predatory interactions with fishes such as Pacific halibut (Hippoglossus stenolepis) (Jones, 2006), salmon (Oncorhynchus spp.), sculpin (Cottus spp.) (Heise et al., 2003), blue sharks (Prionace glauca), and opah (Lampris guttatus) (Dahlheim et al., 2008) have been documented, but there are few direct observations of offshore killer whale foraging. Only the consumption of the livers of Pacific sleeper sharks (Somniosus pacificus) has been directly observed and confirmed by DNA analysis of prey remains found near actively feeding offshore killer whales (Ford et al., 2011). These are large sharks that can reach lengths of more than 7 m, but do not usually measure more than 3 m, and those sampled from Prince William Sound where the feeding events were observed ranged from 0.6 to 2.8 m (Hulbert et al. 2006 cited in Ford et al. 2011), which would range in mass from 15-215 kg (Sigler et al. 2006, cited in Ford et al. 2011).  Sharks have a cartilaginous endoskeleton that supports the large axial muscles. Their integument is thin but tough, extensively covered with abrasive placoid scales. Overall their morphology is more similar to salmonids, than it is to marine mammals. So in the context of this study, offshore killer whales will be considered piscivorous, with a diet that includes sharks.   1.2.4.3 Transient (Bigg’s) killer whales hunt marine mammals  The diet of transient (Bigg’s) killer whales is diverse in prey sizes and species composition, but they prefer hunting marine mammals. Transient (Bigg’s) killer whale prey range in size from sea birds, small porpoise calves and harbour seal (Phoca vitulina) pups (about 10 kg), to 1,600 kg sea lions (Eumatopius jubatus or Zalophus californianus) (Baird and Dill, 1995), and even large whales that are many times the mass of the killer whale. Not   8 many transient killer whales have been weighed, but the maximum recorded size was a 4,000 kg, 6.04 m male from an unknown population (Hoyt, 1981), which is much smaller than a 35 000 kg gray whale (Eschrictius robustus) (Bigg et al., 1987; Jefferson et al., 1991), though killer whales are more likely to target younger, smaller or otherwise more vulnerable prey.  Diet preferences differ depending on location and season (e.g., Saulitis et al., 2000; Maniscalco et al., 2007). In British Columbia, harbour seals, with masses up to approximately 130 kg, are the predominant prey (Baird and Dill, 1995; Ford and Ellis, 1999; Deecke et al., 2005).  Mammals have robust ossified endoskeletons such that the proportion of hard ossified tissues to soft tissues is higher than that of most fishes, and thus likely more difficult to tear apart or bite through than either teleost fishes or cartilaginous sharks. The soft tissues are encased in a thick blubber layer and skin and in most pinnipeds, the skin is also covered with fur.  Transient (Bigg’s) killer whales use their bodies as weapons: they ram prey repeatedly with their rostrum, slap them with tail flukes or flippers, breach upon them, or push the prey under the water to drown them (Bigg et al., 1987; Baird and Dill, 1995; Ford et al., 2005; Maniscalco et al., 2007). In addition to visual observations of these hunting events, collection of prey remains and audio recordings of these events verify the violence of these tactics. Transient (Bigg’s) killer whales have been observed cooperatively dismantling harbour seals in pairs. These data include descriptions of two whales holding opposite ends of a seal, backing away from each other, and also of pairs of whales swimming together while holding a seal between them, tearing it in two by moving their heads apart (Baird and Dill, 1995). Deecke (2002) deployed hydrophones to detect and identify transient (Bigg’s) killer whale foraging behaviours at a distance, and named the percussive sounds of predation Killing, Ramming and Crushing Sounds (KRaCS). These KRaCS suggest that osteophagy (the biting and consumption of bones) is indeed involved in hunting events. The biting of large prey and crushing of hard mammalian bones likely requires greater bite forces than those required for feeding on fish (Scott et al., 2012). Transient (Bigg’s) killer whales usually employ cooperative strategies to either ambush, herd, or chase prey to exhaustion, whereupon the prey succumb by drowning, or from injuries incurred during the attack (Baird and Dill, 1995; Ford and Ellis, 1999; Ford et al., 2005; Barrett-Lennard et al., 2011; Matkin et al., 2011). Thus, the agility, large size, and the greater overall toughness and   9 stiffness of mammalian prey likely increase the difficulty of prey handling, and the magnitude of bite forces required for feeding (Christiansen and Wroe, 2007; Scott et al., 2012).  Many marine mammals have predator-detection and evasion strategies. For example, harbour seals can use their maneuverability to evade transient (Bigg’s) killer whales, and swim towards near shore waters (Womble et al., 2007), and minke whales (Balaenoptera acutorostrata) can sometimes use their impressive speed to entirely out swim predators that are in pursuit (Ford et al., 2005). Some dolphins are able to ram the killer whales with the rostrum, while baleen whales can use flippers and flukes to deter them, e.g., humpback whales (Megaptera novaeangliae) (Pitman et al., 2016), or gray whales (Eschrictius robustus) (Barrett-Lennard et al., 2011). A marine mammal is capable of inflicting serious injury to the vulnerable eyes or soft skin of a killer whale if it were to scratch or bite the attacking predator in self-defense. Even upon being captured, large prey such as aggressive adult male sea lions can struggle and possibly harm the killer whale (Saulitis et al., 2000). Thus, it is likely important for the transient (Bigg’s) killer whales to minimize direct prey handling time, unless they are training calves to hunt, in which case prolonged harassment of prey has been observed where transient (Bigg’s) killer whales will catch, release, and recapture birds, pinnipeds, and otters, to either demonstrate or practice hunting techniques (Maniscalco et al., 2007). This supports the idea that these risky hunting skills must be transmitted and mastered at an early age (Saulitis et al., 2000).  1.2.5 Dentition differences between ecotypes O. orca are known for their long rows of large, conical, caniniform teeth, which curve postero-medially from the maxillae, and are laterally compressed at the base (Heyning and Dahlheim, 1988). Teeth in the alveolar dentary bones are also recurved but have a straighter profile than those from the maxillae (personal observation). Teeth from adults of each of the resident, transient (Bigg’s), and offshore ecotypes show evidence of wear – possibly due to tooth-to-tooth contact between opposing tooth rows (attrition), or from abrasion against prey. In most offshore killer whales, however, many teeth are so severely worn that the apices are completely removed down to the gingiva, leaving smooth, flattened stubs and exposed pulp cavities. This has been attributed to abrasion from shark integument, since the placoid scales of sharks each have enamel projections that are hard enough to destroy mammalian dentition (Ford et al., 2011). The tooth wear observed in resident and transient (Bigg’s) killer whales   10 also can be substantial, but spalling (where the tooth surface has failed and detached from the main tooth) occurs on different aspects of the teeth rather than mostly the apical region (personal observation), and they do not show nearly the same extent of apical wear, which is further evidence that their diets are likely different from that of offshore whales (Foote et al., 2009; Ford et al., 2011).   1.2.6 Inference of jaw function from simplified lever system and muscle attachment  Many aspects of skull morphology are linked to the biomechanical and behavioural challenges of hunting prey, as documented in many studies of vertebrates. As this was an exploratory study, the cranium and jaw were considered as a simple third-class lever system, where the lower jaw (the left and right dentary bones) pivots at the temporo-mandibular joints at the squamosal bones, and is elevated by jaw adductor muscles, including the masseter, temporalis and pterygoideus muscle complexes. Predictions and inferences based on this simplified lever system and mammalian jaw muscle attachments are explained in more detail as the results are discussed in Chapter 2. The tooth-bearing maxilla and dentary bones, the facial bones of the rostrum, and the bones of the temporal and vault regions were of particular interest due to their importance in prey capture and processing. Hyoid bones were not usually available, and so inferences pertaining to suction feeding were not made here.  1.3 Summary The remarkable pattern of numerous ecological divergences in O. orca has led to considerable debate regarding the phylogenetic history and taxonomic status of this group. Natural history and genetic data are accumulating quickly, and killer whales off the west coast of Canada and the United States have been meticulously studied, but the cranial morphology of the three ecotypes has not yet been characterized.  Therefore the goal of this thesis was to measure the skulls and jaws of three killer whales ecotypes from the northeast Pacific ocean, and to consider their morphology in the context of the strict ecotype-specific hunting strategies of mammal-eating transient (Bigg’s), salmon-specialist resident, and piscivorous/shark-hunting offshore killer whales. I assume that transient (Bigg’s) killer whales eating larger marine mammal prey with relatively massive endoskeletons practice osteophagy, whereas resident and offshore killer whales feeding on fishes primarily bite and tear through muscle and soft organ tissues and much less bone.   11 Chapter 2: Comparison of cranial shape among northeast Pacific killer whale ecotypes 2.1 Introduction  Detailed longitudinal studies of the killer whale ecotypes residing in the northeast Pacific have raised numerous practical and theoretical questions regarding their evolutionary history. Their wildly divergent diet preferences and cultural practices seem incongruent with their current status categorized as a single taxon, Orcinus orca, along with many other populations around the world (Heyning and Dahlheim, 1988; Berzin and Vladimirov, 1983; Pitman and Ensor, 2003; Krahn et al., 2004; Morin et al., 2010; Foote et al., 2009; Barrett-Lennard, 2011; Committee on Taxonomy, 2016). The resident, transient (Bigg’s) and offshore killer whale ecotypes that are included in this study are well-characterized, but rapidly growing research in other oceans have revealed other sympatric populations with morphological, dietary, and genetic diversification e.g., in the Antarctic (Pitman and Ensor, 2003; LeDuc et al., 2008) and north Atlantic oceans (Foote et al., 2009; Morin et al., 2010). Currently, the taxonomic status of O. orca is undergoing scrutiny and possible revision (Krahn et al., 2004; Reeves et al., 2004; Committee on Taxonomy, 2016). Baseline natural history data including diet, population structure, and behavioural observations have been illuminated with insights into their phylogenetic history based on genetic studies (e.g., Barrett-Lennard, 2000; Hoelzel et al., 2007; LeDuc et al., 2008; Foote et al., 2009; Morin et al., 2010; Moura et al., 2014). As a contribution, the goal of this research was to test the hypothesis that morphological differences in the cranium would be found in northeast Pacific resident, offshore and transient (Bigg’s) killer whale ecotypes, and if appropriate, to consider the results in the context of their diet specializations: the extreme preference of resident killer whales for salmon (primarily Chinook salmon), piscivory by offshore killer whales who feed on Pacific sleeper shark, and marine mammal hunting by transient (Bigg’s) killer whales. I hypothesized that transient (Bigg’s) whales have wider skulls, and more robust facial bones and dentary bones than the piscivorous resident killer whales that prefer hunting smaller salmonid prey, and the piscivorous offshore killer whales whose diet includes sharks.  Quantifying shape differences in complex articulated structures such as killer whale skulls is a challenge particularly in light of the statistical challenges of low sample size, and unequal representation from each   12 demographic. Here, traditional morphometrics methods were employed to most simply quantify the differences in shape of the cranial skeleton with linear measurements. Also, as an exploratory step, and in the interest of capturing subtle shape changes, landmark data were collected, and methods such as principal components analysis (PCA) and canonical variate analysis (CVA) were attempted to address the questions of whether there are particular regions of the cranial skeleton that contribute most to shape variance, and whether there are particular combinations of shape components that might best distinguish among the three groups.   2.2 Materials and methods 2.2.1 Measuring cranial morphology: geometric morphometrics and traditional morphometrics Geometric morphometrics has gained traction as a popular method to quantify and visualize shape and shape changes. When landmarks are collected, the two-dimensional (x,y) or three-dimensional (x,y,z) coordinates retain the geometric properties of shape information, while still allowing for statistical analysis of shape differences or shape changes. The graphical representation of the results is a powerful tool to visualize observations or experimental results in an intuitive way (Adams et al., 2004; Rohlf, 1998). One disadvantage, however, is that incomplete or damaged specimens cannot be included because one strict requirement is for all landmarks to be present on every specimen (e.g., Sztencel-Jabłonka et al., 2009). In this study I use geometric morphometrics to document the two-dimensional landmark data collected from standardized photographs of the cranium and dentary to identify the anatomical regions of greatest variation in northeast Pacific killer whale skulls. In addition, a traditional morphometrics approach was employed to most simply describe and compare the overall size and shape of skulls and jaws (e.g., Perrin, 1975). In traditional morphometrics, anatomy is characterized by measurements such as lengths, angles, counts, areas, or proportions. One disadvantage to traditional morphometrics is that there is autocorrelation among most of the measurements in each specimen (i.e., size very likely contributes to most of the measurements). Also, the resulting graphs provide less intuitive depictions of shape and shape differences, e.g., ratios, or principal components analysis (PCA) of many individual measurements (Rohlf and Marcus, 1993). A compelling advantage is that specimens with missing structures can be included, and therefore it is possible to include a larger number of specimens (e.g., Sztencel-Jabłonka et al., 2009). This is especially   13 important for studies with small sample sizes from species with small populations, which was the scenario in this study. Curated O. orca skulls are found only in relatively small collections dispersed across large distances, and are frequently damaged or fragmented resulting in the loss of many data points for geometric morphometrics methods.  2.2.2 Cranial skeleton anatomy  The nomenclature of structures and anatomical descriptions of the O. orca cranial skeleton (excluding the hyoids) (Figures 3-7) were adapted from delphinid literature with particular reference to Perrin's (1975) descriptions of spinner and spotted dolphins (Stenella spp.) and Mead and Fordyce (2009), whose work on bottlenose dolphin (Tursiops spp.) was consulted to clarify confusing terminology, especially where different names have been used to identify homologous structures in odontocetes (the toothed whales). These two monographs are based on smaller delphinid species, but the differences between these taxa and Orcinus were noted (Mead & Fordyce, 2009).   To illustrate the intraspecific variation among the skulls and dentary bones of O. orca, some examples of specimens of each northeast Pacific ecotype and sex are displayed in Appendix F.      14                          Figure 3. Photograph and illustration of adult male resident killer whale (O. orca), left view of skull (RBCM8386). Bones and bony features are indicated as follows: 1 maxilla, 2 premaxilla, 3 frontal, 4 parietal, 5 squamosal,  6 preorbital process of lacrimal, 7 jugal, 8 pterygoid, 9 pterygoid sinus, 10 temporal fossa, 11 temporal crest,  12 occipital condyle, 13 exoccipital.     15                                     Figure 4. Photograph and illustration of adult male resident killer whale (O. orca), dorsal view of skull (RBCM8386). Bones and bony features are indicated as follows: 1 mesorostral canal, 2 premaxilla, 3 maxilla,  4 vomer, 5 premaxillary foramen, 6 antorbital notch, 7 dorsal infraorbital foramen, 8 external bony nares, 9 nasal fossa, 10 frontomaxillary suture, 11 squamosal, 12 frontal suture, 13 supraoccipital, 14 exoccipital, 15 nuchal crest, 16 temporal crest, 17 frontal, 18 nasal, 19 ethmoid, 20 nasal septum, 21 maxillary crest, 22 rostrum.   16           Figure 5. Photograph and illustration of adult male resident killer whale (O. orca), ventral view of skull (RBCM8386). Bones and bony features are indicated as follows: 1 palatine process of premaxilla, 2 palatine process of maxilla, 3 palatal vomer, 4 palatal surface, 5 jugal, 6 ventrolateral crest, 7 intercondyloid notch, 8 basioccipital crest, 9 paraoccipital process,10 mandibular fossa, 11 pterygoid hamulus, 12 lacrimomaxillary fossa, 13 alveolar groove, 14 vomerine crest/nasal septum.   17   Figure 6. Photograph and illustration of adult male resident killer whale (O. orca), occipital view of skull (RBCM8386). Bones and bony features are indicated as follows: 1 temporal crest, 2 interparietal, 3 supraoccipital,  4 squamosal, 5 occipital condyle, 6 basioccipital, 7 foramen magnum, 8 intercondyloid notch, 9 basion,  10 paroccipital process, 11 exoccipital.      18 A  B    Figure 7. Photograph and illustration of adult male resident killer whale (O. orca), (A) labial view and (B) lingual view of left dentary bone (RBCM8386). Bony features are indicated as follows: 1 coronoid process, 2 mandibular condyle, 3 angular process, 4 ventral margin, 5 gnathion, 6 pogonion, 7 alveolar groove, 8 mandibular fossa,  9 mandibular foramen, 10 symphyseal surface.  2.2.3 Materials 2.2.3.1 Specimens Data from 84 specimens originating from populations from Alaska, British Columbia, Washington State, California, Mexico, Russia, Baffin Island, and the north Atlantic were collected. For this study 70 animals of known ecotype, North Pacific resident, transient (Bigg’s), or offshore, were included in the analysis: 38 belonging to the resident ecotype (16 males, 11 females, 11 unknown sex); 20 belong to the transient ecotype (11 males, 5 females, 4 unknown sex); and 12 belong to the offshore ecotype (3 males, 5 females, 4 unknown sex) (see Appendix A for specimen details). 14 skulls were measured and examined for reference purposes, but excluded from analysis because they were not assigned to a northeast Pacific ecotype. The geographic distributions of the animals for which collection locations are known are mapped in Figures 8 and 9. Depending on availability and completeness of the structures, data were collected from the cranium, or left dentary, or both. Total body length measurements (from caudal fin to tip of rostrum) were also recorded from accession data and necropsy reports and compared across ecotypes.    19 2.2.3.2 Ecotype assignment Where possible, to assign ecotype, genetic identifications of specimens were determined from published literature (e.g., Morin et al., 2006; Newsome et al., 2009) and corroborated with information from museum accession records. Any contradictions between different data sets were noted (e.g., for sex, or ecotype diagnosis). Consensus among field identification data (e.g., photographic identification of individuals from censused/catalogued populations, stomach contents, tooth wear patterns, acoustic identifications) and genetic data were considered strong evidence for ecotype identification. If genetic data was not available, ecotype was inferred from field observations or necropsy data. For specimens where genetic sampling was required, permission for destructive sampling was acquired from the institution. A power drill and sterilized drill bits were used to collect small amounts of bone, either from the occipital condyles or other intact regions of the skull, or from teeth (e.g., if the condyles were too difficult to sample due to the heat generated by friction from drilling into very hard bone). The initial bone fragments removed from the exterior surface of the specimen were discarded in favour of deeper layers of bone less likely to be contaminated. Bone shavings were collected in sterile aluminum foil, and sealed in small centrifuge tubes.  Mitochondrial DNA extraction and sequencing were completed in ancient DNA laboratories by collaborators who determined ecotype membership based on methods established by Barrett-Lennard (2000) and Morin et al., (2006). Sequences and population identification results were sent to museum curators where required as a condition for destructive sampling permissions (see Appendix A.1 for haplotype and ecotype assignments).   20  Figure 8. Collection locations of specimens from the northeast Pacific (Alaska, British Columbia, and Washington State) with known coordinates were plotted with GPSVisualizer.com; square = male; circle = female; diamond = unknown sex; red = transient (Bigg’s); blue = resident; green = offshore.      21  Figure 9. Collection locations of specimens from the northeast Pacific (California and Mexico) with known coordinates were plotted with GPSVisualizer.com; square = male; circle = female; diamond = unknown sex; red = transient (Bigg’s); blue = resident; green = offshore   22 2.2.4 Traditional morphometrics measurements 2.2.4.1 Linear measurements  Linear measurements were collected from the cranium and the left dentary bone (Figures 10-13). The measurements chosen were adapted from Perrin (1975) and selected if they could be readily identifiable on the killer whale skull and measured with available instruments with reasonable repeatability. At the time of the study they were also selected to best describe overall shape and size, and also as reference measurements for the standardized photographs that could be used if more detailed comparisons of finer-scaled structures was necessary. Most importantly, since the research question pertains to the ecotypic differences between the populations, particular interest was paid to structures involved in feeding, i.e., the tooth bearing maxilla and dentaries, the facial bones that form the rostrum of the skull (premaxilla and maxilla), and the temporal region of the skull, which forms the site of origin of the jaw adductor muscles and also the space through which these large muscles pass to insert on the proximal dentary bone. Other structures, e.g., the foramen magnum, occipital condyles and respiratory structure (the nares) were included as regions that would not be expected to exhibit much plasticity or evidence of selection. Scant linear measurements of the dentaries were taken because there was a great deal of variation in shape and it was difficult to ensure consistency in measurements (e.g., a measurement of the “greatest depth of jaw” varied in location along the longitudinal axis of the jaw and was not strictly repeatable), but also because it was more efficient to take reliable standardized photographs of these smaller structures (in contrast to the much larger, variable crania). Total body length measurements were recorded from necropsy reports and/or museum accession records and converted to centimetres. Each cranial and dentary measurement was taken a minimum of three times, to the closest 0.1 cm (straight ruler, meter stick, steel measuring tape, inside calipers, or 60 cm calipers), or closest 0.001 cm with the digital calipers. The values used for the analyses are the mean of the measurements.   2.2.4.2 Traditional morphometrics - analysis Each measurement from the cranium and dentary was divided by condylobasal length (SKL), so that each measurement expresses shape as a ratio. One-way analysis of variance (ANOVA) was performed (JMP 9.0.2) to determine the effect of ecotype on each shape or measurement. Groups were divided by Northeastern Pacific   23 ecotype: resident, transient (Bigg's), or offshore. Sex determination was not possible for a large proportion of the samples, and for those ecotypes that included animals known to be male or female, the sample sizes were too small and too unequal to conduct appropriate two-way ANOVAs for each ecotype and sex. Therefore, for each ecotype group, males and females were pooled with individuals of unknown sex. To assess the assumptions underlying the use of ANOVA, Shapiro-Wilk tests were performed on residuals to test for normal distribution, and the Levene’s test was conducted to determine whether variance was homogeneous (JMP 9.0.2). If assumptions of normality or homogeneity of variance were violated, a transformation was attempted, or a non-parametric alternative test (Kruskal-Wallis) was conducted. After these analyses, the two-stage Benjamini & Hochberg (TSBH) step-up false discovery rate (FDR)-controlling procedure was conducted (Benjamini et al., 2006) with the R package multtest (Pollard et.al., 2005) to control for a false discovery rate (FDR) of 0.02. The TSBH was chosen among many possible family-wise error rate and FDR-controlling methods because it appeared to have the most balanced and reasonable effect on power, Type I, and Type II errors for a study of this size (Roslyn Dakin, pers. comm.). If there were any differences found among the ecotypes, a post hoc Tukey Honestly Significant Difference test (Tukey HSD) was performed to determine the pairwise differences between the ecotypes for each measurement; significance was accepted at p<0.05. The body length (TLF) measurements were gleaned from accession and necropsy records or literature, and condylobasal lengths (SKL) were available for animals not measured during this study. Therefore, TLF and SKL were analyzed separately, and statistical significance was accepted at p<0.05 in these instances.     24 A      B       C        Figure 10. Measurements from the dorsal (A), ventral (B), and left lateral (C) views of cranium. The anterior of the skull is to the left of the page. (A) SKL (condylobasal length, cm); POS (greatest postorbital width measured at squamosals, cm); WAN (width of rostrum at antorbital notch, cm); WRP (greatest width of rostrum anterior to antorbital notch; from processes, cm); (B) WAP (greatest width at antorbital process, cm); WAL (distance between last two alveoli in maxillae, cm); LMT (length of upper left tooth row, the maxillary alveolar groove, cm); (C) LPF (greatest length left temporal fossa, cm); WPF (greatest width left temporal fossa, cm).   25   Figure 11. Measurements from the occipital view of cranium. OCW (maximum width of occipital view of cranium, cm); OCH (total height cranium occipital view, cm); WOC (width across occipital condyles, cm); FMW (width foramen magnum, cm); FMH (height foramen magnum, cm); LOC (greatest length left occipital condyle, cm).  A  B    Figure 12. Measurements from the labial view (A) and the lingual view (B) of the left dentary. AMN (length lower left tooth row, the mandibulat alveolar groove, cm); LLR (greatest length left dentary, cm); LSY (length mandibular symphysis, cm); LLF (length left mandibular fossa, cm).    26  Figure 13. The left postero-dorsal view of this skull shows the temporal fenestre through which the temporalis muscles pass. Linear measurements of this site included MAD (major diameter of left temporal fenestre, cm), and MID (minor diameter of left temporal fenestre, cm) as measured with spreading (inside) calipers.   2.2.5 Geometric morphometrics 2.2.5.1 Standardized photography protocol Standardized photographs were taken with a digital camera (Nikon Coolpix 995). A spirit level was placed on the appropriate surfaces on the camera to ensure consistent lens position (e.g., parallel to the floor and the camera lens for photos of the cranium dorsal view, cranium ventral view, dentary labial view, and dentary lingual view). A scale bar was placed next to each specimen, e.g., on the same plane as the rostrum for the dorsal view of skull, or on same plane as the occipital condyles for occipital and lateral views. A black fabric backdrop was used where MADMID  27 possible to increase the contrast between the specimen and the background. Photos were saved as *.tiff files and converted to *.jpg files for use with the TPS software suite for landmark digitization (Rohlf, 2010). For photographs of the labial side of the dentary, the flat lingual surface of the dentary rested on a horizontal surface (table, platform or floor), the scale bar was placed on the supporting surface, and the camera lens was positioned horizontally, parallel to the supporting surface. A spirit level was placed on the camera body parallel to the lens, to judge the placement of the camera lens. Photographs of the lingual view of the mandibles were taken with the mandible positioned with the lingual plane as horizontal as possible as indicated with a level, with the edges of the coronoid and angular processes on the same horizontal plane if possible. When the left and right dentaries were attached together during preparation such that they could not be separated and photographed according to the protocol, the dentaries were photographed with the scale bar in the frame and with the camera lens positioned as close to parallel to the lingual plane of the dentary as possible. Photographs of the dorsal, left and occipital views of the skull were taken with the cranium resting on the ventral edges of the basioccipitals and the rostrum positioned such that the distal tip (formed by the premaxillae) was aligned on the horizontal plane with the occipital condyles, to approximate the condylobasal axis, which was defined as when the distal tip of the rostrum and the occipital condyles were aligned on the same horizontal plane as closely as possible. Photographs of the dorsal view and the ventral view of the skull were taken with the camera lens parallel to the condylobasal axis of the skull. The camera was positioned such that the antorbital notches aligned with the midline of the exposure and focus guidelines in the liquid crystal display (LCD) of the camera. The scale bar was placed at the same height as the condylobasal axis.  For photographs of the occipital view, the lens was aligned with the occipital condyles and the tip of the rostrum such that the condylobasal axis was perpendicular to the camera lens. The scale bar was placed perpendicular to the condylobasal axis at the occipital condyles.  For the left lateral view, photographs were taken from left side of animal at a distance, the lens positioned parallel to condylobasal axis and level with the antorbital notches (as measured in height from the floor or table and the camera positioned on a tripod). The scale bar was placed at the condylobasal axis under the skull if room permitted, or either anterior or posterior to the skull. Here it was expected that parallax errors would be introduced,   28 but it was anticipated that parameters such as the measured length of skull (condylobasal length) could be used for scale in the photographs when necessary, especially in the case of very large skulls.  2.2.5.2 Landmark digitization from standardized photographs and analyses of landmark coordinates Cetaceans have directionally asymmetrical skulls. O. orca skulls are relatively less asymmetrical in comparison to other cetaceans (Heyning and Dahlheim, 1988; Mead and Fordyce, 2009). Asymmetry can be observed in the nasal sutures and the telescoping of the facial bones (maxillae and premaxillae) over the vault of the cranium. Landmarks were collected from the left side and “midline” of each skull (Figure 14) and from the labial side of each left dentary bone (Figure 15).  The digital photographs were managed with tpsUtil, and tpsDig2 (Rolhf, 2010) was used to digitize the landmarks and scale bar for each cranium and dentary bone. The (x,y) coordinates were analyzed with the PAST (PAleontological STatistics, Hammer et al., 2001) and MorphoJ (Klingenberg, 2011) software packages.   2.2.5.3 Geometric morphometrics - analysis Centroid size was calculated, and for shape analysis, Procrustes fitting or superimposition was performed to translate the centroids to a single origin, scale the landmarks to the same centroid size, and to rotate each shape about the centroid. This manipulation allows for direct comparison of shapes among ecotypes (Rohlf and Slice 1990). Principal components analysis (PCA) was conducted to summarize which aspects of cranial shape contribute the most to differences among the ecotypes. Canonical variate analysis (CVA) was conducted on the Procrustes superimposed landmarks to calculate which shape features could be used to distinguish among the resident, transient and offshore ecotypes (Klingenberg, 2011) in a manner similar to linear discriminant function analysis.       29  Figure 14. Twenty-one landmarks were digitized from standardized photographs of the skull (dorsal view), along with the dimensions of a scale bar placed at the same height as the condylobasal axis.   Figure 15. Fourteen landmarks were digitized from standardized photographs of each left dentary, (labial view) along with the dimensions of a scale bar placed within the frame of each photograph on the same plane as the lingual surface.    30 2.3 Results Each boxplot in the following section (Figure 16 - Figure 32) indicates the median (line), the first and third quartiles (the bounds of the box), the 95% confidence interval (whiskers), and the outliers (open circles). Means are indicated as values and plotted as black dots. Different lower case letters indicate significant differences between ecotype groups as calculated from post hoc Tukey HSD tests. Where p-values are reported, statistically significant values are denoted with an asterisks (*).  2.3.1 Total body length and condylobasal length measurements  The mean body lengths of the transient (Bigg’s) were significantly longer than resident killer whales. The offshore killer whales did not have significantly different mean body lengths from either the resident or transient (Bigg’s) killer whales (Figure 16). The same was true of the condylobasal lengths (Figure 17).  2.3.2 Results from traditional morphometrics of cranium and dentary bone One-way ANOVA was conducted for 21 measurements of the cranium and dentary to determine whether there are differences in skull shapes among resident, offshore or transient (Bigg’s) ecotypes. A two-stage Benjamini & Hochberg step-up false discovery rate-controlling procedure (Benjamini et al., 2006) was conducted to control for a false discovery rate of 0.02. This determined that results were considered significant if p ≤ 0.01049, which is the p-value at which the corresponding 'adjusted p-values' are less than 0.02 (the false discovery rate cut-off value). All raw and adjusted p-values are reported in Appendix C. In the text of this thesis, the raw p-values are stated, and are marked with an asterisk (*) if they are considered statistically significant.    31  Figure 16. Boxplot showing the effect of ecotype on the total body length, TLF [F(2,42)=3.6277, p=0.035*]. The post hoc comparison (Tukey HSD) shows that the body length of transient (Bigg’s) killer whales is greater than that of residents (p=0.0427*) and that offshores were no different in length from residents nor transients (p>0.05).   32  Figure 17. Boxplot showing the effect of ecotype on condylobasal length, SKL [F(2,45)= 3.6306, p=0.03298*]. Post hoc comparison (Tukey HSD) shows that transients (Bigg’s) killer whales have a longer skull than residents (p=0.0352*) and that offshores do not have a different length than either residents or transients (p>0.05).     33 2.3.2.1 Overall skull width and height across the vault and occipital regions  Relative cranial width was measured in two ways: as the greatest postorbital width of the cranium at the squamosals (POS, Figure 18), and as the maximum width of the cranium from the occipital view across the exoccipitals (OCW, Figure 19). Both measurements were significantly different among the three ecotypes, and post hoc tests indicated that the shape of transient (Bigg’s) killer whales skulls is on average about 6% broader than the resident killer whales (p<0.01049*). The widths of offshore killer whale skulls are intermediate to, but not significantly different from the residents or transients (Bigg’s). In contrast, the overall height of the cranium as measured from the occipital view (OCH/SKL) was not significantly different among the ecotypes (p>0.01049, Figure 20).  2.3.2.2 Rostrum (facial bones and the palate)   Of four measurements related to the width of the rostrum, three showed small but significant differences among the three ecotypes, whereas one did not show any effect related to ecotype. WAP is a measurement taken across the anterior margins of the antorbital processes of the maxillae (Figure 21), and did not show a relationship with ecotype (p=0.09274). In contrast, the relative maximum rostral width (WRP/SKL, Figure 22) depends on ecotype (p=0.0086*). Offshore killer whales have a broader rostrum compared to resident killer whales (p=0.0120*), whereas there was no difference between transient killer whales and either offshore or resident killer whales (p>0.05). The relative rostral width measured at the antorbital notches (WAN/SKL) did depend on ecotype (p=0.0016*, Figure 23). For this measurement transient are broader than resident killer whales (p=0.0012*), whereas offshores are not different from either transients or resident killer whales (p>0.05). Residents have a narrower palate (WAL/SKL) than both the transients (p=0.0019*) and the offshore killer whales (p=0.0312*, Figure 24).  2.3.2.3 Bones associated with jaw adductor muscles (temporalis)  The temporal fossa (site of temporalis muscle complex origin attachment) was characterized by measuring the greatest length (LPF) and width (WPF) of the space bounded by the temporal crest, temporal arch, parietal bone, and squamosal bone (Figure 10). The temporal fenestre, the space through which the temporalis muscles pass to insert upon the proximal end of the dentary bones, was also measured (MAD and MID, Figure 13). No differences were found among the ecotypes for any of these measurements (p>0.01049, Figures 25 and 26).  34   Figure 18. Boxplot showing the effect of ecotype on the relative post-orbital width of the cranium, POS/SKL [F(2,45)=8.3011, p=0.0009*]. The post hoc comparison (Tukey HSD) shows that the overall skull width of transient (Bigg’s) killer whales is greater than that of residents (p=0.0010*). Offshore killer whales were not significantly different from either residents or transients (Bigg's) killer whales (p>0.05).   35   Figure 19. Boxplot of the relative width of the cranium as measured on the exoccipitals for each ecotype, OCW/SKL [F(2,38)=9.2161, p=0.0005*]. The post hoc comparison (Tukey HSD) shows that transient (Bigg’s) killer whales have wider occipital bones than residents (p=0.0003*) and that offshores are not different from either residents or transients (Bigg’s) (p>0.05).     36   Figure 20. Boxplot showing no difference among ecotypes for measurements of the relative height of the cranium, OCH/SKL [F(2,39)=3.1063, p=0.0560].     37  Figure 21. Boxplot of the relative distance between the anterior-most edges of the antorbital processes for each ecotype, WAP/SKL [F(2,41)=2.5213, p=0.0927].    38  Figure 22. Boxplot showing the effect of ecotype on the greatest relative width of the rostrum, WRP/SKL [F(2,43)=5.3277, p=0.0086*]. Post hoc tests showed that offshore killer whales have a broader rostrum compared to resident killer whales (p=0.0120*), and that transient (Bigg's) killer whales are not significantly different from either residents or offshores (p>0.05).    39  Figure 23. The relative width of the rostrum at the antorbital notches was significantly different among the ecotypes, WAN/SKL [F(2,30)=8.0432, p=0.0016*]. The rostrum is wider in transients compared to residents (p=0.0012*).   40  Figure 24. Ecotype had an effect on the relative width of the palate, WAL/SKL [F(2,42)=7.8831, p=0.0012*]. Residents have a narrower palate than both the transient (p=0.0019*) and offshore killer whales (p=0.0312*).   41 A         B   Figure 25. Boxplots showing that measurements of the (A) relative length, LPF/SKL [F(2,42)=4.1595, p=0.0225] and (B) relative width, WPF/SKL [F(2,41)=1.7537, p=0.1859] of the temporal fossae were similar across all ecotypes.    42 A         B  Figure 26. Boxplots showing that measurements of the relative size of the temporal fenestrae were not significantly different among ecotypes: neither (A) the relative major diameter of the left temporal fenestre, MAD/SKL [F(2,41)=3.1198, p=0.0548], nor (B) the relative minor diameter of left temporal fenestre, MID/SKL [F(2,41)=0.3746, p=0.6899] showed a significant difference.     43 A         B                        Figure 27. (A) Boxplot showing the significant relationship between ecotype and the relative length of the upper left tooth row, LMT/SKL [F(2,43)=10.1176, p=0.0003*], where offshore killer whales have a longer relative LMT compared to resident (p=0.0002*) and transient (Bigg's) killer whales (p=0.0472*). (B) The relationship between ecotype and the relative length of the dentary tooth row (AMN/SKL) is significant [F (2,36)=6.9992, p=0.0027*], where the mean in resident killer whales is shorter compared to both offshore killer whales (p=0.0075*) and transient killer whales (p=0.0249*).  44 2.3.2.4 Length of tooth-bearing alveolar bones on maxilla and dentary  On the maxilla, the relative length of the tooth-bearing alveolar bone (tooth row) (LMT/SKL) differed depending on ecotype (p=0.0003*, Figure 27 A). Specifically, it was longer in offshore killer whales than both the resident (p=0.0002*), and transient (Bigg’s) killer whales (p=0.0472*). The relationship between ecotype and the relative length of the dentary tooth row (AMN/SKL) is significant [F(2,36)=6.9992, p=0.0027*, Figure 27 B], where it is shorter in resident killer whales compared to both offshore killer whales (p=0.0075*) and transient killer whales (p=0.0249*).  2.3.2.5 Dentary bone shape  For the linear measurements that characterize overall size and shape of the dentary bone, the relative length of the dentary bone (LLR/SKL, Figure 28) was found to be longer in transient killer whales compared to resident killer whales (p=0.0075*), but offshore killer whales are not significantly different from either residents or transient killer whales (p>0.05). Neither the relative length of the mandibular symphysis (LSY/SKL, Figure 29 A), nor the relative size of the opening to the mandibular foramen (LLF/SKl, Figure 29 B), were significantly different among the ecotypes.  2.3.2.6 The foramen magnum, occipital condyles, and internal nares  All the measurements for the foramen magnum and the occipital condyles were similar among all the ecotypes. Neither foramen magnum width (FMW) nor foramen magnum height (FMH) were different, (Figure 30), and also the length and width of the occipital condyles (LOC and WOC) were similar across ecotypes (Figure 31). Finally, the relative width of the internal nares (WIN) was not different among the three ecotypes (Figure 32).   45  Figure 28. Boxplot showing that the relative length of the left dentary (LLR/SKL) depends on ecotype [F(2,36)=5.1864, p=0.0105*]. They are significantly longer in transient killer whales compared to resident killer whales (p=0.0075*) but offshore killer whales are not significantly different from either residents or transient killer whales (p>0.05).  46 A         B                 Figure 29. Boxplots showing that (A) the relative length of the mandibular symphysis, LSY/SKL [F(2,36)=3.6298, p=0.0366], and (B) the relative length of the left mandibular fossa, LLF/SKL [F(2,36)=0.2663, p=0.7677] are not significantly different among the ecotypes.    47 A         B                  Figure 30. Boxplots showing that the measurements for (A) the relative foramen magnum width, FMW/SKL [F(2,38)=0.1871, p=0.8301] and (B) relative foramen magnum height, FMH/SKL [F(2,37)=0.1758, p=0.8394] all showed no differences in shape among ecotypes.     48 A           B                  Figure 31. Boxplots showing that (A) measurements of the relative length of the left occipital condyle (LOC/SKL) did not vary with ecotype (Kruskal-Wallis, d.f.=2, Chi-squared=2.1431, p=0.3425). (B) The relative width across both occipital condyles (WOC/SKL) did not significantly differ among ecotypes [one-way ANOVA, F(2,38)=3.9207, p=0.0283].    49   Figure 32. Boxplots showing there was no significant difference among ecotypes in the relative width of the internal nares, WIN/SKL [F(2,40)=2.5209, p=0.0931].     50 2.3.3 Results from geometric morphometrics For landmark data collected from the dorsal view of the cranium and from the left labial view of the dentary, Procrustes fitting was performed to scale, translate and rotate the landmarks to directly compare shape. Alignment of the Procrustes fitting was along the primary axis of the mean configuration (Klingenberg, 2011). Principal components analysis (PCA) was then conducted on the Procrustes coordinates to identify regions that contribute the greatest amount of shape variation of the cranium. Canonical variate analysis (CVA) was conducted to find the anatomical regions that might best distinguish between the three killer ecotypes.  2.3.3.1 Cranium, dorsal view 2.3.3.1.1 PCA of Procrustes coordinates, dorsal view of cranium Twenty-one landmark coordinates were collected from photographs of skulls of 6 offshore, 17 resident, and 13 transient (Bigg’s) killer whale crania (total n=36). A shape PCA was conducted on the Procrustes transformed coordinates. The first four principal components calculated from the Procrustes coordinates of the dorsal skull contributed together 59% of the variance in dorsal skull shape (Figure 33; see Appendix D1 for a table of eigenvalues and percent variance). To visualize the shape changes observed between the resident, offshore and transient killer whale ecotypes, wireframe diagrams were drawn to represent the shape changes that occur along the first and second principal components (PC1 and PC2, Figure 34), and the third and fourth principal components (PC3 and PC4, Figure 35), and the corresponding PC scores are plotted for each specimen in the scatter plots. These results show that the ecotypes overlap in the shape space. However, there is some degree of separation between the transient and resident killer whales, and the data suggest that the offshores exhibit somewhat more variability in shape and also overlap with each of the resident and transient shape space, which is consistent with the results from the traditional morphometrics analyses. The wireframe graphs (Figure 34) depict variations in the vault and occipital regions of the skull that contribute to PC1, where at the positive extreme of PC1, the size of the braincase appears larger relative to the rostrum. For PC2 there is a relative narrowing of the entire skull, primarily across the braincase region, and also the distal regions of the maxillae and premaxillae.   51  Figure 33. The percentage contribution to the variance in shape from each principal component, for the shape PCA of the dorsal view of the skull. The first four principal components account for 59% of the variance.   05101520255 10 15 20 25 30 35principal components% variance  52   Figure 34. Scatterplot of PCA 1 and PCA 2 scores for Procrustes coordinates from the dorsal view of the skull, and 90% confidence ellipses for the mean PC scores of each ecotype. Ecotype membership is indicated by colour (offshore = green, resident = blue, transient = orange). The wireframe graphs show the shape changes that are associated with PC 1 (22.2%) and PC 2 (15.9%). The light blue wireframe indicates the reference shape, the dark blue wireframe represents the target shape; scale factor 10.0. The numbered dark blue dots and light blue open circles correspond to the 21 landmarks on the left side of the skull from the dorsal view. The anterior (or distal) tip of the rostrum is to the left.    53   Figure 35. Scatterplot of PCA 3 and PCA 4 scores for Procrustes coordinates from the dorsal skull, and 90% confidence ellipses for the mean PC scores of each ecotype. Ecotype membership is indicated by colour (offshore = green, resident = blue, transient = orange). The wireframe graphs show the shape changes associated with PC 3 (11.3% variance) and PC 4 (9.7%). The light blue wireframe indicates the reference shape, the dark blue wireframe represents the target shape; scale factor 10.0. The numbered dark blue dots and light blue open circles correspond to the 21 landmarks on the left side of the skull from the dorsal view. The anterior (or distal) tip of the rostrum is to the left.   54 2.3.3.1.2 CVA of Procrustes coordinates, dorsal view of cranium Canonical variate analysis (CVA) was conducted on the cranial landmarks. CVA gives an indication of which anatomical regions and combinations of shape variables might best distinguish between groups by maximizing between-group means (relative to within-group variation). Here, the ordination of Procrustes superimposed shape variables that results in the best separation between the group means includes variations in the vault, occipital region and overall width of the cranium (CV1 and CV2), and changes in shape of the distal end of the rostrum (CV2) (Figure 36). Mahalanobis distances (representing statistical distance among groups, not morphological distance) were calculated, and pairwise distances among the ecotypes (resident – offshore, transient – offshore, and resident - transient) were all significant (p<0.0001, calculated from 10000 rounds of permutations). Procrustes distances (indicating a difference in shape) were calculated among groups, showing support for morphological difference between residents and transient killer whales (p=0.0001*) and between residents and offshores (p=0.0309*), but not between offshores and transient killer whales (p=0.1632). In each test, p was calculated from 10000 rounds of permutations.      55   Figure 36. This scatter plot of CV 1 and CV 2 values for skulls of individuals from each ecotype (offshore = green, resident = blue, transient = orange) is the ordination of shape variables that best separate the ecotypes; percent variance is stated in parentheses. The wireframe graphs indicate the shape changes associated with CV1 and CV2. The light blue wireframe represents the reference shape; the dark blue wireframe represents the target shape (scale factor is 20.0). The numbered dark blue dots and light blue open circles correspond to the position of 21 landmarks taken from the left side of the dorsal skull.   56 2.3.3.2 Left dentary, labial view 2.3.3.2.1 PCA of Procrustes coordinates, dentary bone Fourteen landmark coordinates were collected from photographs of the labial view of the dentary bones of 21 resident, 12 transient (Bigg’s), and 8 offshore killer whales (total n=41). A shape PCA was conducted on the Procrustes transformed coordinates. The first four principal components account for 74.3% of the variation in dentary shape (Figure 37; see Appendix E.1 for a table of eigenvalues and percentage variance). In contrast to the traditional morphometrics shape comparisons that indicated that only relative jaw length varied among the three ecotypes, the geometric morphometric analysis of the landmark coordinates on the dentaries indicates that there are regions of shape variation that invite further examination. Scatterplots of PCA scores and the corresponding shape changes were plotted for PC1 and PC2 (Figure 38), and PC 3 and PC 4 (Figure 39). The PCA of the landmark data from the dentaries indicate that there is overlap in shape space among the ecotypes, however the shape of the proximal end of the dentary as well as the overall shape of the ramus are important contributors to overall shape variation. The most important observations from this analysis are that: at the positive extreme of PC1 there is an overall curvature of the jaw (where the ventral edge is convex), an increase in depth of the dentary, a decrease in size and change in shape of the coronoid process, and an increase in size of the angular process. PC 2 corresponds to an enlargement of the distal tip of the dentary, and changes in shape of the coronoid process and mandibular condyle. PC3 indicates that residents have a narrower dentary overall, in contrast to transients and offshores which overlap in shape space on this PC axis (Figure 39).    57  Figure 37. The percentage contribution to the variance in shape from each principal component, for shape PCA of the left dentary. The first four principal components account for 74.3% of the variation in dentary shape.  2.3.3.2.2 CVA of Procrustes coordinates, dentary bone CVA was conducted to determine which shape features of the dentary bone best distinguish between resident, transient (Bigg’s), and offshore ecotypes (Figure 40). Mahalanobis distances (representing statistical distance between groups) were calculated, and pairwise distances among the ecotypes (resident – offshore, transient – offshore, and resident-transient) were all significant (p<0.0001, calculated from 10000 rounds of permutations). Procrustes distances (indicating a difference in shape) were calculated, showing statistical support for morphological distance between each pair (resident – offshore, transient – offshore, and resident - transient), where p <0.0001 from 10000 rounds of permutations. The shape of the coronoid process, the depth and curvature of the dentary, and the shape of the distal end of the dentary are variable traits that could be used to distinguish among the three ecotypes. principal components%variance30201005 10 15 20  58   Figure 38. Scatterplot of PC1 and PC2 scores for the Procrustes coordinates of each dentary bone along with 90% confidence ellipses for the means of each ecotype. Ecotype membership is indicated by colour (offshore = green, resident = blue, transient = orange). The wireframe graphs show the shape changes that are associated with each of PC 1 (27.9%) and PC 2 (21.6%), which together account for 49.5% of the shape variation of the dentary bones. The light blue wireframe indicates the reference shape and the dark blue wireframe represents the target shape, scale factor 0.10. The numbered dark blue dots and light blue open circles correspond to the position of 14 landmarks taken from the lateral side of the left dentary.   59  Figure 39. Scatterplot of PC3 and PC4 scores for the Procrustes coordinates of each dentary bone along with 90% confidence ellipses for the means of each ecotype. Ecotype membership is indicated by colour (offshore = green, resident = blue, transient = orange). The wireframe graphs show the shape changes that are associated with each of PC 3 (15%) and PC 4 (10%). The light blue wireframe indicates the reference shape, the dark blue wireframe represents the target shape, scale factor 0.10. The numbered dark blue dots and light blue open circles correspond to the position of 14 landmarks taken from the lateral side of the left dentary.  60   Figure 40. This scatter plot of CV 1 and CV 2 values for the left dentary bones from individuals from each ecotype (offshore = green, resident = blue, transient = orange) is the ordination of shape variables that best separate the ecotypes; percent variance is stated in parentheses. The wireframe graphs for CV1 and CV2 indicate the shape changes associated with each canonical variate. The light blue wireframe represents the reference shape, the dark blue wireframe represents the target shape (scale factor 10.0). The numbered dark blue dots and light blue open circles correspond to the position of 14 landmarks taken from the lateral side of the left dentary.  61 2.4 Discussion 2.4.1 Total body length and condylobasal length measurements For the individuals included in this study, the mean total body length of transient (Bigg’s) killer whales exceeded that of resident killer whales, whereas the mean body length of offshore killer whales was not significantly different from either the resident or transient (Bigg’s) killer whales (Figure 16). These data, although collected by numerous observers under different conditions and degrees of preciseness, indicate that the mean, median, and maximum total body length of transient (Bigg’s) killer whales exceeds those of resident killer whales. This supports field observations that transient (Bigg’s) killer whales appear much larger (longer and more robust) than resident killer whales (Ford et al., 2000; Barrett-Lennard and Heise, 2006). The field observations that offshore killer whales are smallest in size (Ford et al., 2000; Dahlheim et al., 2008) is not supported by the ANOVA of the means, nor the post hoc comparisons of these data, but the longest offshore killer whale in this study (620 cm) is shorter than the longest resident (701 cm) and transient (Bigg’s) killer whale (760 cm) individuals (Figure 16). The range of lengths observed do not represent all age classes for each ecotype. In particular, the abbreviated range of length measurements for transient (Bigg’s) killer whales likely only reflects the poor sampling from the youngest age classes, which can only be resolved over time when more specimens become available.    Transient (Bigg’s) killer whales also had longer skulls (SKL) than resident killer whales, whereas offshore killer whales were not significantly different from either resident or transient (Bigg’s) killer whales (Figure 17). Again, lack of representation from all age categories including neonates necessitates caution in interpretation of these data because no data was available for the smallest (neonatal or juvenile) offshore killer whales. However, the maximum skull lengths indicate that transient (Bigg’s) killer whales can achieve a larger skull length than the resident and offshore killer whales.   2.4.1.1 Functional inferences: do differences in overall body length reflect differences in dietary specialization Differences in body length can affect swimming velocity and agility. Other factors, however, such as fineness ratio, fluke size, stroke amplitude and frequency, and drag are also important (Fish, 1998; Rohr and Fish,   62 2004; Sato et al., 2007), making it impossible to infer the consequences of the direction or magnitude of difference in either parameter without direct measurement or appropriate modeling. Killer whales that feed on smaller prey (e.g., resident killer whales feeding on salmonids) may gain an advantage by being smaller and more agile. Increases in overall body size, if correlated to an increase in muscle mass, could confer an advantage to mammal-hunting killer whales since their prey are larger and more powerful, and require greater strength to subdue. However, it is not possible to rule out nutritional status (which depends on variables including prey availability, prey quality, and hunting success) as a possible factor in explaining overall size differences among the ecotypes. Also, any functional differences attributed to maximum body sizes found in each ecotype could be possibly confounded by group demographics where animals of different age classes travel and hunt together with sophisticated cooperative strategies (e.g., Baird and Whitehead, 2000), which can mitigate disadvantages related to the sizes of individuals.  2.4.2 Traditional morphometrics  Most of the traditional measurements of the cranium and jaws revealed that over all, the shape of the skull is similar among the three ecotypes. However, there were small but significant differences in features that could be correlated with diet differences. There were differences in the overall relative width of the braincase (POS/SKL, Figure 18, and OCW/SKL, Figure 19), the relative width of the proximal rostrum where it projects as a dorso-ventrally flattened shelf from the cranium (WRP/SKL and WAN/SKL), and the relative width of the palate (WAL/SKL). In each of these cases, the mean values for the resident killer whales were narrower than both the transient (Bigg’s) and offshore killer whales. Transients overall had the greatest breadth, and offshore killer whales appear in general to be intermediate to the other two, but offshores had a relatively wide rostrum (WRP/SKL) and palate (WAL/SKL) similar to that of transient (Bigg's) killer whales.  2.4.2.1 Possible morphological correlates to bite strength: skull width, and size of temporal fenestrae and fossae Bite strength in general is positively correlated with skull width in vertebrates due to the increase in space for attachment and expansion of jaw closing musculature (Thomason, 1991; Wroe et al., 2005). In mammals, the jaw adductor muscles include the temporalis, masseter and pterygoid muscle groups. The temporalis muscles   63 typically originate from an extensive area on the cranium (the temporal fossae), pass through the temporal fenestre, and insert upon the proximal end of the dentary on the dorsally projecting coronoid process. Masseter muscles originate on the zygomatic arch and insert on the angle of the dentary (Mead and Fordyce, 2009). In O. orca, however, the zygomatic arches are very thin rods rather than robust sites of muscle attachment, and were quite frequently missing or broken in the skulls in this study. Where the small, fatty masseters of delphinids insert on the lateral sides of the dentaries, there are no identifiable margins, in contrast to the well-defined masseteric fossa as seen in other mammals (Mead and Fordyce, 2009). Also, the pterygoid bones are fragile and were often found broken or missing on specimens; putative insertion sites for the pterygoideus muscles on the medial aspects of the dentary bone are also difficult to discern (Mead and Fordyce, 2009). Thus the primary adductor muscles considered here are the large temporalis muscles (Heyning and Dahlheim, 1988) due to their prominence in Orcinus, but also the presence of clear bony features that potentially allow inferences about muscle size and position.  Although it is best to measure physiological cross-sectional area of muscle directly (accounting for factors such as mass, fibre type, fibre length, and pennation pattern) to estimate performance measures such as contraction force or contraction speed, here the fenestrae bounded by bone serve as a proxy to compare the relative muscle sizes. The inner margins of the bones bounding the temporal fenestrae indicate the cross-sectional area of the muscle in that particular region of muscle, and the bony features forming the edges of the temporal fossae allow estimation of the size and position of the muscle origin. The traditional measurements of the temporal fenestrae and fossae however, did not indicate any differences among the ecotypes (Figures 25 and 26). However, these linear measurements did not capture the shape and size of the flared edges of the temporal region (the temporal crests), which likely serve as a large surface area and edge for the origin attachment for the temporalis muscles, in a manner similar to the sagittal crests observed in taxa such as Carnivora and in some primates. Also, the borders of the temporal fenestrae may not accurately represent the maximum cross-sectional area of the adductor muscles. Dissection of one neonatal killer whale revealed that the temporalis muscles bulge prominently in the temporal region of the head dorsal to the temporal arch (personal observation). In neonates, the edges of the temporal crest have not fused or elevated into ridges yet. This suggests that as the animals enlarge and mature and the temporal crests develop, the temporalis muscles can grow much larger in size than indicated by the bony features alone, and could in fact have relatively unencumbered space for expansion during maximal contraction. Therefore in this case,   64 only direct measurement of the jaw adductor muscles and further inspection of the temporal crests can resolve this question.   Thus while the greater width of the skull in the transient (Bigg’s) killer whales suggest they have greater bite strength, the data on the size of temporal fenestrae and fossae and other bony features do not allow any unequivocal conclusions to be drawn about the size of the masseter, pterygoid or temporalis muscles. Differences in the shape and the increased size of the flared edges of the temporal region (the temporal crests) would likely correlate with increased temporalis muscle size and increased bite forces in the transient (Bigg’s) killer whales, but measurements of muscle cross-sectional area as well as of the other static and dynamic factors that affect bite force (such as the velocity of jaw adduction, the optimal force-length relationship of each muscle, and the position of the prey in the mouth), will be required to adequately address this.   2.4.2.2 Resistance to torsional and bending forces: the rostrum and palate   The overall breadth of the rostrum and the palate would also likely affect feeding. The initial hypothesis was that a rostrum of greater width would be able to withstand greater torsional forces imposed by struggling prey in the jaws, which is an advantage for the transient (Bigg’s) killer whales that feed on relatively large mammalian prey, and possibly likewise for offshore killer whales that must grab and dismantle relatively large sharks. It is unclear whether feeding on salmonids would require reinforcement of the rostrum to the same extent, since the resident killer whales probably quickly snap their jaws to trap, kill, and divide fishes.  The structure of the rostrum, however, is quite different from carnivorous terrestrial mammals because there is a complete lack of fusion between the left and right sides of the maxilla and premaxilla bones on the distal (anterior) half of the rostrum. Perhaps, then, the structure of the proximal end of the rostrum, where the facial bones protrude from the cranium at the antorbital notches, deserves closer inspection. The linear measurements in this study did not quantify the ‘depth’ of the maxilla and premaxilla bones along the rostrum, nor the size and shape of the vomer, which protrudes into the rostrum between the maxillae and premaxillae. In this study the dimensions of the vomer were not accessible, and so must be further investigated.  In addition to torsion, the bending forces that are imposed on the skull during biting would also likely be quite pronounced in the dorso-ventrally flattened bones of the proximal rostrum. Bending forces are incurred on   65 skulls and jaws because of the forces imposed by contracting jaw adductor muscles. The mandibles pivot at the temporo-mandibular joint, and the biting forces are imposed on the skull where the prey is held, distal to the origin and insertion sites of the jaw adductor muscles; this arrangement is a third class lever (Thomason, 1991). Therefore the morphology of the bones in proximal rostrum should be considered in the future, and the size and shape of the rostral components measured directly.  To further illustrate the potential importance of the proximal rostrum: other top predators with a superficially similar skull are the tyrannosaurs, except that tyrannosaur skulls are more laterally flattened overall, rather than dorso-ventrally flattened at the rostrum, and also have a greater amount of fusion in the rostrum in contrast to killer whales. The highly fused nasal bones and lateral flattening of the skull results in a remarkable amount of rigidity, which allows for huge biting forces during jaw adductor contraction, and resistance to bending forces in the mid skull (Snively et al., 2006). Although nasal bones in delphinids do not contribute to the rostrum due to the telescoping of the skull, there are still likely massive forces acting on the other bones of the mid-skull region. Therefore one would predict that, particularly for animals feeding on very large prey and biting very hard components (osteophagy), the shape and depth of the rostrum is likely to be under strong selection pressure and should be scrutinized more closely.   2.4.2.3 Alveolar (tooth-bearing) bone The relative length of the upper left tooth-bearing row (LMT/SKL, Figure 27 A) in the maxilla was longest in the offshore killer whales compared to both the resident and the transient (Bigg’s) killer whales. In the dentary, the lower tooth row (AMN/SKL) was longer in the offshores and transient (Bigg's) killer whales compared to the resident killer whales (Figure 27 B). A general trend observed across vertebrate taxa is that very elongated skulls with numerous teeth function well for piscivorous hunters that use raptorial feeding (e.g., gavials among the crocodilians, or river dolphins), due to the relatively fast speed of jaw closure, and the traction from many pointed teeth. Also, biomechanical models of mammalian feeding predict that a mammal with a shorter rostrum will have a slower but more powerful bite than one with a longer rostrum, which have a faster but weaker bite (e.g., Aguirre et al., 2002). Unfortunately, rostrum length was not measured directly in this part of the study. The observation of piscivorous/shark-eating offshores possessing a relatively long maxillary (LMT/SKL) and dentary tooth row   66 (AMN/SKL), however, is confounded by the strange morphology of their teeth, where the cusps are worn down flat, in contrast to the pointed recurved teeth observed in transient (Bigg’s) and resident killer whales. A strong argument for the functional significance of tooth row length, therefore, cannot be made in this context, but could be clarified with direct measurements of rostrum length, and examination of tooth number, morphology, size, and distribution in the maxilla and dentary.   2.4.2.4 Dentary characteristics  Transient (Bigg's) killer whales have relatively longer dentary bones, but the traditional morphometrics approach did not reveal variations in size of the mandibular symphysis or the mandibular foramen (Figures 28 and 29). This was surprising since jaw structure is thought to be relatively plastic and subject to selection in animals with diets of different hardness and size (e.g., Scott et al., 2014; Santana et al., 2012). These data did not support the hypothesis that the transient (Bigg’s) killer whales would have more robust jaws compared to the resident and offshore killer whales. However, very few linear measurements were taken, and none of those quantified the shape variations of the different functional regions such as the coronoid process or mandibular symphysis.   2.4.2.5 Measurements of the foramen magnum, occipital condyles and nares   In regions of the skull not directly associated with jaw closure and feeding, it was found that there were no differences among the ecotypes for measurements of either the foramen magnum, the occipital condyles, or the nares (Figures 30-32). These data supported the hypothesis that these features would not vary across ecotypes since they were likely not subject directly to forces incurred during feeding events.  2.4.3 Geometric morphometrics In contrast to the traditional morphometrics measurements, where features of interest were chosen a priori, geometric morphometrics allows a more exploratory approach to characterizing morphological differences. For instance, angles or lengths or combinations of shape variables can be explored even if they were not strictly   67 identified during the data collection stage, and auto-correlated shape components that influence integrated structural features can be more easily visualized (Adams and Rohlf, 2000).  2.4.3.1 PCA and CVA on skull landmarks The greatest amount of variation in the skulls, as identified by the PCA of landmarks from the left side of the skull, is attributed to the braincase region (in the vault bones and also width of the skull where PC 1 is described by a relative expansion of the brain case) (Figure 34), which is consistent with the findings from the traditional morphometrics. In addition, there was variation in the shape and size of the premaxilla bones at the anterior half of the skull that was not quantified by the linear measurements from the traditional morphometrics analysis. PC 2 is associated with narrowing of the rostrum and braincase. The CVA for the skulls indicates that when the shape variables are mathematically maximally separated, there is successful clustering of individuals into each of the ecotypes.   2.4.3.2 PCA and CVA on dentary landmarks More shape differences in the dentary bones were detected with the geometric morphometrics approach than the traditional morphometrics method. Variations in shape and depth of the ramus of the dentary are described by PCs 1 to 4, as are relative size and shape differences in the coronoid process and angular process (Figures 38 and 39), all of which can be correlated to functional differences in jaw structure related to feeding. The piscivorous ecotypes (residents and offshores) appear to be distinguished by higher coronoid processes with larger surface areas relative to the transients (Bigg’s). The role of the coronoid process as an insertion site for jaw adductor muscles (jaw closing muscles) suggests that there could be an increase in the magnitude of forces applied by the temporalis muscles on the lever arm of the jaw. Examination of temporalis muscle insertion sites, which could inform models of the lever system of the dentary, would further clarify the functional significance of the large coronoid process. From the geometric morphometric analyses, and visual inspection of the bones, the body of the dentary in transient (Bigg’s) killer whales, on average, has a deeper curvature of the jaw, with a pronounced convex ventral edge. Greater depth (robustness) and bending in this direction would help resist the considerable bending and torsional forces placed on the dentary during feeding, particularly at the mid-point of the jaw during maximum bite   68 force. Reinforcement of the jaw in these ways could be correlated with the demands of gripping struggling animals, and from the bite forces imposed during biting large prey, in contrast to the presumably less demanding tasks of piscivory where an elongate gracile dentary is sufficient during rapid snap feeding (i.e., a faster but weaker bite that is sufficient for piercing fish). As it was for the CVA of the skull landmarks, dentary shape variation was great enough that it was possible to separate the ecotypes very clearly based on maximum ordination of jaw shape differences (Figure 40).   2.4.4 The effect of diet on cranial morphology – plasticity, genetics, and age  In morphological studies that consider dietary differences, one concern is that the shape and size differences observed could arise due to plasticity or age. Under ideal, experimental, conditions it would be possible to infer the relative roles of plasticity and genetic control on the shape differences observed among the ecotypes (e.g., a common garden experiment). For this species, it is only possible to consider opportunistic data and anecdotal evidence.  One female transient (Bigg’s) killer whale specimen in this study was live-captured in 1970 as a juvenile (estimated age 7 years), and kept in captivity for 20 years (LACM 84291, “Knootka”). The skull and dentaries of this adult individual are robust (Figure 41), and in my opinion, exhibit a ‘typical’ transient (Bigg’s) killer whale morphology. The cranium (Figure 41 B) was not included in the geometric morphometrics analysis because of the removal of much of the braincase presumably during necropsy. It is likely that this animal was fed fishes or other small prey and implausible that it hunted and ate marine mammals during its time in captivity. Therefore I would consider this reasonable anecdotal evidence that the robustness and shape of the mandible is in large part genetic, or, at least, canalized during development in juveniles.   Another factor to consider is the effect of age, particularly in such long-lived animals where shape changes in bone can be attributed to illnesses such as osteoarthritis. RBCM 16006 is a mature (estimated age 62-years) female resident killer whale whose dentary bones exhibit a gracile structure along with a large coronoid process (Figure 42). This shape is consistent with what I consider to be ‘typical’ of resident killer whales, and is certainly consistent with the resident-type shapes as shown in Figures 38 and 40, and resembles the gracile shape of most adult resident killer whale jaws observed in this study. The left dentary of this jaw was not included in any of the   69 geometric morphometrics analyses due to the profound damage at the distal end. This provides some evidence that ‘robustness’ does not correlate with advanced age, and that at least differences between transient (Bigg’s) and resident killer whale shapes cannot be attributed to age.  A       B  Figure 41. (A) Left dentary, labial view, and (B) cranium, dorsal view, of LACM84291, a 27-year-old female transient (Bigg’s) killer whale that was kept in captivity for 20 years.   A      B        Figure 42. (A) Right (top) and left (bottom) dentaries, labial view, and (B) cranium, dorsal view of RBCM16006, a southern resident, ID L66, female, estimated age 62-years).   2.4.5 Sexual dimorphism and allometry Sexual dimorphism was not quantified in this study due to the unequal and inadequate representation of known males and females from each ecotype. Likewise, age data was not available for many samples, and for many skulls, the teeth are missing which precludes estimation of age via tooth sectioning. These fundamental issues must   70 be resolved as more data become available over time. Sex is likely to be an important factor, considering the profound sexual dimorphism observed in dorsal fin and flipper size where the males have larger structures (Heyning and Dahlheim, 1988). Field observations suggest that sexual dimorphism is not as pronounced in offshore killer whales (Dahlheim et al., 2008). When I conducted preliminary two-way ANOVAs on a few shape variables with ecotype and sex, sex did appear to affect cranial shape. Hopefully as more data accumulates the appropriate statistical analyses accounting for sex can be applied. It remains a possibility that sexually dimorphic traits will be significant, but may not affect the degree of morphological difference observed among the three ecotypes (i.e., the intra-ecotype differences will not obscure the inter-ecotype differences that are observed). Mammals do not grow isometrically - that is, changes in shape occur as animals grow in size. In this thesis, for the traditional morphometrics measurements, size was adjusted by expressing each measurement as a ratio of skull length. The geometric morphometrics approach incorporated scaling of each animal during generalized Procrustes superimposition to directly compare shape. This issue was a concern here because all age classes were included in the analysis, which encompasses a large range of skull sizes. However, there were not enough data points to appropriately characterize allometry for each ecotype, and so the most simple size adjustment was used in this analysis.  2.4.6 Summary/conclusions The goal of this thesis was to compare the cranial morphology of killer whales from the northeast Pacific in the context of the strict ecotypic differences among mammal-eating transient (Bigg’s), and piscivorous resident and offshore killer whales. I hypothesized that mammal-eating transient (Bigg’s) whales would have more robust structures in response to selection pressures related to hunting and eating larger marine mammal prey with massive endoskeletons, in contrast to piscivorous resident killer whales that prefer hunting smaller salmonid prey, or offshore killer whales whose diet includes teleost fishes and sharks.   Traditional morphometric measurements of resident, offshore and transient (Bigg’s) killer whale skulls revealed a number of small but significant differences in features that support this hypothesis. A geometric morphometrics analysis of cranial landmarks indicated that when the shape variables are mathematically maximally separated, there is a successful clustering of individuals into each ecotype grouping, adding further support to the   71 hypothesis that there is morphological distance between the groups. A landmark PCA revealed regions of variation in the skull and the dentary that were not apparent from the traditional morphometrics data such as in the increased size of the coronoid processes of the dentaries in the piscvorous ecotypes, the overall curvature in shape of the dentary in transient (Bigg’s), and the overall greater depth of the dentary in transients relative to the resident and offshore killer whales. The relative overall robustness of transient (Bigg’s) killer whale jaws and rostra, particularly in contrast to the more gracile structures observed in the resident killer whales, is consistent with the hypothesis that the rigors of handling larger marine mammal prey would correlate with bony features that could resist the accompanying bending and torsional forces.  The cranial morphology of offshore killer whales is difficult to clearly discern from both resident and transient (Bigg’s) killer whales, and exhibited large variation in shape. The small sample size is a plausible contributing factor to the high variance, but this conclusion is consistent with qualitative visual inspection of the skulls. I found no obvious identifying features in the cranium or the dentaries, except for the flattened dentition and the length of the maxillary tooth row, that is, on average, longer than in resident and transient (Bigg's) killer whales.  The variability of offshores that I observed is interesting considering their likely broad diet diversity (which is not yet well understood), particularly in contrast to the strong preference of resident killer whales for Chinook salmon. Inferring a clear relationship between morphological variance and diet breadth is not possible here, however, since transient (Bigg’s) killer whales are also likely to hunt a diversity of species of different sizes and age classes. More specimens will be required to adequately characterize the range of cranial shapes in offshore killer whales. The observations from this study invite biomechanical analyses of the effect of increased cranial width, increased rostral/palatal width, and the differences in dentary shape in O. orca. Based on the extreme dietary difference between these three groups, it was not surprising to find morphological differences in the cranial skeleton. Genetic drift cannot be ruled out as a mechanism for the evolution of these morphological differences, but there is reason to suspect natural selection: the differences between the transient (Bigg’s) and resident killer whales, in particular, that were found in the braincase, rostrum and lower jaws are all likely functionally relevant to the process of acquiring and processing their diet, which is a known driver of diversification in many animals. All of the   72 proposed anatomical and biomechanical correlates described in this thesis rely on simple predictions based on the principles of functional morphology, which require further investigation and testing of the assumptions.   73 Chapter 3: Conclusion Killer whales are incredible predators. Their size and strength in addition to the imposing architecture of their skulls and teeth enable them to perform impressive tasks, whether cooperatively hunting baleen whales, or precise capture of fast-moving salmonids. It is thought that killer whales evolved as predators of marine mammals (Springer et al., 2003) but some populations of killer whales have specialized to hunt and consume fish. Due to the profound differences in dietary preference among the resident, offshore, and transient (Bigg’s) killer whales, I hypothesized that there would be morphological differences in the cranial skeleton, and proposed possible morphological correlates with their dietary specializations. Traditional and geometric morphometrics methods were employed to characterize the cranial morphology of the three ecotypes of Northeast Pacific killer whales (Orcinus orca) and to test these hypotheses.  The most important differences found among the three northeast Pacific ecotypes were of the cranial width and the overall shape of the lower jaw. The mammal-hunting transient (Bigg's) whales had the broadest skulls overall, and the most robust dentary bones. The shape of the dentary, at least in adult resident killer whales, is more gracile, and straighter in shape compared to the robust and curved dentaries of most adult transient (Bigg's) killer whales. Also, the coronoid process is different in shape and larger in the piscivorous offshore and resident killer whales compared to transient (Bigg's) killer whales. Offshore killer whales were quite variable in shape, but were found, on average, to have relatively longer rows of teeth compared to the other two ecotypes in this study.  3.1 The trouble with offshores Of the few measured, offshore killer whales are quite variable in cranial structure, and were not clearly morphologically distinct from the resident and transient (Bigg's) ecotypes. Perhaps this was not surprising in the context of the uncertainty of their genetic relationship to the other groups. Northeast Pacific resident, offshore, and transient (Bigg’s) killer whales are unequivocally genetically distinct with little to no gene flow between them. However, from mitochondrial (haplotype) data, offshore killer whales are calculated to be more closely related to resident killer whales (i.e., have a more recent maternal ancestor), whereas in phylogenies constructed with nuclear (microsatellite) data, offshore killer whales cluster more closely with transient (Bigg’s) killer whales (Barrett-Lennard, 2000; Pilot, 2010; Morin, 2015). Thus the patterns of ancestry and relatedness of offshore killer whales to   74 the other ecotypes may be obscured by periods of backcrossing, hybridization, or introgression (Barrett-Lennard, 2000) that could obscure patterns of morphological divergence as well.  Also, although the anecdotal evidence presented in this thesis would suggest that factors such as plasticity and age are not likely to explain the morphological differences found between the resident and transient (Bigg’s) killer whales, there is no evidence that plasticity, age, diet and other environmental factors do not impact development of cranial morphology in offshore killer whales. Consider the very odd challenge of hunting with, apparently, progressively flattened and destroyed dentition over time: this must affect their feeding efficiency and hunting strategies. The plasticity of bones in general, and the very plausible changes in static and dynamic bite forces caused by dentition changes could lead to the population-level variation observed in this ecotype. This deserves consideration as more is learned about offshore killer whales.  3.2 False discovery rate control A number of analytical problems regarding sex, age and allometry were raised in Chapter 2. In addition, with the traditional morphometrics analysis, statistical tests were conducted with multiple comparisons made within the same data set. A TSBH false discovery rate (FDR) - controlling procedure was used, where FDR was set at 0.02. This, in retrospect, could be judged to be rather conservative, because in an exploratory morphometrics study such as this, it may not be necessary- a false positive is not desirable, but is not detrimental. In cases where a higher FDR is allotted, and with the caveat that future studies can be used to either support or refute the observations, it will also decrease the chance of type II error (i.e., failing to find a significant effect), which is arguably more important. Bearing this in mind, the traditional morphometrics results were considered again with a higher FDR of 0.05 (Appendix C). In this calculation, three more measurements would have been found to be significantly different among ecotypes: the relative length of the left temporal fossa (LPF/SKL, p=0.02248, Figure 25B), the relative length of the mandibular symphysis (LSY/SKL, p=0.03664, Figure 29A), and the relative width across the occipital condyles (WOC/SKL, p=0.02831, Figure 31B). Post hoc tests revealed, however, that there were inter-ecotype differences only for WOC/SKL, where resident killer whales had wider occipital condyles than offshore killer whales (Tukey HSD, p=0.026*), but transient killer whales showed no difference from either offshore or resident   75 killer whales (p >0.05). For post hoc pair-wise comparisons in either LPF/SKL and LSY/SKL, there were no significant differences among any of the ecotypes (p>0.05). The finding that there may be differences in relative width across the occipital condyles should be verified in future work. The functional significance of this morphological difference is unclear. The occipital condyles serve as the articulation between the cranium and the first cervical vertebra (the atlas) and as an attachment site for the alar ligaments (Vishteh et al., 1999). Overall neck mobility in odontocetes is highly restricted, however, due to the profound fusion of the cervical vertebrae, except in the Monodontidae (the belugas (Delphinapterus leucas), and narwhals (Monodon monoceros) (Jefferson et al., 1993)). The effect of occipital condyle size could potentially have an impact on all aspects of head movement, however small these movements may be (e.g., axial rotation, lateral bending, flexion, and extension) (Vishteh et al., 1999), particularly if there are few regions of flexibility in the cervical region.  3.3 Future directions The observed differences in shapes among the three ecotypes were each relatively small. However, taken altogether with theoretical changes in muscle size and positioning, and therefore jaw leverage, these data beg investigation into bite force performance in these groups, along with further study of the temporal fossa and temporal crest, the structure of the rostral bones and dentary bones. The geometrics morphometrics analysis was illuminating, however, because of the ability to quantify and summarize overall shape changes of the entire cranium and dentary. Soft tissue was not available in this study, but where possible, muscle and connective tissue anatomy would be fundamental for any firm conclusions to be drawn about biting performance. These features along with skeletal morphology all deserve a proper biomechanical treatment to infer how these shape changes impact feeding efficiency in the context of hunting different prey sizes and types. Also, to infer whether any differences among the groups are due to processes such as ecological character displacement, drift, or natural selection, there must be systematic comparison between other sympatric and allopatric pairs of populations with either similar or divergent diet specializations, such as those found in the Antarctic or the North Atlantic oceans. In particular, if any parallel patterns of cranial morphological divergence are found in   76 sympatric or parapatric pairs of piscivorous and mammal-hunting specialists, this would give further insight in the patterns of diversification in this taxon. Although small in magnitude, the morphological differences among the ecotypes in this study are intriguing because they are likely maintained by the culturally-enforced ecotypic reproductive barriers among members of the resident, offshore and transient (Bigg's) ecotypes.  3.4 Killer whale taxonomy  Taxonomic studies require morphological data in concordance with genetic and behavioural evidence (Reeves et al., 2004). Of killer whale populations genetically sequenced, transient (Bigg’s) killer whales form a monophyletic group that diverged early on from the lineage. It has been inferred from genetic data that sympatric contact between the piscivorous ecotypes in the Pacific and the transient (Bigg’s) ecotype could have arisen secondarily, subsequent to transient (Bigg's) killer whale migration from the Atlantic into the Pacific ocean (Morin et al., 2010), but other geneticists argue for a scenario where transients and residents diverged in sympatry (Moura et al., 2014). Nonetheless, there is interest in formally naming the transients (“Bigg’s killer whale”) and residents as separate taxa (Committee on Taxonomy, 2016). These morphometric data provide evidence that there are indeed shape differences between sympatric northeast Pacific transient (Bigg’s) and resident killer whales in the cranial skeleton. More anatomical and behavioural data from the offshore ecotype is required to draw firm conclusions, but the results from this thesis show that there is indeed morphological distance among the three groups, with the most support for morphological divergence between the gracile resident and robust transient (Bigg's) killer whale ecotypes.    77 References  Avery, A., Straley, J., Atkinson, S., and O’Connell, V. (2011) Skeletal preparation and measurements of a young transient killer whale (Orcinus orca) from Southeast Alaska. Sitka Sound Science Center, Sitka, Alaska. Adams, D.C. and Rohlf, F.J. (2000). Ecological character displacement in Plethodon: biomechanical differences found from a geometric morphometric study. Proc. Natl. Acad. Sci. U.S.A., 97, 4106–4111. Adams, D.C., Rohlf, F.J., and Slice, D.E. (2004). Geometric morphometrics : Ten years of progress following the “revolution.” Ital. J. Zool., 71, 5–16. Au, W.W.L., Ford, J.K.B., Horne, J.K., and Allman, K.A.N. (2004). Echolocation signals of free-ranging killer whales (Orcinus orca) and modeling of foraging for chinook salmon (Oncorhynchus tshawytscha). J. Acoust. Soc. Am., 115, 901–909. Baird, R.W., Abrams, P.A., and Dill, L.M. (1992). Possible indirect interactions between transient and resident killer whales: implications for the evolution of foraging specializations in the genus Orcinus. Oecologia, 89, 125–132. Baird, R.W. and Dill, L.M. (1995). Occurrence and behaviour of transient killer whales: seasonal and pod-specific variability, foraging behaviour, and prey handling. Can. J. Zool., 73, 1300–1311. Baird, R.W. and Stacey, P.J. (1988). Variation in saddle patch pigmentation in populations of killer whales (Orcinus orca) from British Columbia, Alaska, and Washington State. Can. J. Zool., 66, 2582–2585. Baird, R.W. and Whitehead, H. (2000). Social organization of mammal-eating killer whales: group stability and dispersal patterns. Can. J. Zool., 78, 2096–2105. Barrett-Lennard, L.G. (2000). Population structure and mating patterns of killer whales (Orcinus orca) as revealed by DNA analysis. Ph.D. Dissertation. University of British Columbia, Vancouver. Barrett-Lennard, L.G. and Heise, K.A. (2006). The natural history and ecology of killer whales. In, Brownell, R.L., DeMaster, D.P., Doak, D.F., and Williams, T.M. (eds), Whales, whaling and ocean ecosystems. University of California Press, Berkeley, CA, pp. 163–173. Barrett-Lennard, L.G. (2011). Killer whale evolution: Populations, ecotypes, species, oh my! Whalewatcher J. Am. Cetacean Soc., 40, 48–53.    78 Barrett-Lennard, L.G., Ford, J.K.B., and Heise, K.A. (1996). The mixed blessing of echolocation: differences in sonar use by fish-eating and mammal-eating killer whales. Anim. Behav., 51, 553–565. Barrett-Lennard, L.G., Matkin, C.O., Durban, J.W., Saulitis, E.L., and Ellifrit, D. (2011). Predation on gray whales and prolonged feeding on submerged carcasses by transient killer whales at Unimak Island, Alaska. Mar. Ecol. Prog. Ser., 421, 229–241. Benjamini, Y., Krieger, A.M., and Yekutieli, D. (2006). Adaptive linear step-up procedures that control the false discovery rate. Biometrika, 93, 491–507. Berzin, A.A. and Vladimirov, V.L. (1983). Novyi vid kosatki (Cetacea, Delphinidae) iz vod Antarktiki (A new species of killer whale (Cetacea, Delphinidae) from Antarctic waters). Zool. Zhurnal, 62, 287–295. Bigg, M. (1982). An assessment of killer whale (Orcinus orca) stocks off Vancouver Island, British Columbia. Rep. Int. Whal. Comm., 32, 655–666. Bigg, M.A., Ellis, G.M., Ford, J.K.B., and Balcomb, K.C. (1987). Killer whales: A study of their identification, genealogy, and natural history in British Columbia and Washington State 1st ed. Phantom Press, Nanaimo, B.C. Christiansen, P. and Wroe, S. (2007). Bite forces and evolutionary adaptations to feeding ecology in carnivores. Ecology, 88, 347–358. Committee on Taxonomy. 2014. List of marine mammal species and subspecies. Society for Marine Mammalogy, www.marinemammalscience.org, accessed on 2016 June 03. COSEWIC. 2008. COSEWIC assessment and update status report on the Killer Whale Orcinus orca, Southern Resident population, Northern Resident population, West Coast Transient population, Offshore population and Northwest Atlantic / Eastern Arctic population, in Canada. Committee on the Status of Endangered Wildlife in Canada. Ottawa. viii + 65 pp., www.sararegistry.gc.ca/status/status_e.cfm Dahlheim, M.E., Schulman-Janiger, A., Black, N., Ternullo, R., Ellifrit, D., and Balcomb III, K.C. (2008). Eastern temperate North Pacific offshore killer whales (Orcinus orca): Occurrence, movements, and insights into feeding ecology. Mar. Mammal Sci., 24, 719–729. Deecke, V.B., Slater, P.J.B., and Ford, J.K.B. (2002). Selective habituation shapes acoustic predator recognition in harbour seals. Nature, 420, 171–3.   79 Deecke, V., Ford, J., and Slater, P. (2005). The vocal behaviour of mammal-eating killer whales: communicating with costly calls. Anim. Behav., 69, 395–405. Deecke, V., Ford, J., and Spong, P. (2000). Dialect change in resident killer whales: implications for vocal learning and cultural transmission. Anim. Behav., 60, 629–638. Deecke, V., Nykänen, M., Foote, A., and Janik, V. (2011). Vocal behaviour and feeding ecology of killer whales Orcinus orca around Shetland, UK. Aquat. Biol., 13, 79–88. Fish, F.E. (1998). Comparative kinematics and hydrodynamics of odontocete cetaceans: morphological and ecological correlates with swimming performance. J. Exp. Biol., 201, 2867–77. Foote, A.D., Morin, P.A., Durban, J.W., Pitman, R.L., Wade, P., Willerslev, E., Gilbert, M.T.P., and da Fonseca, R.R. (2011). Positive selection on the killer whale mitogenome. Biol. Lett., 7, 116–8. Foote, A.D., Newton, J., Piertney, S.B., Willerslev, E., and Gilbert, M.T.P. (2009). Ecological, morphological and genetic divergence of sympatric North Atlantic killer whale populations. Mol. Ecol., 18, 5207–5217. Foote, A.D., Vilstrup, J.T., De Stephanis, R., Verborgh, P., Abel Nielsen, S.C., Deaville, R., Kleivane, L., Martín, V., Miller, P.J.O., Oien, N., et al. (2011). Genetic differentiation among North Atlantic killer whale populations. Mol. Ecol., 20, 629–41. Foote, A., Vijay, N., Avila-Arcos, M., Baird, R., Durban, J., Morin, P., Fumagalli, M., Gibbs, R., Hanson, B., Korneliussen, T., et al. (2016). Genome-culture coevolution promotes rapid divergence in the killer whale. Nat. Commun., 1–48. Ford, J.K.B. (2011). Killer whales of the Pacific northwest coast: From pest to paragon. Whalewatcher J. Am. Cetacean Soc., 40, 15–23. Ford, J.K.B. (1991). Vocal traditions among resident killer whales (Orcinus orca) in coastal waters of British Columbia. Can. J. Zool., 69, 1454–1483. Ford, J.K.B. and Ellis, G.M. (2006). Selective foraging by fish-eating killer whales Orcinus orca in British Columbia. Mar. Ecol. Prog. Ser., 316, 185–199. Ford, J.K.B. and Ellis, G.M. (1999). Transients: Mammal-hunting killer whales 1st ed. UBC Press, Vancouver, B.C. Ford, J.K.B., Ellis, G.M., and Balcomb, K.C. (2000). Killer Whales 2nd ed. UBC Press, Vancouver, B.C.    80 Ford, J.K.B., Ellis, G.M., Barrett-Lennard, L.G., Morton, A.B., Palm, R.S., and Balcomb III, K.C. (1998). Dietary specialization in two sympatric populations of killer whales (Orcinus orca) in coastal British Columbia and adjacent waters. Can. J. Zool., 76, 1456–1471. Ford, J.K.B., Ellis, G.M., Matkin, C.O., Wetklo, M.H., Barrett-Lennard, L.G., and Withler, R.E. (2011). Shark predation and tooth wear in a population of northeastern Pacific killer whales. Aquat. Biol., 11, 213–224. Ford, J.K.B., Ellis, G.M., Matkin, D. R., Balcomb, K.C., Briggs, D., and Morton, A.B. (2005). Killer whale attacks on minke whales: Prey capture and antipredator tactics. Mar. Mammal Sci., 21, 603–618. Ford, J.K.B. and Fisher, H.D. (1982). Killer whale (Orcinus orca) dialects as an indicator of stocks in British Columbia. Rep. Int. Whal. Comm., 32, 671–679. Hammer, Ø., Harper, D.A.T., and Ryan, P.D. (2001). PAST : Paleontological statistics software package for education and data analysis. Palaeontol. Electron., 4, 1–9. Hanson, M.B., Baird, R.W., Ford, J.K.B., Hempelmann-Halos, J., Van Doornik, D., Candy, J., Emmons, C., Schorr, G., Gisborne, B., Ayres, K., et al. (2010). Species and stock identification of prey consumed by endangered southern resident killer whales in their summer range. Endanger. Species Res., 11, 69–82. Heise, K., Barrett-Lennard, L.G., Saulitis, E., Matkin, C., and Bain, D. (2003). Examining the evidence for killer whale predation on Steller sea lions in British Columbia and Alaska. Aquat. Mamm., 29, 325–334. Herman, D.P., Burrows, D.G., Wade, P.R., Durban, J.W., Matkin, C.O., Leduc, R.G., and Krahn, M.M. (2005). Feeding ecology of eastern North Pacific killer whales Orcinus orca from fatty acid, stable isotope, and organochlorine analyses of blubber biopsies. Mar. Ecol. Prog. Ser., 302, 275–291. Heyning, J.E. and Dahlheim, M.E. (1988). Orcinus orca. Mamm. Species, 304, 1–9. Hoelzel, A.R. and Dover, G.A. (1991). Genetic differentiation between sympatric killer whale populations. Heredity (Edinb)., 66, 191–195. Hoelzel, A.R., Hey, J., Dahlheim, M.E., Nicholson, C., Burkanov, V., and Black, N. (2007). Evolution of population structure in a highly social top predator, the killer whale. Mol. Biol. Evol., 24, 1407–15. Hoyt, E. (1990). Orca, the whale called killer 3rd ed. Robert Hale Limited, London. Jefferson, T.A., Stacey, P. J., and Baird, R.W. (1991). A review of killer whale interactions with other marine mammals: predation to co-existence. Mamm. Rev., 21, 151–180.   81 Jefferson, T.A., Leatherwood, S. and Webber, M.A. (1993) FAO species identification guide. Marine mammals of the world. Rome, FAO. 1993.320. p. 587  JMP®, Version 9.0.2. SAS Institute Inc., Cary, NC, 1989-2007. Jones, I.M. (2006). A northeast Pacific offshore killer whale (Orcinus orca) feeding on a Pacific halibut (Hippoglossus stenolepis). Mar. Mammal Sci., 22, 198–200. Klingenberg, C.P. (2011). MorphoJ: an integrated software package for geometric morphometrics. Mol. Ecol. Resour., 11, 353–7. Krahn, M., Ford, M.J., Perrin, W.F., Wade, P.R., Angliss, R.P., Hanson, M.B., Taylor, B.L., Ylitano, G.M., Dahlheim, M.E., Stein, J.E., and Waples, R.S.. (2004). Status Review of Southern Resident Killer Whales (Orcinus orca) under the Endangered Species Act. NOAA Technical Memorandum, NMFS-NWFSC-62. 73pp. Krahn, M.M., Herman, D.P., Matkin, C.O., Durban, J.W., Barrett-Lennard, L., Burrows, D.G., Dahlheim, M.E., Black, N., LeDuc, R.G., and Wade, P.R. (2007). Use of chemical tracers in assessing the diet and foraging regions of eastern North Pacific killer whales. Mar. Environ. Res., 63, 91–114. LeDuc, R.G., Robertson, K.M., and Pitman, R.L. (2008). Mitochondrial sequence divergence among Antarctic killer whale ecotypes is consistent with multiple species. Biol. Lett., 4, 426–9. Maniscalco, J.M., Matkin, C.O., Maldini, D., Calkins, D.G., and Atkinson, S. (2007). Assessing killer whale predation on Steller sea lions from field observations in Kenai Fjords, Alaska. Mar. Mammal Sci., 23, 306–321. Matkin, C.O., Matkin, D. R., Ellis, G.M., Saulitis, E., and McSweeney, D. (1997). Movements of resident killer whales in southeastern Alaska and Prince William Sound, Alaska. Mar. Mammal Sci., 13, 469–475. Mead, J.G. and Fordyce, R.E. (2009). The Therian Skull - A Lexicon with Emphasis on the Odontocetes. Smithson. Contrib. to Zool., x, 261. Morin, P.A., Archer, F. I., Foote, A. D., Vilstrup, J., Allen, E.E., Wade, P., Durban, J., Parsons, K., Pitman, R., Li, L., et al. (2010). Complete mitochondrial genome phylogeographic analysis of killer whales (Orcinus orca) indicates multiple species. Genome Res., 20, 908–16.     82 Morin, P.A., LeDuc, R.G., Robertson, K.M., Hedrick, N.M., Perrin, W.F., Etnier, M., Wade, P., and Taylor, B. L. (2006). Genetic analysis of killer whale (Orcinus orca) historical bone and tooth samples to identify western U.S. ecotypes. Mar. Mammal Sci., 22, 897–909. Morin, P.A., Parsons, K.M., Archer, F.I., Ávila-Arcos, M.C., Barrett-Lennard, L.G., Dalla Rosa, L., Duchêne, S., Durban, J.W., Ellis, G.M., Ferguson, S.H., et al. (2015). Geographic and temporal dynamics of a global radiation and diversification in the killer whale. Mol. Ecol., 24, 3964–3979. Moura, A.E., Kenny, J.G., Chaudhuri, R.R., Hughes, M.A., Reisinger, R.R., de Bruyn, P.J.N., Dahlheim, M.E., Hall, N., and Hoelzel, A.R. (2014). Phylogenomics of the killer whale indicates ecotype divergence in sympatry. Heredity (Edinb)., 1–8. Moura, A., Kenny, J., Chaudhuri, R., Hughes, M., Welch, A., Reisinger, R., de Bruyn, P., Dahlheim, M., Hall, N., and Hoelzel, A. (2014). Population genomics of the killer whale indicates ecotype evolution in sympatry involving both selection and drift. Mol. Ecol., 23, 5179–5192. Newsome, S., Etnier, M., Monson, D., and Fogel, M. (2009). Retrospective characterization of ontogenetic shifts in killer whale diets via δ13C and δ15N analysis of teeth. Mar. Ecol. Prog. Ser., 374, 229–242. Perrin, W.F. (1975). Variation of spotted and spinner porpoise (genus stenella) in the Eastern Pacific and Hawaii. Pilot, M., Dahlheim, M.E., and Hoelzel, A.R. (2010). Social cohesion among kin, gene flow without dispersal and the evolution of population genetic structure in the killer whale (Orcinus orca). J. Evol. Biol., 23, 20–31. Pitman, R.L., Deecke, V.B., Gabriele, C.M., Srinivasan, M., Black, N., Denkinger, J., Durban, J.W., Mathews, E.A., Matkin, D.R., Neilson, J.L., et al. (2016). Humpback whales interfering when mammal-eating killer whales attack other species: Mobbing behavior and interspecific altruism? Mar. Mammal Sci., 1–52. DOI: 10.1111/mms.12343 Pitman, R.L. and Ensor, P. (2003). Three forms of killer whales (Orcinus orca) in Antarctic waters. J. Cetacean Res. Manag., 5, 131–139. Pollard, K.S., Dudoit, S., van der Laan, M.J. (2005). Multiple Testing Procedures: R multtest Package and Applications to Genomics, in Bioinformatics and Computational Biology Solutions Using R and Bioconductor, Gentleman, R., Carey, V., Huber, W., Irizarry, R., Dudoit, S. (eds.). Springer (Statistics for Biology and Health Series), pp. 251-272.   83 R Core Team (2015). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/ Reeves, R.R., Perrin, W.F., Taylor, B.L., Baker, C.S., and Mesnick, S.L. (eds.). (2004). Report of the Workshop on Shortcomings of Cetacean Taxonomy in Relation to Needs of Conservation and Management, 30 April to 2 May, La Jolla, California. NOAA-Technical Memorandum-NMFS-SWFSC-363, 94 pp. Riesch, R., Barrett-Lennard, L.G., Ellis, G.M., Ford, J.K.B., and Deecke, V. B. (2012). Cultural traditions and the evolution of reproductive isolation: ecological speciation in killer whales? Biol. J. Linn. Soc., 106, 1–17. Rohlf, F. J. (1998). On applications of geometric morphometrics to studies of ontogeny and phylogeny. Syst. Biol., 47, 147–158. Rohlf, F. J. and Marcus, L. F. (1993). A revolution in morphometrics. Trends Ecol. Evol., 8, 129–132. Rohlf F.J. (2010) Tps Series. Department of Ecology and Evolution, State University of New York, Stony Brook, New York. http://life.bio.sunysb.edu/morph/, accessed on 2015 June 8. Rohr, J.J. and Fish, F.E. (2004). Strouhal numbers and optimization of swimming by odontocete cetaceans. J. Exp. Biol., 207, 1633–1642. Santana, S.E., Dumont, E.R., and Davis, J.L. (2010). Mechanics of bite force production and its relationship to diet in bats. Funct. Ecol., 24, 776–784. Santana, S.E., Grosse, I.R., and Dumont, E.R. (2012). Dietary hardness, loading behavior, and the evolution of skull form in bats. Evolution, 66, 2587–98. Sato, K., Watanuki, Y., Takahashi, A., Miller, P.J.O., Tanaka, H., Kawabe, R., Ponganis, P.J., Handrich, Y., Akamatsu, T., Watanabe, Y., et al. (2007). Stroke frequency, but not swimming speed, is related to body size in free-ranging seabirds, pinnipeds and cetaceans. Proc. R. Soc. B Biol. Sci., 274, 471–477. Saulitis, E., Matkin, C., Barrett-Lennard, L., Heise, K., and Ellis, G. (2000). Foraging strategies of sympatric killer whale (Orcinus orca) populations in Prince William Sound, Alaska. Mar. Mammal Sci., 16, 94–109. Scheel, D., Matkin, C.O., and Saulitis, E. (2001). Distribution of Killer Whale Pods in Prince William Sound, Alaska 1984–1996. Mar. Mammal Sci., 17, 555–569. Scott, J.E., Hogue, A.S., and Ravosa, M.J. (2012). The adaptive significance of mandibular symphyseal fusion in mammals. J. Evol. Biol. 25, 661-673   84 Scott, J.E., McAbee, K.R., Eastman, M.M., and Ravosa, M.J. (2014). Teaching an old jaw new tricks: diet-induced plasticity in a model organism from weaning to adulthood. J. Exp. Biol., 217, 4099–4107. Snively, E., Henderson, D.M., and Phillips, D.S. (2006). Fused and vaulted nasals of tyrannosaurid dinosaurs : Implications for cranial strength and feeding mechanics. Acta Palaeontol. Pol., 51, 435–454. Springer, A., Estes, J., van Vliet, G., Williams, T., Doak, D.F., Danner, E., Forney, K., and Pfister, B. (2003). Sequential megafaunal collapse in the North Pacific Ocean: an ongoing legacy of industrial whaling? Proc. Natl. Acad. Sci. U.S.A., 100, 12223–12228. Stevens, T.A., Duffield, D.A., Asper, E.D., Hewlett, K.G., Bolz, A., Gage, L.J., and Bossart, G.D. (1989). Preliminary findings of restriction fragment differences in mitochondrial DNA among killer whales (Orcinus orca). Can. J. Zool., 67, 2592–2595. Sztencel-Jabłonka, A., Jones, G., and Bogdanowicz, W. (2009). Skull morphology of two cryptic bat species: Pipistrellus pipistrellus and P. pygmaeus - a 3D geometric morphometrics approach with landmark reconstruction. Acta Chiropterologica, 11, 113–126. Thomason, J.J. (1991). Cranial strength in relation to estimated biting forces in some mammals. Can. J. Zool., 69, 2326–2333. Towers, J.R., Ellis, G.M., and Ford, J.K.B. (2012). Photo-identification catalogue of Bigg’s (transient) killer whales from coastal waters of British Columbia, northern Washington, and southeastern Alaska. Can. Data Rep. Fish. Aquat. Sci. 1241: v + 127 p.  Vishteh, A.G., Crawford, N.R., Melton, M.S., Spetzler, R.F., Sonntag, V.K.H., and Dickman, C.A. (1999). Stability of the craniovertebral junction after unilateral occipital condyle resection: a biomechanical study. J. Neurosurg. (Spine 1), 90, 91–98. Womble, J.N., Gende, S.M., and Blundell, G.M. (2007). Dive behavior of a harbor seal (Phoca vitulina richardii) in the presence of transient killer whales (Orcinus orca) in Glacier Bay National Park, Alaska. Mar. Mammal Sci., 23, 203–208. Wroe, S., McHenry, C., and Thomason, J. (2005). Bite club: comparative bite force in big biting mammals and the prediction of predatory behaviour in fossil taxa. Proc. Biol. Sci., 272, 619–25. Ylitalo, G.M., Matkin, C.O., Buzitis, J., Krahn, M.M., Jones, L. L., Rowles, T., and Stein, J.E. (2001). Influence of   85 life-history parameters on organochlorine concentrations in free-ranging killer whales (Orcinus orca) from Prince William Sound, AK. Sci. Total Environ., 281, 183–203. Young, M.T., Brusatte, S.L., de Andrade, M.B., Desojo, J.B., Beatty, B.L., Steel, L., Fernández, M.S., Sakamoto, M., Ruiz-Omeñaca, J.I., and Schoch, R.R. (2012). The cranial osteology and feeding ecology of the metriorhynchid crocodylomorph genera Dakosaurus and Plesiosuchus from the late Jurassic of Europe. PLoS One, 7, e44985. Yurk, H. (2002). Cultural transmission within maternal lineages: vocal clans in resident killer whales in southern Alaska. Anim. Behav., 63, 1103–1119. Von Ziegesar, O., Ellis, G., Matkin, C.O., and Goodwin, B. (1986). Repeated sightings of identifiable killer whales (Orcinus orca) in Prince William Sound, Alaska, 1977-1983. Cetus, 6, 9–13.                  86 Appendices Appendix A  Specimen information A.1 General location, collection year, ecotype, and ecotype evidence Specimen ID LOG YR ECO Evidence for ecotype assignment SEX CAS 20749 California 1977 offshore tooth wear, stomach contents (shark vertebrae) f CAS 16464 Washington 1968 unk   m CAS 23814 California 1992 unk   f CAS 24294 California 1996 unk   f CAS 5574 California 1926 unk   unk Cowan China Hat 51 British Columbia   resident northern resident (A. Miscampbell) unk Cowan China Hat 124     unk   unk JSKWICS OROR-2     resident northern resident (A. Miscampbell) unk JSKWICS OROR-1     transient transient haplotype (A. Miscampbell) unk KHS British Columbia unk transient transient haplotype (A. Miscampbell) unk LACM 52455 California 1961 offshore offshore haplotype (Morin et al., 2006) unk LACM 72550 California 1985 offshore offshore haplotype (Morin et al., 2006) f LACM 52479 British Columbia 1969 resident vocal dialect A5 (J. Ford in Hoyt, 1990); northern resident (A. Miscampbell) f     87 A.1 General location, collection year, ecotype, and ecotype evidence continued (page 2) Specimen ID LOG YR ECO Evidence for ecotype assignment SEX LACM 52480 British Columbia 1969 resident vocal dialect A5 (J. Ford in Hoyt, 1990); northern resident (A. Miscampbell)  m LACM 72453 captive born 1983 resident northern resident: SWFSC sequenced, from northern resident haplotype (SWFSC and A. Miscampbell)  f LACM 84249 British Columbia 1968 resident vocal dialect A5 (J. Ford in Hoyt, 1990); northern resident (A. Miscampbell)  m LACM 54444 California 1973 transient transient haplotype (Morin et al., 2006); stomach contents partially digested remains of a young sea lion  f LACM 84291 British Columbia 1970 transient M-pod dialect (J.Ford in Hoyt 1990); transient haplotype (A. Miscampbell)  f LACM 30461 California 1927 unk SWFSC unable to sequence (Morin et al., 2006) unk LACM 72577 British Columbia 1985 unk  f LACM 22791 Baja, Mexico 1951 offshore offshore haplotype (Morin et al., 2006) unk MVZ 129686 California 1962 transient transient haplotype (Morin et al., 2006) m MVZ 134462 California 1966 transient transient haplotype (Morin et al., 2006) f   88 A.1 General location, collection year, ecotype, and ecotype evidence continued (page 3) Specimen ID LOG YR ECO Evidence for ecotype assignment SEX NMML 0077 Alaska 1960 resident Southern Resident/ICE/TAAF/Brazil (A. Miscampbell)  m NMML 0078 California 1961 transient transient haplotype (Morin et al., 2006) m NMML 0079 California 1963 transient transient haplotype (Morin et al., 2006) m NMML 0080 California 1964 offshore offshore haplotype (Morin et al., 2006) f NMML 0081 California 1965 transient transient haplotype (Morin et al., 2006) m NMML 0082 California 1966 transient transient haplotype (Morin et al., 2006) m NMML 0083 Washington 1967 resident southern resident haplotype (Morin et al., 2006) m NMML 0084 California 1967 transient transient haplotype (Morin et al., 2006) m NMML 0085 California 1967 transient transient haplotype (Morin et al., 2006) f NMML 0086 Washington 1970 unk SWFSC unable to obtain sequence (Morin et al., 2006) m NMML 0087 California 1966 offshore offshore haplotype (Morin et al., 2006) m NMML 0088 Washington 1967 resident southern resident haplotype (Morin et al., 2006) m NMML 0089 Washington 1967 resident southern resident haplotype (Morin et al., 2006) m NMML 0090 Alaska 1968 resident northern resident (A. Miscampbell) f NMML "NMFS Enforcement"  unk  unk NMML unlabelled   unk  unk NMML1850 Washington 2002 transient field identification CA 189 f      89 A.1 General location, collection year, ecotype, and ecotype evidence continued (page 4) Specimen ID LOG YR ECO Evidence for ecotype assignment SEX RBCM 5214 British Columbia 1945 offshore* southern resident haplotype (A. Miscampbell), tooth wear; stranded with 5319 (offshore haplotype); Ford et al. (2010)*   f RBCM 5319 British Columbia 1945 offshore offshore haplotype (A. Miscampbell) m RBCM 10833 British Columbia 1982 resident southern resident haplotype (or Iceland/TAAF/Brazil) (A. Miscampbell)  f RBCM 16006 British Columbia 1986 resident southern resident haplotype (L. Barrett-Lennard)   f RBCM 16196 British Columbia 1977 resident resident (A. Miscampbell) m RBCM 16630 British Columbia 1986 resident southern resident (A. Miscampbell)  unk RBCM 16639 British Columbia 1987 resident resident (A. Miscampbell) m RBCM 16814 British Columbia 1989 resident field identification L14, southern resident haplotype (A. Miscampbell)  m RBCM 5106 British Columbia 1944 resident resident haplotype (A.Miscampbell) f RBCM 5655 British Columbia 1949 resident northern resident (A. Miscampbell) m RBCM 6721 British Columbia 1949 resident northern resident haplotype (A. Miscampbell) unk RBCM 8386 British Columbia 1973 resident resident haplotype (A. Miscampbell) m RBCM 8861 British Columbia 1975 resident resident haplotype (A. Miscampbell) m RBCM 10001 British Columbia 1979 transient field identification transient pod 01 m RBCM 10402 British Columbia 1981 transient transient haplotype (A. Miscampbell) m      90 A.1 General location, collection year, ecotype, and ecotype evidence continued (page 5) Specimen ID LOG YR ECO Evidence for ecotype assignment SEX RBCM 12844  British Columbia 1976 transient transient haplotype (A. Miscampbell) m RBCM 9716 British Columbia 1977 resident field identification L8 (southern resident) m ROM 62742 Baffin Island, Nunavut, Canada  1969 Baffin  unk SBMNH 979 California 1981 offshore offshore haplotype (Morin et al., 2006) m SBMNH 1546 California 1977 transient transient haplotype (Morin et al., 2006) m SIRS 009808 British Columbia 1997 offshore offshore haplotype (L. Barrett-Lennard); field identification O-120 (G. Ellis); tooth wear  f SIRS N009605d British Columbia 1996 resident SIRS records f SSSC 201108 Alaska 2011 transient stomach contents harbor seal claws and hair (Avery et al., 2011)  m USNMA A13018 California  transient offshore haplotype (Morin et al., 2006) unk USNMA A16487 California  offshore offshore haplotype (Morin et al., 2006) unk USNMA 219326 unk  resident southern resident haplotype (or Iceland/TAAF/Brazil), (A. Miscampbell)  unk USNMA 239357 unk  resident northern resident haplotype (A. Miscampbell) unk      91 A.1 General location, collection year, ecotype, and ecotype evidence continued (page 6) Specimen ID LOG YR ECO Evidence for ecotype assignment SEX USNMA A16488 California  resident northern resident haplotype (Morin et al., 2006) unk USNMA A16625 Washington  resident southern resident haplotype (Morin et al., 2006) unk USNMA A22068 Russia  R/O Res/Offshore (A. Miscampbell) unk USNMA A21330 Russia  Russia  unk USNMA A37166 Alaska 1895 transient transient haplotype (A. Miscampbell) unk USNMA 238112 New Jersey 1909 unk  m USNMA 550857 Massachusetts 1986 unk  unk USNMA 571360 Massachusetts 1989 unk  m USNMA 267617 South Atlantic Ocean  1938 Atlantic  f USNMA 504925 Newfoundland 1971 Atlantic  f USNMA A23004 Norway 1887 Atlantic  unk VAMSC "Hyak" British Columbia; captive   resident A5 pod dialect (J.Ford in Hoyt, 1990) m VAMSC J18 British Columbia 2000 resident photographic identification J18 m VAMSC Moby Doll British Columbia 1964 resident J1 pod dialect m VAMSC Skana Washington; captive 1967 resident K dialect (J. Ford in Hoyt 1990); southern resident haplotype (A. Miscampbell) f   92 A.2 Specimen source or collection locations Specimen ID ecotype sex longitude latitude CAS 20749 offshore female -122.7333 38.0667 Cowan China Hat 51 resident unknown -131.3211111 55.2447222 JSKWICS OROR-1 transient unknown -128.344162 50.765976 Kruzof Island, Sitka Alaska transient male -135.674722 57.170556 LACM 22791 offshore unknown -114.5 28 LACM 52455 offshore unknown -117.9181 33.5906 LACM 52479 resident female -124.0667 49.6333 LACM 52480 resident male -124.0667 49.6333 LACM 54444 transient female -119.7167 34.0167 LACM 72550 offshore female -118.0036 33.6547 LACM 84249 resident male -124.0667 49.6333 LACM 84291 transient female -123.55 48.3333 MVZ 129686 transient male -123.76 37.73 MVZ 134462 transient female -123.1253 38.4467 NMML 0078 transient male -122.9667 37.4833 NMML 0079 transient male -123.0667 37.95 NMML 0080 offshore female -123.5167 33.8333 NMML 0081 transient male -121.2667 34.3833 NMML 0082 transient male -122.75 37      93 A.2 Specimen source or collection location continued (page 2) Specimen ID ecotype sex longitude latitude NMML 0084 transient male -124.0333 37.8667 NMML 0085 transient female -123.75 38.1667 NMML 0087 offshore male -120.9167 35.3167 NMML 0088 resident male -122.5166667 47.53333333 NMML 0089 resident male -122.5166667 47.53333333 NMML 0090 resident female -165.9 54.55 NMML1850; CA 189  transient female -123.1394 48.168604 RBCM 10001 transient male -122.9333 49.0333 RBCM 10402 transient male -125.3833 48.8333 RBCM 10833 resident female -125.3833 50.1167 RBCM 12844 transient male -125.6667 49.15 RBCM 16006 resident female -124.4 48.5667 RBCM 16196 resident male -123.233333 48.416667 RBCM 16630 resident unk -123.4 48.4167 RBCM 16639 resident male -125.5333 48.95 RBCM 16814 resident male -125.7167 49.05 RBCM 5106 resident female -123.55 48.716667 RBCM 5214 resident female -126.55 49.3833      94 A.2 Specimen source or collection location continued (page 3) Specimen ID ecotype sex longitude latitude RBCM 5319  offshore male -126.55 49.3833 RBCM 5655 resident male -125.1833 50.1 RBCM 6721 resident unknown -123.6167 49 RBCM 8386 resident male -126.9333 50.5833 RBCM 8861 resident male -124.266667 49.483333 SBMNH 1546 transient male -120.831667 35.353889 SBMNH 4074 transient unknown -119.716667 34.016667 SBMNH 979 offshore male -119.716667 34.016667 SIRS N009605d resident female -128.4 50.7833 USNMA A37166 transient unknown -169.9340668 56.8848877 VAMSC "Hyak" resident male -124.0667 49.6333 VAMSC J18 resident male -122.9399824 49.0351587 VAMSC Moby Doll resident male -123.1500015 48.7833328 VAMSC Skana resident female -122.5342 47.5308   95 Appendix B  Geometric morphometrics landmarks  B.1 Landmarks on dorsal aspect of cranium   Landmark  number Location   1 medial edge of proximal portion of premaxilla on concavity bordering the external nares 2 medial edge of proximal portion of premaxilla at widest margins of nasal opening 3 posterior-most margin of premaxilla 4 lateral margin of widest part of premaxillae 5 intersection of margin of premaxilla and line across antorbital processes 6 lateral edge of premaxilla at midpoint of rostrum 7 lateral margin of premaxilla, widest point of distal portion  8 anteriormost point of rostrum 9 intersection between frontal sutures and nuchal crest 10 lateral margin of maxilla at widest point of rostrum 11 lateral margin of maxilla at widest portion of proximal rostrum (anterior to antorbital notch) 12 concavity of antorbital notch 13 anteriormost point lateral to antorbital notch 14 lateral edge of maxilla at widest part of facial bones 15 posteriormost edge of maxilla 16 margin of temporal crest at intersection with frontal bones and occipital bones 17 posteriormost margin of temporal crest 18 widest point of skull at squamosal bones 19 posteriormost edge of occipital condyles 20 concavity of intercondyloid notch   96 B.2 Landmarks on labial view of left dentary Landmark  number Location   1 pogonion (anteriormost tip of left dentary) 2 alveolar border at midpoint of dentary (between 1 and 8). 3 dorsal-most point of coronoid process 4 posterior-most point of coronoid process (convex portion) 5 concavity of mandibular notch 6 dorsal constriction of mandibular condyle at mandibular notch (in concavity if present) 7 dorsal point of mandibular condyle 8 posterior-most point of mandibular condyle 9 ventral edge of mandibular condyle 10 constriction (concavity) of mandibular condyle at angular process 11 posterior edge of angular process 12 ventral-most point of angular process 13 ventral margin at midpoint of dentary (between landmarks 1 and 8) 14 gnathion  * The line between landmarks 2 and 13 is perpendicular to the line between landmarks 1 and 8.     97 Appendix C  Two-stage Benjamini & Hochberg step-up false discovery rate - controlling procedure: raw p-values and adjusted p-values Adjusted p-values were calculated for the 21 traditional morphometrics cranial measurements, with a FDR of 0.02. This procedure was repeated with a FDR of 0.05 for reference purposes. Significant adjusted p-values are presented in bold font. The text of the thesis reports the raw p-values. Measurement Rank of raw p Raw p Adjusted p value, FDR=0.02 Adjusted p value, FDR=0.05 LMT/SKL 1 0.0002506 0.0037590 0.0032578 OCW/SKL 2 0.0005456 0.0040920 0.0035464 POS/SKL 3 0.0008538 0.0042690 0.0036998 WAL/SKL 4 0.001239 0.0046463 0.0040268 WAN/SKL 5 0.001597 0.0047910 0.0041522 AMN/SKL 6 0.002705 0.0067625 0.0058608 WRP/SKL 7 0.008568 0.0183600 0.0159120 LLR/SKL 8 0.01049 0.0196688 0.0170463 LPF/SKL 9 0.02248 0.0374667 0.0324711 WOC/SKL 10 0.02831 0.0424650 0.0368030 LSY/SKL 11 0.03664 0.0499636 0.0433018 MAD/SKL 12 0.0548 0.0646154 0.0560000 OCH/SKL 13 0.056 0.0646154 0.0560000 WAP/SKL 14 0.09274 0.0930800 0.0806693 WIN/SKL 15 0.09308 0.0930800 0.0806693 WPF/SKL 16 0.1859 0.1742813 0.1510438 LOC/SKL 17 0.3425 0.3022059 0.2619118 MID/SKL 18 0.6899 0.5749167 0.4982611 LLF/SKL 19 0.7677 0.5995714 0.5196286 FMW/SKL 20 0.8301 0.5995714 0.5196286 FMH/SKL 21 0.8394 0.5995714 0.5196286      98 Appendix D  Geometric morphometrics analysis for dorsal skull D.1 Eigenvalues and percent variance (dorsal skull landmarks) Principal component Eigenvalues % Variance  Cumulative % 1 0.00085973 22.208 22.208 2 0.00061485 15.882 38.090 3 0.00043561 11.252 49.342 4 0.00037528 9.694 59.036 5 0.00033687 8.702 67.737 6 0.00020073 5.185 72.923 7 0.00017627 4.553 77.476 8 0.00015816 4.085 81.561 9 0.00013337 3.445 85.006 10 0.00008406 2.171 87.178 11 0.00007814 2.018 89.196 12 0.00006689 1.728 90.924 13 0.00005705 1.474 92.398 14 0.00004348 1.123 93.521 15 0.00003646 0.942 94.462 16 0.00003573 0.923 95.385 17 0.0000304 0.785 96.170 18 0.00002848 0.736 96.906 19 0.00002329 0.602 97.508 20 0.00001843 0.476 97.984 21 0.00001558 0.403 98.386 22 0.00001407 0.363 98.750 23 0.00001045 0.270 99.019 24 0.00000963 0.249 99.268 25 0.00000651 0.168 99.436 26 0.00000606 0.157 99.593 27 0.00000492 0.127 99.720 28 0.00000352 0.091 99.811 29 0.00000266 0.069 99.880 30 0.00000177 0.046 99.925 31 0.00000117 0.030 99.956 32 0.00000096 0.025 99.980 33 0.00000048 0.013 99.993 34 0.00000016 0.004 99.997 35 0.00000011 0.003 100   99 D.2 Principal component coefficients (dorsal skull landmarks)  PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9  x1 -0.090687 -0.115885 0.167469 0.104392 0.020899 0.203269 0.022281 0.069722 -0.012977  y1 -0.015509 -0.115111 -0.021954 0.145145 -0.050390 0.044231 0.067993 0.019069 0.108801  x2 -0.202521 -0.072163 0.196315 -0.033297 0.160720 -0.086209 -0.169175 0.085555 -0.060471  y2 0.040937 -0.009103 -0.022635 0.079270 -0.078145 0.063350 0.067103 0.034650 0.069358  x3 -0.318524 0.126137 0.202657 -0.408039 -0.343748 -0.003934 0.211153 -0.166368 0.110919  y3 -0.078997 0.067619 0.110391 -0.245989 -0.367392 0.103515 0.373970 -0.170097 -0.054724  x4 -0.136489 -0.124228 0.054163 0.125402 -0.127270 -0.566701 -0.152283 -0.406903 -0.285900  y4 0.034172 -0.026170 -0.020329 0.006852 0.019760 0.117402 0.044962 0.105812 0.065231  x5 -0.021022 0.062818 0.016329 0.001952 0.120068 0.087392 -0.061106 -0.260262 0.117394  y5 -0.017031 -0.156644 -0.011377 0.124624 -0.024437 0.105342 0.051272 0.098039 -0.056562  x6 0.101871 0.137964 -0.053231 0.170438 0.020930 -0.035857 0.198942 0.035273 0.036375  y6 -0.000570 -0.102735 -0.065597 0.065045 -0.029787 0.057843 0.045601 -0.029807 -0.059265  x7 -0.112640 -0.197688 -0.114689 -0.031061 -0.487940 0.040465 -0.334372 0.200133 0.273900  y7 0.060171 -0.038223 -0.018450 -0.055247 0.055577 -0.082787 -0.030373 -0.142290 -0.096872  x8 0.190256 0.144157 -0.154005 0.037810 -0.062870 -0.032028 0.228012 0.042947 -0.248467  y8 -0.034202 -0.153798 -0.066205 0.024935 -0.015225 0.068945 0.005059 -0.069419 -0.202950  x9 -0.133134 0.035006 -0.040519 0.087727 0.154863 -0.408670 0.085744 0.342492 0.177324  y9 0.096325 0.028467 -0.294316 0.021190 -0.219453 -0.418598 -0.022694 0.107896 0.194182  x10 0.109143 0.149986 -0.065135 0.188033 0.070864 -0.078713 0.206164 0.051745 0.013742  y10 0.021388 -0.019182 -0.004142 -0.068677 0.077333 -0.021585 -0.022150 -0.121353 0.135170  x11 -0.058526 0.004976 0.037435 -0.034061 0.160248 0.107597 -0.097022 -0.239832 0.014009  y11 0.016870 0.018041 0.025802 -0.018927 0.010011 -0.095863 -0.107721 0.055742 0.037115  x12 -0.074267 -0.064109 0.018914 0.038106 0.037860 0.207239 -0.083897 0.016426 0.071468  y12 0.093158 0.051304 -0.014603 0.000025 0.019550 -0.003926 -0.058483 -0.007280 0.094560  x13 -0.073457 -0.017523 0.087054 -0.015947 0.123177 0.103500 -0.175795 0.039808 0.000099  y13 -0.020863 0.084663 0.025104 -0.086130 0.025496 -0.012921 -0.097123 -0.041704 0.141780    100 D.2. Principal component coefficients (dorsal skull landmarks) continued, page 2  PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9  x14 0.627808 -0.531854 0.332467 -0.269212 0.055855 -0.113094 0.072482 0.007796 0.005068  y14 0.054504 0.188034 0.112651 -0.201428 0.176516 -0.036140 -0.113150 0.082478 0.066526  x15 -0.083743 -0.033841 0.073085 0.039117 0.027863 0.129048 -0.071812 0.140666 -0.122277  y15 -0.062372 0.103444 0.007088 -0.007918 0.030978 0.052007 0.047530 0.169899 -0.434812  x16 -0.076283 0.055771 -0.031689 0.052985 -0.124998 0.064447 -0.304578 0.155656 -0.310209  y16 0.010166 0.051008 0.089167 -0.098684 -0.051131 -0.028170 0.044566 0.219263 -0.393485  x17 0.198830 0.035026 -0.164854 0.365028 -0.199189 0.256101 -0.056325 -0.282433 -0.027596  y17 -0.048149 0.127118 0.201492 0.328904 0.119684 -0.104063 0.086516 -0.279377 0.074633  x18 -0.084902 -0.186641 -0.704729 -0.373408 0.272706 0.059531 0.120035 -0.119342 -0.068217  y18 -0.104877 0.170861 0.061578 -0.209881 0.257885 -0.015905 -0.211028 -0.040970 0.017959  x19 0.274526 0.376662 0.017384 -0.076914 0.022444 0.070295 -0.012507 -0.033131 0.119118  y19 -0.046478 -0.146228 -0.053023 0.098361 0.060181 0.050359 -0.026917 -0.052336 0.112894  x20 0.235815 0.367421 0.013938 -0.108114 -0.051897 -0.003317 -0.081118 0.110027 0.054411  y20 -0.077781 -0.098764 0.001352 0.032628 0.111252 0.083516 -0.043352 -0.096170 0.155418  x21 -0.272052 -0.151993 0.111643 0.139065 0.149415 -0.000360 0.455176 0.210025 0.142288  y21 0.079138 -0.024603 -0.041995 0.065901 -0.128263 0.073448 -0.101583 0.157955 0.025045      101 D.2. Principal component coefficients (dorsal skull landmarks) continued, page 3  PC10 PC11 PC12 PC13 PC14 PC15 PC16 PC17 PC18  x1 -0.077266 0.138260 -0.538668 -0.046053 0.051052 -0.321145 -0.337882 -0.140356 0.007626  y1 -0.078672 0.040784 -0.007607 -0.027536 0.134722 0.06089 0.163966 0.030094 -0.151204  x2 -0.062740 0.014282 -0.005972 -0.383931 -0.271155 -0.064637 0.319198 -0.235083 -0.009666  y2 -0.123126 -0.094305 0.081449 0.080754 0.070664 0.092848 0.049553 -0.079771 -0.223475  x3 0.215508 0.011685 0.163914 0.021380 0.233387 0.142577 -0.208109 -0.250085 0.148954  y3 0.022761 -0.168017 -0.162974 -0.059528 -0.165111 -0.234556 0.402403 0.147562 -0.184351  x4 -0.193622 0.137365 -0.142824 -0.069480 0.039644 0.255177 0.022349 0.012931 -0.116865  y4 -0.123511 -0.060258 0.088921 -0.102191 0.002764 0.195049 -0.003583 -0.038756 -0.003798  x5 0.032580 -0.037772 0.114169 -0.098982 0.029009 -0.101705 -0.099481 0.312011 0.334254  y5 -0.032751 0.013839 0.116386 -0.134237 0.132799 0.316657 0.060157 -0.180568 -0.149041  x6 0.300334 -0.057425 -0.194685 0.031120 -0.082236 0.217155 -0.008380 -0.160569 0.091432  y6 0.049238 0.041412 0.020108 0.101861 -0.066516 0.002819 0.034936 0.027297 0.092935  x7 -0.117604 0.246925 0.010645 0.275443 -0.063542 -0.014229 0.029969 0.135644 -0.050994  y7 0.012954 -0.033480 -0.009647 0.095703 -0.171155 -0.147388 0.088485 0.092139 0.199142  x8 -0.209568 -0.063795 0.195621 -0.154606 0.154538 -0.168970 0.093146 0.103330 0.085276  y8 0.202684 0.139120 0.116442 0.109355 -0.197624 0.039608 -0.155719 -0.080411 -0.024363  x9 0.373045 -0.004478 0.183629 0.086101 0.220534 -0.294457 0.064132 0.163154 -0.290939  y9 -0.078414 -0.135065 0.021837 -0.276987 -0.141207 -0.135338 -0.289586 -0.119509 0.260669  x10 0.262611 -0.014323 -0.195560 0.028102 -0.041712 0.237370 0.139349 -0.105748 0.073406  y10 -0.036824 -0.177584 0.045648 -0.075088 0.050595 -0.188616 -0.109333 -0.051865 0.127286  x11 -0.021093 -0.015358 0.350962 -0.170694 0.048053 -0.000521 -0.148881 0.240118 -0.284456  y11 -0.012531 -0.001052 0.020953 0.144608 0.079062 -0.037513 0.155131 -0.013862 0.010428  x12 -0.018759 -0.084710 0.096565 -0.088798 -0.133840 -0.018773 -0.091059 -0.085547 -0.109716  y12 -0.088050 -0.050483 -0.063430 -0.121040 0.244741 -0.019997 0.127709 0.014637 0.016281  x13 0.103472 -0.044915 0.253167 0.028376 -0.224216 -0.113827 0.149217 -0.370918 0.005047  y13 0.044605 -0.167313 -0.167405 -0.011359 0.073465 0.168343 -0.143834 0.025210 -0.145585   102 D.2. Principal component coefficients (dorsal skull landmarks) continued, page 4  PC10 PC11 PC12 PC13 PC14 PC15 PC16 PC17 PC18  x14 0.050782 -0.016735 0.001177 0.101488 0.103657 -0.024701 -0.003761 -0.016668 0.018412  y14 -0.064082 -0.125797 -0.151937 0.039902 0.074359 0.108961 -0.245800 -0.033157 -0.256145  x15 -0.095841 -0.391774 0.043309 -0.005454 0.122045 0.193364 0.001205 0.087782 0.242113  y15 -0.087043 0.381657 0.157573 0.084625 0.380890 -0.129060 -0.056020 -0.173276 0.171542  x16 0.035525 -0.427918 -0.149455 0.289966 0.099371 0.009819 0.095294 0.159075 0.102672  y16 0.212498 0.056199 0.040355 -0.051618 -0.401602 0.029523 -0.317191 0.299999 -0.156358  x17 0.219642 0.079479 -0.071647 -0.190390 0.153337 -0.170636 -0.064393 0.011441 -0.206260  y17 -0.191937 -0.137075 0.112754 0.487930 -0.141952 -0.208985 -0.088570 -0.198906 -0.048123  x18 -0.100401 -0.026008 -0.113432 0.177433 -0.047408 -0.015357 0.005668 -0.139637 -0.124474  y18 0.057626 0.187373 -0.293417 -0.107499 0.147652 -0.044047 0.155792 0.135470 0.000383  x19 -0.210581 0.206795 0.064997 0.119898 -0.153626 0.200165 -0.024986 0.116119 -0.054386  y19 0.166222 0.129096 0.038370 -0.031168 -0.069117 0.235065 0.054709 0.200350 0.220401  x20 -0.123741 0.247584 0.001108 0.044165 -0.148203 -0.012451 0.136596 -0.085537 0.061590  y20 0.216232 0.203883 -0.008893 0.082563 0.016938 -0.007803 0.147359 0.094105 0.238316  x21 -0.362283 0.102836 -0.067022 0.004915 -0.088688 0.065784 -0.069191 0.248545 0.076973  y21 -0.067878 -0.042935 0.004511 -0.229052 -0.054367 -0.096460 -0.030567 -0.096783 0.005061                 103 D.2. Principal component coefficients (dorsal skull landmarks) continued, page 5  PC20 PC21 PC22 PC23 PC24 PC25 PC26 PC27 PC28  x1 0.302546 -0.102419 -0.124078 -0.074184 0.084895 0.095164 -0.003067 -0.109685 -0.004490  y1 0.220283 0.521511 -0.142477 0.312524 -0.199744 0.033277 -0.029789 0.215752 -0.144717  x2 -0.315041 -0.086163 -0.283532 -0.231963 -0.151033 0.158732 0.112729 0.085176 0.040672  y2 0.115283 -0.313807 -0.239475 -0.116591 0.070511 -0.054960 -0.073683 0.013956 -0.105954  x3 -0.077017 -0.138798 -0.157939 0.223436 0.019296 0.024283 -0.074082 0.026135 0.006511  y3 0.122507 0.098322 0.183761 -0.216709 -0.028787 0.051702 0.078808 -0.000813 0.002593  x4 0.108040 0.109977 0.055500 0.041077 0.126154 -0.090896 -0.057692 -0.136051 -0.078110  y4 0.046720 -0.144230 -0.090386 -0.220457 0.290168 -0.186857 -0.273594 -0.259425 -0.337134  x5 0.058619 0.239700 -0.230751 -0.178114 -0.264201 0.047533 -0.325126 -0.110958 0.120815  y5 0.058433 -0.187553 0.125926 -0.028237 -0.473468 -0.068127 -0.026081 -0.129650 0.314375  x6 0.083779 0.044834 0.131714 0.003398 -0.076332 -0.077851 -0.010654 0.028299 -0.097311  y6 0.048834 -0.042004 -0.085207 0.055291 0.035029 0.055977 -0.040631 -0.121114 0.397337  x7 -0.211707 -0.007374 0.064862 -0.060740 -0.100515 0.038848 0.047700 -0.040108 -0.072004  y7 -0.013594 -0.157944 -0.120046 0.200219 -0.043542 0.209961 -0.132213 0.199488 -0.260678  x8 -0.103324 -0.119484 -0.156953 0.115863 0.183949 -0.143678 0.125674 0.153383 0.057794  y8 -0.049301 0.003022 0.118464 -0.126157 0.094602 -0.132027 0.081001 0.333181 -0.133720  x9 0.011128 -0.001505 -0.013941 -0.093615 0.032805 -0.013746 -0.117105 -0.025208 -0.041424  y9 0.150855 -0.092292 0.177673 -0.084982 -0.076346 0.050484 0.074024 0.200881 0.143887  x10 0.005664 -0.009981 -0.114227 -0.003109 -0.086208 0.109249 0.092244 -0.073057 -0.028625  y10 -0.027537 0.094697 0.201850 -0.039256 -0.128739 -0.205415 0.379189 -0.423966 -0.163901  x11 0.301202 -0.159501 -0.003666 -0.030552 -0.043084 0.017803 0.256573 0.029386 -0.064097  y11 0.072316 -0.008847 -0.130704 0.121956 0.110181 0.156467 0.086505 -0.324512 0.207433  x12 0.170096 0.020121 0.215566 -0.019956 -0.109093 -0.043436 -0.310921 0.264746 -0.023111  y12 -0.070079 -0.069976 0.305057 0.038540 0.157222 0.402035 -0.349762 -0.054365 -0.020960  x13 0.189415 0.199677 0.244802 0.200667 0.381769 0.063304 0.035166 -0.099576 0.128074  y13 -0.178967 0.335281 -0.151767 -0.385424 0.315351 -0.183849 -0.018918 0.194202 0.280709    104 D.2. Principal component coefficients (dorsal skull landmarks) continued, page 6  PC20 PC21 PC22 PC23 PC24 PC25 PC26 PC27 PC28  x14 -0.044280 -0.013430 0.018022 -0.034808 -0.025987 -0.055107 -0.022838 0.071417 0.051029  y14 -0.114490 0.067863 -0.057860 0.141075 -0.131959 0.279712 0.292758 0.090219 -0.195131  x15 -0.276826 0.148362 0.233404 0.024392 0.002785 0.018177 0.000874 -0.041931 -0.155514  y15 0.012748 0.202995 0.074768 -0.281292 -0.056899 0.130632 0.059783 0.076777 -0.030921  x16 0.240538 -0.157703 -0.076186 -0.012379 -0.044777 -0.086411 0.115840 0.118119 0.085244  y16 -0.129527 0.056793 0.006837 0.133883 -0.062675 0.020083 -0.174215 -0.278430 0.041228  x17 -0.386583 -0.048407 0.081334 0.073499 0.061408 0.058481 0.068634 -0.051738 -0.007962  y17 -0.223061 -0.086440 -0.011864 -0.019304 -0.134615 -0.046175 -0.121219 0.035020 0.027420  x18 -0.088595 0.042974 -0.068843 0.012847 -0.048662 0.035146 -0.073013 -0.096371 0.006144  y18 -0.082366 -0.267844 0.258888 0.213196 -0.041117 -0.324482 -0.136620 0.162613 0.098771  x19 0.057933 -0.001553 0.058031 0.004747 0.099901 0.147470 0.094835 0.064815 0.289157  y19 0.091253 -0.126439 -0.055511 -0.107142 0.181984 0.316932 0.169799 0.049192 -0.161625  x20 0.093508 0.056648 -0.015386 -0.097909 -0.153205 -0.237714 -0.067277 -0.110621 -0.278194  y20 -0.007086 -0.003344 -0.046586 0.037641 0.031393 -0.292284 0.203277 0.015305 0.029462  x21 -0.119095 -0.015975 0.142269 0.137404 0.110133 -0.065350 0.111503 0.053828 0.065401  y21 -0.043226 0.120236 -0.321340 0.371225 0.091449 -0.213085 -0.048419 0.005689 0.011525                  105 D.2. Principal component coefficients (dorsal skull landmarks) continued, page 7  PC30 PC31 PC32 PC33 PC34 PC35  x1 -0.082686 0.031598 -0.008223 0.040423 -0.019484 -0.040919  y1 0.105835 0.144701 -0.264972 -0.069363 0.031289 0.025055  x2 0.084400 0.081939 -0.125982 -0.136794 0.176694 0.002944  y2 -0.004776 0.341169 0.005088 -0.186973 -0.080502 -0.463533  x3 0.074718 -0.002984 -0.037862 -0.021535 0.029527 0.046809  y3 -0.164959 -0.068651 0.018092 0.054295 -0.040756 0.031455  x4 -0.087404 -0.068999 0.125864 -0.044178 0.049610 -0.009462  y4 -0.066988 -0.137131 -0.205614 0.129182 0.109425 0.327367  x5 -0.244296 -0.094493 -0.120214 0.033074 -0.029518 -0.186812  y5 -0.036901 -0.288140 0.080194 0.315197 -0.042668 -0.127305  x6 0.059731 0.023602 -0.111406 -0.084592 0.315413 -0.362425  y6 -0.018347 -0.051545 0.173733 -0.568991 -0.257896 0.151901  x7 0.044779 -0.024940 -0.077612 0.002398 -0.026295 0.006537  y7 0.273694 -0.045609 0.133302 0.407133 -0.059621 -0.046837  x8 -0.127602 -0.110729 0.131733 -0.029227 0.119196 -0.056671  y8 -0.317416 0.081672 -0.202307 -0.096981 0.019061 0.023540  x9 -0.120856 -0.141328 0.014949 -0.013495 0.026970 0.014418  y9 -0.016071 0.154432 -0.058478 0.072363 -0.109762 0.023712  x10 0.101875 0.135345 -0.035715 0.038266 -0.307532 0.464235  y10 0.178991 -0.075061 -0.111681 -0.151828 0.184920 0.058859  x11 0.248192 0.190546 -0.062070 -0.009110 -0.200750 0.092980  y11 -0.286526 0.456330 0.077814 0.333777 0.164432 0.067480  x12 -0.086083 0.094111 0.435584 -0.035622 0.273314 0.327078  y12 0.289182 -0.012955 0.040873 -0.263666 0.143550 -0.046002  x13 0.016604 -0.161567 -0.058369 0.130860 -0.209310 -0.220943  y13 0.281706 -0.021052 0.125907 0.188201 -0.124883 -0.061782    106 D.2. Principal component coefficients (dorsal skull landmarks) continued, page 8  PC30 PC31 PC32 PC33 PC34 PC35  x14 0.024092 0.001406 -0.060247 -0.008413 -0.030562 0.028052  y14 -0.282508 -0.231572 0.249705 -0.075674 0.001569 -0.032403  x15 -0.248667 0.311315 0.053382 -0.021396 -0.218285 -0.047745  y15 0.112197 0.027342 -0.027904 0.044318 0.001689 0.056016  x16 0.185838 -0.187489 -0.118321 -0.023524 0.132227 0.059267  y16 0.185568 0.115318 -0.010629 -0.011418 0.068807 -0.078728  x17 0.012129 0.007782 0.009139 0.139071 -0.044151 -0.014720  y17 0.001755 -0.074572 -0.099996 -0.019233 -0.032132 0.086055  x18 0.016676 -0.001534 -0.026272 0.004211 0.018116 -0.019676  y18 -0.120696 0.059113 -0.280857 -0.004012 -0.181314 0.017113  x19 -0.063819 -0.012044 -0.237695 0.013952 0.313947 0.095702  y19 -0.140529 -0.300136 0.002268 -0.037314 -0.072754 -0.079520  x20 0.055529 -0.029310 0.283208 -0.065457 -0.307973 -0.118832  y20 0.098772 0.132774 0.397372 -0.005583 0.288638 -0.018409  x21 0.136850 -0.042227 0.026131 0.091089 -0.061156 -0.059818  y21 -0.071985 -0.206430 -0.041910 -0.053431 -0.011092 0.085967     107 D.3 Canonical coefficients for Procrustes coordinates of the dorsal skull.  CV1 CV2   CV1 CV2  x1 56.9253 -31.0473   x12 137.3714 448.1557  y1 -45.3443 -80.6401   y12 -140.1966 -81.0003  x2 0.2885 58.4551   x13 201.6226 -308.2658  y2 238.6277 -645.5028   y13 274.7760 -263.2403  x3 -82.9681 13.3622   x14 -37.7304 8.3790  y3 -9.2970 97.5246   y14 -169.0925 257.2751  x4 -0.7084 57.5459   x15 -67.1948 -374.1724  y4 -618.8565 512.7917   y15 -69.6568 108.5532  x5 230.3296 -13.0453   x16 -20.6249 348.3921  y5 392.3598 0.1954   y16 237.6172 -293.2921  x6 248.4238 -109.1352   x17 62.9214 -102.2127  y6 193.1733 -142.8486   y17 -138.2524 149.0296  x7 18.2560 22.4463   x18 44.8130 -37.0885  y7 -249.5322 -144.4110   y18 -219.7707 35.4834  x8 -96.7618 250.7932   x19 -367.1541 347.6828  y8 -98.1246 377.6493   y19 2.4856 284.0026  x9 -59.0284 126.9733   x20 280.7832 -324.1651  y9 49.9151 24.0153   y20 280.9363 -153.9103  x10 -291.6192 -104.5624   x21 -42.4397 -44.8754  y10 67.8579 123.7697   y21 -220.0129 181.0278  x11 -215.5051 -233.6154      y11 240.3876 -346.4722         108 Appendix E  Geometric morphometrics analysis of left dentary E.1 Eigenvalues and percent variance (left dentary landmarks) Principal component Eigenvalues % Variance  Cumulative % 1 0.00046641 27.908 27.908 2 0.00036042 21.566 49.473 3 0.00025002 14.96 64.433 4 0.0001647 9.855 74.288 5 0.00010869 6.503 80.792 6 0.00007615 4.557 85.348 7 0.00004873 2.915 88.264 8 0.00004071 2.436 90.700 9 0.00003429 2.052 92.752 10 0.00002675 1.601 94.352 11 0.0000226 1.352 95.705 12 0.00001463 0.875 96.580 13 0.00001314 0.786 97.366 14 0.00001051 0.629 97.995 15 0.00000952 0.57 98.564 16 0.00000628 0.376 98.940 17 0.000005 0.299 99.240 18 0.00000413 0.247 99.487 19 0.00000358 0.214 99.701 20 0.00000238 0.142 99.843 21 0.00000151 0.09 99.934 22 0.00000058 0.035 99.969 23 0.00000041 0.025 99.993 24 0.00000011 0.007 100.000      109 E.2 Principal component coefficients (left dentary landmarks)    PC1   PC2   PC3   PC4   PC5   PC6   x1 0.035088 -0.141020 0.253752 0.018723 0.049645 0.090712  y1 0.241247 0.174901 0.297205 0.237419 0.117860 -0.470451  x2 0.038583 -0.241041 0.079329 0.046416 0.028884 -0.033117  y2 -0.186577 -0.090982 -0.115310 -0.184499 0.351455 0.070974  x3 -0.223431 0.468997 0.045560 0.109176 -0.278998 0.205706  y3 -0.177243 -0.098900 0.127003 -0.356283 -0.134317 -0.044005  x4 -0.269149 0.264790 0.173780 0.026418 -0.106649 0.168491  y4 -0.149150 -0.054344 0.118016 -0.305416 -0.139859 -0.128553  x5 -0.278501 0.074687 0.211841 -0.001488 -0.039249 0.077169  y5 0.012443 -0.019425 0.112174 -0.205406 -0.007096 -0.223184  x6 -0.180547 -0.161502 0.082118 -0.000291 0.031126 -0.107624  y6 0.040942 -0.039267 0.328876 -0.061039 -0.075847 -0.019140  x7 -0.192280 -0.294341 -0.134344 0.052845 0.098177 -0.292509  y7 0.140433 -0.046986 0.223361 -0.015232 0.035572 0.264954  x8 0.053403 -0.365716 -0.095161 0.129482 0.014118 0.045461  y8 0.161242 0.076991 0.100360 0.071828 0.180057 0.306037  x9 0.150572 -0.055725 -0.094233 0.245700 0.104354 0.047874  y9 0.129477 0.202687 0.016973 0.202359 0.250404 0.085737  x10 0.147915 0.074676 0.010522 0.120130 0.084968 0.108791  y10 0.042905 0.034639 -0.030705 0.207587 0.063229 -0.112345  x11 0.288496 0.217624 -0.172407 -0.114704 0.093150 -0.039323  y11 -0.075748 0.008073 -0.255848 0.376993 -0.298714 -0.234920  x12 0.418704 0.148647 -0.239519 -0.374623 -0.114867 -0.152793  y12 -0.164138 -0.048401 -0.269801 0.233468 -0.369042 -0.037875  x13 0.023536 -0.255434 0.040035 0.012722 0.052872 0.110788  y13 -0.304142 0.116113 -0.448627 -0.081902 0.428532 0.153004  x14 -0.012389 0.265358 -0.161273 -0.270506 -0.017532 -0.229626  y14 0.288310 -0.215098 -0.203677 -0.119875 -0.402233 0.389768      110 E.2 Principal component coefficients (left dentary landmarks) continued, page 2    PC7   PC8   PC9   PC10   PC11   PC12   x1 0.285926 0.261006 -0.082370 0.139775 -0.119076 0.100228  y1 0.162262 -0.109912 0.023370 -0.077243 -0.015314 0.073966  x2 0.134669 -0.030017 0.058965 -0.128698 0.049957 -0.131437  y2 0.023251 -0.226881 -0.043099 -0.034871 0.029459 0.278082  x3 -0.093841 0.094067 0.129849 -0.193664 -0.298318 0.008810  y3 0.144923 0.175178 0.086531 -0.090441 0.216839 0.212747  x4 0.143534 0.042799 -0.141912 -0.211568 0.212369 0.035232  y4 0.058777 0.243420 0.169639 0.122197 0.090728 0.180888  x5 -0.055224 -0.165346 -0.167399 0.180888 0.228813 -0.265854  y5 -0.052123 0.185287 0.221701 -0.005476 -0.206077 -0.556744  x6 -0.134882 -0.186374 -0.332745 0.416905 0.076004 -0.158058  y6 -0.189093 -0.223766 0.130504 0.009563 -0.314118 -0.116603  x7 -0.311847 0.300512 -0.298533 -0.109581 -0.382054 0.302325  y7 -0.011061 -0.388597 0.016301 0.090006 -0.065236 0.292105  x8 -0.006477 -0.267200 0.256410 -0.392761 0.048686 -0.105685  y8 -0.095352 -0.028302 -0.280270 -0.036648 -0.024820 -0.000836  x9 -0.174026 0.073746 0.349958 0.065851 0.176038 0.098052  y9 -0.223118 0.230957 -0.207067 -0.111895 -0.107117 -0.088015  x10 0.097950 0.190387 0.330549 0.283701 -0.000103 0.216178  y10 -0.075822 0.231892 -0.104918 -0.266481 0.530111 -0.041243  x11 0.232034 0.042534 0.001871 0.276840 -0.115145 0.022645  y11 0.015109 -0.086100 -0.010212 0.375489 0.134088 -0.029531  x12 0.210462 -0.113948 -0.344839 -0.179373 0.029626 -0.066552  y12 0.331628 -0.190090 -0.072967 -0.150014 -0.248020 0.105894  x13 0.139204 0.016977 0.041182 -0.100600 -0.021824 -0.195664  y13 0.187144 0.007560 0.139428 0.072821 -0.078719 -0.240635  x14 -0.467481 -0.259142 0.199014 -0.047714 0.115027 0.139780  y14 -0.276526 0.179354 -0.068941 0.102992 0.058195 -0.070076      111 E.2 Principal component coefficients (left dentary landmarks) continued, page 3    PC13   PC14   PC15   PC16   PC17   PC18   x1 -0.049736 0.065670 0.183236 0.020352 0.152213 0.069977  y1 0.176613 -0.006284 -0.033472 -0.028896 -0.056236 0.071100  x2 -0.037142 -0.105709 -0.045558 0.055618 -0.008376 -0.020824  y2 -0.431690 -0.088224 -0.099532 0.207094 0.064661 -0.042845  x3 -0.094495 -0.131439 0.114802 0.221278 0.274203 0.327663  y3 0.194805 -0.153898 -0.011831 0.131126 0.112761 -0.086479  x4 0.064387 -0.065782 0.045586 -0.185175 -0.029885 -0.509040  y4 0.262826 0.096630 0.011272 0.038235 -0.118009 0.302228  x5 -0.089905 0.141809 -0.020531 -0.236198 -0.383804 0.102649  y5 -0.067157 -0.024674 -0.102499 -0.095867 0.061933 -0.268627  x6 0.142523 0.146704 -0.195642 0.378843 0.337365 0.080116  y6 -0.348330 0.111785 0.138660 -0.083627 0.005219 0.110066  x7 -0.044647 -0.062945 0.109536 -0.283015 -0.097362 -0.056649  y7 0.052510 0.074018 0.276530 -0.122218 0.119846 -0.259215  x8 0.165375 -0.279997 -0.010204 0.181311 -0.010278 0.070812  y8 0.292222 -0.309346 -0.229376 -0.274333 -0.112282 0.394648  x9 0.206877 0.452244 0.150382 -0.203344 0.266010 0.007704  y9 0.200841 0.121205 -0.008716 0.484782 -0.135053 -0.310273  x10 -0.260224 0.008920 -0.182573 0.221188 -0.507119 -0.003913  y10 -0.438049 0.091660 -0.074956 -0.097540 0.224484 0.179538  x11 -0.073792 -0.265934 -0.328241 -0.244746 0.345900 -0.128770  y11 -0.024057 -0.422437 0.378307 0.043676 -0.066359 -0.072913  x12 -0.059729 0.189658 0.363122 0.124542 -0.145761 0.126488  y12 0.038603 0.394186 -0.429157 -0.053514 -0.061891 -0.033754  x13 0.021010 -0.065695 -0.005623 -0.004794 -0.100535 -0.001768  y13 0.135940 0.101526 0.251697 -0.119182 -0.023142 0.103589  x14 0.109498 -0.027504 -0.178289 -0.045859 -0.092570 -0.064446  y14 -0.045076 0.013853 -0.066926 -0.029737 -0.015933 -0.087061      112 E.2 Principal component coefficients (left dentary landmarks) continued, page 4    PC19   PC20   PC21   PC22   PC23   PC24   x1 -0.093172 0.199360 0.130936 -0.331148 -0.194452 -0.030201  y1 0.038007 -0.008459 -0.025332 -0.011823 0.070176 0.108679  x2 -0.045268 -0.261664 -0.068382 0.418413 0.192530 -0.675518  y2 -0.510746 0.073597 0.018372 0.090503 -0.072075 0.074006  x3 -0.090886 -0.067969 -0.164440 -0.031851 0.178540 0.004857  y3 0.125806 0.004165 0.418887 -0.127593 0.487216 0.029071  x4 0.092331 0.258536 0.012624 0.276618 -0.318807 0.022105  y4 -0.099933 -0.242895 -0.286007 0.157827 -0.478482 -0.012008  x5 -0.220813 -0.320033 0.125475 -0.372259 0.100411 -0.039324  y5 -0.242359 0.315379 -0.268408 -0.152448 0.092134 -0.049425  x6 0.233307 0.208117 -0.106276 0.093725 0.005565 0.025017  y6 0.274648 -0.024133 0.463600 0.255885 -0.205744 0.034086  x7 0.090769 -0.028276 -0.066047 -0.051448 0.077228 -0.020619  y7 0.146351 -0.129304 -0.465453 -0.195900 0.165362 -0.067657  x8 0.086534 0.049568 0.062941 -0.378626 -0.337374 -0.005685  y8 -0.117105 0.321012 0.040895 0.127419 0.069267 -0.064886  x9 -0.340905 0.105684 0.203686 0.140674 0.088661 0.020996  y9 -0.074612 -0.310915 0.154053 -0.059806 -0.095438 -0.016021  x10 0.235089 0.258309 -0.091096 -0.005402 0.115813 -0.039778  y10 0.256025 -0.040745 -0.176630 -0.103402 -0.016640 0.023371  x11 0.046698 -0.367043 0.128125 -0.083000 -0.154002 0.027679  y11 -0.149989 0.037328 0.065494 0.077072 -0.015010 0.031190  x12 -0.062093 0.141106 -0.000222 0.033007 0.035799 -0.019275  y12 -0.030859 0.035112 0.030102 -0.044919 0.029519 -0.015995  x13 -0.004596 -0.224118 -0.158507 0.317356 0.208617 0.705632  y13 0.341862 0.000925 0.021991 -0.002910 0.012842 -0.072505  x14 0.073006 0.048422 -0.008815 -0.026059 0.001473 0.024115  y14 0.042905 -0.031066 0.008435 -0.009904 -0.043126 -0.001906      113 E.3 Canonical coefficients for Procrustes coordinates of the left dentary    CV1   CV2   x1 -36.5625 -102.9993  y1 17.7995 75.2731  x2 -375.5514 52.5881  y2 -125.3041 -175.2244  x3 -39.2882 220.6954  y3 74.4214 -47.1844  x4 89.7238 -320.3756  y4 2.6071 183.6599  x5 101.5354 34.0725  y5 46.5471 -72.9949  x6 -227.5293 95.0839  y6 -383.6105 -274.4855  x7 60.8347 72.7144  y7 183.2015 257.8614  x8 2.1654 -104.8608  y8 -151.6488 22.3484  x9 30.5152 -96.1642  y9 27.5139 -117.1087  x10 -176.9358 84.6611  y10 3.5738 179.7751  x11 182.0594 -174.8040  y11 73.2282 4.1485  x12 -53.3816 93.3585  y12 72.7903 -68.7690  x13 519.9946 169.7639  y13 95.1670 17.1420  x14 -77.5797 -23.7340  y14 63.7136 15.5587         114 Appendix F  Sample photographs of resident, offshore and transient (Bigg's) left dentary bones and skulls (dorsal view). F.1 Offshore left dentary (A) SIRS 0120, female; (B) NMML 0087, male; (C) NMML 0080, female. Scale bars are 30 cm rulers. A    B       C     115 F.2 Resident left dentary (A) NMML 0090, female and (B) RBCM 16814, male.  A        B    116 F.3 Transient left dentary (A) LACM 84291, female and (B) RBCM 10001, male. A         B     117 F.4 Offshore dorsal skull (A) SIRS 0120, female, and (B) NMML 0087, male. A  B    118 F.5 Resident dorsal skull (A) NMML 0090, female and (B) RBCM 16814, male. A           B             119 F.6 Transient dorsal skull (A) MVZ 129686, female and (B) RBCM 10001, male. A            B  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0314137/manifest

Comment

Related Items