QUANTIFYING MORPHOLOGICAL VARIABILITY THROUGH THE LATEST ONTOGENY OF HOPLOSCAPHITES (JELETZKYTES) FROM THE LATE CRETACEOUS WESTERN INTERIOR USING GEOGRAPHIC INFORMATION SYSTEMS AS A MORPHOMETRIC TOOL
Mathew J. Knauss
A Thesis
Submitted to the Graduate College of Bowling Green State University in partial fulfillment of the requirements for the degree of
MASTER OF SCIENCE
August 2013
Committee:
Margaret M. Yacobucci, Advisor
Enrique Gomezdelcampo
Sheila Roberts
© 2013
Mathew J. Knauss
All Rights Reserved iii ABSTRACT
Margaret M. Yacobucci, Advisor
Ammonoids are known for their intraspecific and interspecific morphological variation through ontogeny, particularly in shell shape and ornamentation. Many shell features covary and individual shell elements (e.g., tubercles, ribs, etc.) are difficult to homologize, which make qualitative descriptions and widely-used morphometric tools inappropriate for quantifying these complex morphologies. However, spatial analyses such as those applied in geographic information systems (GIS) allow for quantification and visualization of global shell form. Here, I present a GIS-based methodology in which the variability of complex shell features is assessed in order to evaluate evolutionary patterns in a Cretaceous ammonoid clade.
I applied GIS-based techniques to sister species from the Late Cretaceous Western
Interior Seaway: the ancestral and more variable Hoploscaphites spedeni, and descendant and less variable H. nebrascensis. I created digital models exhibiting the shells’ lateral surfaces using photogrammetric software and imported the reconstructions into a GIS environment. I used the number of discrete aspect patches and the 3D to 2D area ratios of the lateral surface as terrain roughness indices. These 3D analyses exposed the overlapping morphological ranges of H. spedeni and H. nebrascensis, with H. nebrascensis specimens exhibiting similar ornamentation to the most ornate H. spedeni. In order to assess more specific shell characters, I digitized the tubercles (points), ribs (polylines) and shell shape (polygons) of select specimens from photographs for 2D analyses. These 2D analyses revealed that the distribution of ribs and the shape of the body chambers are fairly constrained in at least H. spedeni, and the distribution of tubercles is the most variable feature through ontogeny between both Hoploscaphites species. iv The results of the GIS-based spatial analyses demonstrated that the target for evolutionary change in this clade resides in the macroconch body chamber. Additionally, the results support the hypothesis that H. nebrascensis is a paedomorphic descendant of H. spedeni, derived by means of prolonged tubercle expression through later ontogenetic stages. H. nebrascensis microconchs retain developmental flexibility, and the macroconchs are more constrained. These geospatial analyses not only successfully quantified variability in complex morphologies, but also demonstrated the versatility of this method to address questions related to ontogeny and phylogeny. v ACKNOWLEDGMENTS
I thank P. Gorsevski (Department of Geology, Bowling Green State University) and J.
Haug (the Yale Peabody Museum of Natural History) for technical advice, D. Pavuk (Department of Biology, Bowling Green State University) for advice on the thesis proposal, S. Butts (the Yale
Peabody Museum of Natural History) for access to specimens, and N. Landman (the American
Museum of Natural History) for advice on Hoploscaphites morphology. I would also like to thank E. Gomezdelcampo and S. Roberts (Department of Geology, Bowling Green State
University) for being on my thesis committee and providing guidance for both this research project and the final manuscript. I especially thank my advisor, M.M. Yacobucci, for all her advice and help in obtaining external funding, in conducting this research, and in writing and editing this manuscript. Funding for this research was provided by the Geological Society of
America, the Paleontological Society, and the Department of Geology at Bowling Green State
University. vi
TABLE OF CONTENTS
Page
CHAPTER I. INTRODUCTION ...... 1
Ammonoid Intraspecific Variation ...... 1
Quantifying Intraspecific Variability ...... 4
Heteromorphic Ammonoids and Scaphites ...... 7
Hoploscaphites (Jeletzkytes) spedeni ...... 9
Hoploscaphites (Jeletzkytes) nebrascensis ...... 9
Biostratigraphy ...... 10
Research Objectives ...... 12
CHAPTER II. METHODOLOGY...... 14
Acquisition of Digital Models ...... 14
Digital Models Imported into a GIS Environment ...... 17
3D GIS Methodology ...... 18
Orientation Patch Count ...... 18
Surface-to-Planimetric Area Ratio ...... 19
Generating Coordinate Systems for 2D Methodology...... 20
2D GIS Methodology ...... 21
Rib and Tubercle Spacing ...... 21
Rib Density ...... 22
Tubercle-Rib Spatial Correlation ...... 22
Shell Coiling ...... 23
Body Chamber Shape Distribution ...... 24 vii
Statistical Analyses ...... 24
Sample Size for Statistical Analyses...... 25
CHAPTER III. RESULTS ...... 26
3D GIS Analyses...... 26
Results of OPC Analysis ...... 26
Results of Surface-to-Planimetric Area Ratio Analysis...... 28
2D GIS Analyses...... 30
Results of Rib and Tubercle Spacing Analysis ...... 30
Results of Rib Density Analysis ...... 30
Results of Tubercle-Rib Spatial Correlation Analysis ...... 31
Results of Shell Coiling Analysis ...... 32
Results of Body Chamber Shape Distribution Analysis ...... 33
CHAPTER IV. DISCUSSION...... 35
3D GIS Discussion ...... 35
Interpretation of OPC Analysis ...... 35
Interpretation of Surface-to-Planimetric Area Ratio Analysis ...... 39
2D GIS Discussion ...... 42
Interpretation of Rib and Tubercle Spacing Analysis ...... 42
Interpretation of Rib Density Analysis ...... 44
Interpretation of Tubercle-Rib Spatial Correlation Analysis ...... 46
Interpretation of Shell Coiling Analysis ...... 47
Interpretation of Body Chamber Shape Distribution Analysis ...... 48
Evolutionary Implications ...... 50 viii
GIS as a Morphometric Tool ...... 53
CHAPTER V. CONCLUSIONS ...... 57
CHAPTER VI. FIGURES AND TABLES ...... 59
REFERENCES……...... 98
APPENDIX A. RAW DATA TABLES ...... 105
APPENDIX B. ADDITIONAL FIGURES ...... 116 ix
LIST OF FIGURES/TABLES
Figure/Table Page
1 Figure 1. Example of ammonoid intraspecific variation ...... 59
2 Figure 2. Heterochrony as an evolutionary mechanism...... 60
3 Figure 3. Raup’s parameters ...... 61
4 Figure 4. Example of 3D GIS-based analyses on fossil form ...... 62
5 Figure 5. Generalized scaphitid morphology ...... 63
6 Figure 6. Ammonoid dimorph sizes...... 64
7 Figure 7. Scaphitid shell features ...... 65
8 Figure 8. Dimorphs of Hoploscaphites spedeni ...... 66
9 Figure 9. Dimorphs of Hoploscaphites nebrascensis ...... 67
10 Figure 10. Dimorphs of both Hoploscaphites species ...... 68
11 Figure 11. Intermediate morphology ...... 69
12 Figure 12. Biostratigraphy of the Hoploscaphites taxa ...... 70
13 Figure 13. Materials ...... 71
14 Figure 14. Modified Raup parameters for scaphitids ...... 72
15 Figure 15. Photogrammetric accuracy ...... 73
16 Figure 16. Edited digital model example ...... 74
17 Figure 17. TIN accuracy ...... 75
18 Figure 18. TIN model exhibiting aspect ...... 76
19 Figure 19. 3D:2D area ratio rasters ...... 77
20 Figure 20. Vectorization of ammonoid features ...... 78
21 Figure 21. Transect and buffer vectors ...... 79 x
22 Figure 22. Dispersed, random and clustered features ...... 80
23 Figure 23. Rib density transect and map ...... 81
24 Figure 24. Tubercle buffers and intersecting ribs ...... 82
25 Figure 25. Separated body chamber and whole shell polygons ...... 83
26 Figure 26. Landmark-scaled body chambers ...... 84
27 Figure 27. OPC vs. LMAX ...... 85
28 Figure 28. OPC results 1 ...... 86
29 Figure 29. OPC results 2 ...... 87
30 Figure 30. Morphotype bins ...... 88
31 Table 1. Results of Mann-Whitney U Test comparisons for surface-to-planimetric area
ratio medians ...... 89
32 Figure 31. 3D:2D area median vs. OPC ...... 90
33 Figure 32. 3D:2D area median vs. LMAX ...... 91
34 Table 2. Results of Average Nearest Neighbor analysis for rib spacing and tubercle
spacing ...... 92
35 Figure 33. Rib density spiral transect results ...... 93
36 Figure 34. Rib density map results ...... 94
37 Figure 35. Shell coiling analysis results ...... 95
38 Figure 36. Landmark-scaled body chamber results 1 ...... 96
39 Figure 37. Landmark-scaled body chamber results 2 ...... 97
40 Table A1. Orientation patch count (OPC) for whole shells, body chambers and
phragmocones of both Hoploscaphites species...... 105
41 Table A2. Whole shell surface-to-planimetric (3D:2D) area ratio mean, standard xi
deviation (STD) and median cell values...... 106
42 Table A3. Body Chamber surface-to-planimetric (3D:2D) area ratio mean, standard
deviation (STD) and median cell values...... 107
43 Table A4. Phragmocone surface-to-planimetric (3D:2D) area ratio mean, standard
deviation (STD) and median cell value...... 108
44 Table A5. Results of Average Nearest Neighbor analysis of rib spacing on the left
lateral surface of ammonoids...... 109
45 Table A6. Results of Average Nearest Neighbor analysis of tubercle spacing on the
left lateral surface of ammonoids...... 110
46 Table A7. Rib density calculated by the number of ribs that intersect spiral transects
at 25%, 50% and 75% whorl height on the left lateral surface of
ammonoids...... 111
47 Table A8. Percentage of tubercles with intersecting ribs at 1 mm and 2 mm buffers
on the let lateral surface of ammonoids...... 112
48 Table A9. Body chamber to whole shell area ratios used in the shell coiling analysis
on the left lateral surface of ammonoids...... 113
49 Table A10. Body chamber area data from the first landmark pair used in the body
chamber shape distribtution analysis of the left lateral surface of
ammonoids...... 114
50 Table A11. Body chamber area data from the second landmark pair used in the body
chamber shape distribution analysis of the left lateral surface of
ammonoids...... 115
51 Figure B1. Hoploscaphites spedeni macroconch, YPM 23118...... 117 xii
52 Figure B2. Hoploscaphites spedeni macroconch, YPM 23120...... 118
53 Figure B3. Hoploscaphites spedeni macroconch, YPM 23122...... 119
54 Figure B4. Hoploscaphites spedeni macroconch, YPM 23124...... 120
55 Figure B5. Hoploscaphites spedeni macroconch, YPM 23129...... 121
56 Figure B6. Hoploscaphites nebrascensis macroconch, YPM 23144...... 122
57 Figure B7. Hoploscaphites nebrascensis macroconch, YPM 23145...... 123
58 Figure B8. Hoploscaphites nebrascensis macroconch, YPM 23147...... 124
59 Figure B9. Hoploscaphites nebrascensis macroconch, YPM 23151...... 125
60 Figure B10. Hoploscaphites nebrascensis microconch, YPM 23195...... 126
61 Figure B11. Hoploscaphites nebrascensis microconch, YPM 23198...... 127
62 Figure B12. Hoploscaphites spedeni microconch, YPM 23199...... 128
63 Figure B13. Hoploscaphites spedeni microconch, YPM 23200...... 129
64 Figure B14. Hoploscaphites nebrascensis microconch, YPM 23687...... 130
65 Figure B15. Hoploscaphites spedeni microconch, YPM 23694...... 131
66 Figure B16. Hoploscaphites nebrascensis microconch, YPM 23697...... 132
67 Figure B17. Hoploscaphites spedeni microconch, YPM 23704...... 133
68 Figure B18. Hoploscaphites spedeni microconch, YPM 23706...... 134
69 Figure B19. Hoploscaphites spedeni microconch, YPM 23714...... 135
70 Figure B20. Hoploscaphites spedeni microconch, YPM 23719...... 136
71 Figure B21. Hoploscaphites spedeni microconch, YPM 23727...... 137
72 Figure B22. Hoploscaphites spedeni microconch, YPM 23730...... 138
73 Figure B23. Hoploscaphites spedeni microconch, YPM 23732...... 139
74 Figure B24. Hoploscaphites spedeni macroconch, YPM 24817...... 140 xiii
75 Figure B25. Hoploscaphites spedeni macroconch, YPM 24818...... 141
76 Figure B26. Hoploscaphites spedeni macroconch, YPM 27160...... 142
77 Figure B27. Hoploscaphites spedeni macroconch, YPM 27161...... 143
78 Figure B28. Hoploscaphites spedeni macroconch, YPM 27164...... 144
79 Figure B29. Hoploscaphites spedeni macroconch, YPM 44402...... 145 1
CHAPTER I. INTRODUCTION
Ammonoid Intraspecific Variation
Paleontologists recognize numerous examples of interspecific and intraspecific variation in ammonoid taxa, especially in Mesozoic ammonoids. While many of the previous studies have focused on interspecific variation within ammonoid genera or higher taxonomic groups to understand broader evolutionary relationships between species (Landman et al., 1991; Landman
& Geyssant, 1993; Harada & Tanabe, 2005; Gerber et al., 2007), only more recently have studies closely examined the variation observed within a single species. According to Weitschat (2008), intraspecific variability is common among ammonoids, but this character often is missed in evolutionary analyses.
Several examples of well-documented intraspecific variability in Triassic ammonoids are known from a rock unit in Siberia (Dagys & Weitschat, 1993; Dagys et al., 1999; Dagys, 2001;
Weitschat, 2008). These Triassic ammonoids come from a single nodular bed, which contains
600 ammonoids (90% of which belong to one highly variable species, Svalbardiceras spitzbergensis) at various ontogenetic stages (Weitschat, 2008). Researchers describe the specimens discovered at this locality as ranging from involute, smooth, compressed shells to evolute, ribbed, depressed shells (Fig. 1) (Weitschat, 2008).
Other well-documented examples of intraspecific variability are Placenticeras kaffrarium from the Upper Cretaceous of India (Gangopadhyay & Bardhan, 2007), multiple species of
Neogastroplites from the mid-Cretaceous of the Western Interior (Reeside & Cobban, 1960), and
Dactylioceras from the Early Jurassic (Howarth, 1978). In fact, De Baets et al. (2013) provide an extensive list of described Mesozoic ammonoid taxa that exhibit intraspecific variability
(Reeside & Cobban, 1960; Westermann, 1966; Kennedy & Cobban, 1976; Hohenegger & 2
Tatzreiter, 1992; Dagys & Weitschat, 1993; Checa et al., 1996; Dagys et al., 1999; Morard &
Guex 2003; Monnet & Bucher, 2005; Weitschat, 2008; Monnet et al., 2010). De Baets et al.
(2013) show that some Paleozoic ammonoids (i.e., “Anetoceras” and “Erbenoceras”) from
Morocco also exhibit very large intraspecific variation. These examples express a large spectrum of variability in size, shape and ornamentation throughout the ontogeny of each taxon. If each morphotype of a species was found separately, some paleontologists may classify these specimens as different, closely related species or even genera; however, the morphotypes of each example are often spatially and temporally observed together, which suggests that they belong to the same taxon.
Oversplitting of taxa is a serious problem in paleontology since it can cause evolutionary trees, which are necessary for studying evolutionary relationships and the evolution of new character states, to become unresolved. Due to constraint in some characters (such as septal suture form (Yacobucci & Manship, 2011)) and plasticity of other characters (such as shell shape and ornament), cladistic approaches of ammonoids used by biologists and paleontologists to define the evolutionary relationships often are inconclusive (Yacobucci, 2012). However, by understanding the spread of variation of relatively plastic characters as opposed to constrained characters, researchers can better define and utilize characters and character states they apply in cladistic analyses to develop plausible evolutionary trees. In addition, this variability due to developmental plasticity of certain characters may fuel the high speciation rates that are commonly observed in ammonoids (Yacobucci, 1999).
Recognizing the possibility of drastically variable morphological extremes within a species, many evolutionary biologists and paleontologists seek to explore the mechanisms causing such variability within a taxon. Some researchers conclude that heterochrony may play 3 an important role in creating new species or variants within a species (Landman & Geyssant,
1993; Yacobucci, 1999; Harada & Tanabe, 2005; Gangopadhyay & Bardhan, 2007). Gould
(1977) defined heterochrony as changes in the timing and rate of developmental traits during ontogeny. In other words, as an organism grows to maturity, the timing and duration of expressed characters differs from that expressed by ancestral taxa. These changes in development cause paedomorphosis or peramorphosis to occur in descendants (Fig. 2) (Gould,
1977). In paedomorphy, the adult form of a derived taxon expresses the juvenile characters of its ancestor by means of progenesis or neoteny; conversely, a peramorphic taxon expresses ancestral adult characters in the juvenile form of a descendant through acceleration or hypermorphosis
(Gould, 1977; Alberch et al., 1979). Landman and Geyssant (1993) analyzed several published examples of heterochronous ammonoids and determined that paedomorphosis was the most prevalent type of heterochrony. By altering the timing and duration of specific morphological features through developmental genetic plasticity within a population, a broad spectrum of variability may arise from which new species originate (Yacobucci, 1999). Heterochrony via developmental genetic plasticity may even explain polymorphism, or discrete morphotypes, observed in many ammonoid taxa displaying intraspecific variability (Gangopadhyay &
Bardhan, 2007).
Originally noted by English geologist S.S. Buckman in 1887, ornamentation and shell shape covary in ammonoid taxa (Westermann, 1966). More specifically, ammonoids with compressed, involute forms are lightly ornamented while ammonoids with depressed, evolute forms are heavily ornamented; these observations define the First Buckman Law of Covariance
(Fig. 1) (Westermann, 1966). Many ammonitologists hypothesize that this covariation exists due to genetic controls, constructional constraints or environmental factors influencing different 4 aspects of the shell (Guex et al., 2003; Yacobucci, 2004; Hammer & Bucher, 2005). If covariation is controlled genetically, then heterochrony may only be acting on certain characters, such as ornamentation, with shell shape and size consequentially changing in response due to genetic linkages or constructional constraints. While Buckman's first law seems to prevail within a species, some studies demonstrate that Buckman's first law does not always apply across ammonoid species (Yacobucci, 2004; Hammer & Bucher, 2005). In order to fully conceptualize how intraspecific variability occurs within a species, one must be able to first quantify the degree of variability for each trait before any evolutionary relationships (intra- or interspecific) can be analyzed.
Quantifying Intraspecific Variability
While many previous studies have acknowledged the mechanisms behind intraspecific variation, fewer have attempted to quantify variability of specific morphological characters observed within an ammonoid species. Many ammonoid researchers have utilized morphometric techniques to obtain Raup's (1966, 1967) parameters to describe variations in shape and size of coiling shells (Fig. 3). Traditionally, the four measurements: shell diameter, coiling radius, whorl height and umbilicus diameter (Raup, 1966, 1967; Landman, 1987; Landman & Waage, 1993;
Gerber et al., 2007), are measured to calculate Raup’s parameters. With these parameters, researchers can describe shell shape and size with respect to the various stages of ontogenetic development by constructing ontogenetic trajectories (Yacobucci, 2004; Gerber et al., 2007; De
Baets et al., 2012; Korn, 2012) that compare the possible forms through development. While this method does allow for quantitative analysis of disparity within a species or higher taxonomic group, it fails to factor in ornamentation and its covariation with shape within a species. Also, size and shape alone are unreliable for determining the exact stage of development since 5 heterochrony affects the size and shape expressed at maturation (Landman & Geyssant, 1993).
Although covariation does exist between shell shape and ornamentation, appearance or disappearance of specific ornaments provide a better proxy for developmental stage since these changes coincide with other developmental features (such as closely spaced septa at the obtainment of maturity) (Davis et al., 1996). Additionally, several studies have utilized mathematical models to reconstruct ammonoid shells to quantify shell dimensions and features
(Hariri & Bachnou, 2004; Hammer & Bucher, 2006; Ubukata et al., 2008). Again, some of these studies lack an evaluation of how ornamentation correlates with shape and size of ammonoid shells. In order to truly measure variability within an ammonoid species, ornamentation must be considered and quantified.
Most studies use qualitative descriptions to explain ornamental variability through ontogenetic development and shape (Landman, 1987; Landman & Waage, 1993; Harada &
Tanabe, 2005; Gangopadhyay & Bardhan, 2007; Weitschat, 2008; Nishimura et al., 2010). For example, a paleontologist describes what type of ornamentation (ribs, tubercles, bullae, etc.) is present at any given stage in ontogeny. While qualitative methods can be good descriptors for the variation observed in a given taxonomic group, these descriptions allow for biased results as many descriptions can be subjective (Bocxlaer & Schultheiss, 2010). Researchers need to utilize new methods to describe ornamentation so that variability throughout ontogeny can be closely examined.
More recently, studies have applied geometric morphometrics using landmarks and geographic information systems (GIS) to quantify morphological variability in a number of mollusk species (Stone, 1998; Bocxlaer & Schultheiss, 2010; Bookstein & Ward, 2013).
However, landmark techniques require that specific, ideally homologous, points can be located 6 on every specimen. In the case of ammonoids, ornamentation can be so variable within a species that no landmarks can be determined that encompass the entire shell form (shape, size and ornament). Although some paleontologists have been successful in using landmarks to describe the shape of ammonoid ribs and overall shape of the shell (Neige, 1999), these studies fail to quantify features such as duration of particular ornaments at ontogenetic stages, variations in shape due to strong ornamentation, and the clustering of analogous tubercles throughout ontogeny. Some morphometric researchers have explored the utility of semi-landmarks and closed, outline curves to describe shell morphology (Bocxlaer & Schultheiss, 2010; Bookstein &
Ward, 2013). Semi-landmarks are defined as evenly-spaced, though arbitrary, points along homologous, open or closed curves (Bocxlaer & Schultheiss, 2010). These studies focus mainly on shell shape and shell coiling without addressing ornamentation's effect on these morphological characters. The quantity and placement of ornamentation in ammonoids varies to such a degree that landmark techniques alone are impractical to use to quantify variability.
On the other hand, GIS presents a new methodology that paleontologists can exploit for quantifying morphological variability when homologous landmarks cannot be determined.
Previously, evolutionary biologists and paleontologists have analyzed mammal tooth crowns using GIS techniques (Fig. 4); essentially, researchers use GIS to analyze teeth as topographic landscapes (Ungar & Williamson, 2000; Evans et al., 2007; Eronen et al., 2009; Wilson et al.,
2012). Invertebrate paleontologists have used GIS-based methodologies to quantify morphology in various echinoderm groups (Zachos, 2012; Sheffield et al., 2012). Manship (2004), Waggoner and Manship (2004) and Yacobucci and Manship (2011) have utilitized GIS to analyze spatial constraints and asymmetry of ammonoid sutures. By using GIS to analyze shell morphology as a landscape, researchers can directly compare morphological variability with respect to shape and 7 ornamentation within a species (Knauss, 2012). Thus far, only Knauss (2012) has used GIS to quantify intraspecific variability in regards to shape and ornamentation of the shell through ontogeny; researchers should explore the utility of this technique to quantify ornament and shape variability in ammonoids.
Heteromorphic Ammonoids and Scaphites
Throughout the Mesozoic Era, some ammonoid taxa evolved shells that deviate from the typical planispiral form. Some taxa have a planispiral juvenile phragmocone followed by a straight adult body chamber; other taxa have loosely coiled shell morphologies throughout the entire ontogeny of their shells. One taxon, Nipponites, demonstrates one of the most extreme cases, with its shell loosely coiled in a series of U-bends that form a tangled shell design
(Clarkson, 1986). These heteromorphic ammonoids are first present during the Late Triassic and are observed through the Jurassic; however, extensive development of heteromorphic ammonoids did not occur until after the Jurassic, in the Early Cretaceous (Clarkson, 1986).
Early paleontologists hypothesized that these heteromorphic forms were the result of degeneration of taxa, since the major episodes of heteromorphy in the Triassic and Cretaceous occurred just before intervals of large extinction events (Clarkson, 1986). This idea blossomed during the 1920s and 1930s as paleontologists viewed these heteromorphs as evolving overspecialized forms to adapt to new environmental conditions before their ultimate extinction
(Clarkson, 1986). While this idea persisted for decades, Wiedmann (1969) demonstrates that heteromorphy only evolved in specific taxa and not across higher taxonomic groups at any one time. Additional evidence of uncoiled ancestors from the Early Cretaceous giving rise to normally coiled or near-coiled descendants also supports Wiedmann's (1969) hypothesis that heteromorphy does not represent overspecialization before extinction (Clarkson, 1986). In 8 retrospect, this observation further demonstrates developmental plasticity in ammonoids in at least higher taxonomic classifications. These findings caused paleontologists to reassess heteromorphy in ammonoids since no direct link between heteromorphic form and extinction exists.
Within Ancyloceratina, the superfamily Scaphitaceae contains many of the heteromorphic ammonoids from the Late Cretaceous. Researchers can easily recognize scaphites in the field and in the laboratory due to their planispiral phragmocone and their uncoiled, shaft- shaped body chamber that terminates with a hook projected back towards the phragmocone (Fig.
5 ) (Landman et al., 2012). This dramatic shift in the shell growth program suggests that maturity occurs at the start of the formation of a shaft body chamber (Clarkson, 1986; Landman et al.,
2012). Internally, closely approximated septa indicated the termination of growth and the obtainment of full maturity (Bucher et al., 1996; Landman et al., 2010; Landman et al., 2012).
Davis et al. (1996) describe members of Scaphitaceae as having not only sexual dimorphs called macroconchs and microconchs (females and males, respectively), but also as having broad variability in size, shape and ornamentation within each dimorph (Fig. 6). Between macroconchs and microconchs of scaphites, the juveniles exhibit very similar features until adulthood (Davis et al., 1996). In general, fully adult scaphites exhibit a bimodal distribution in regards to size due to sexual dimorphism occurring at maturity (Landman, 1987; Landman & Waage, 1993; Davis et al., 1996, Landman et al., 2010). If developmental plasticity (i.e., changing the timing of expression of developmental regulatory genes) is the mechanism by which heteromorphic ammonoids, specifically scaphites, change their shell morphology at maturation (Landman et al.,
2012), then this same plasticity in development could lead to variability in other shell features at 9 maturity such as ornamentation and shape (Makowski, 1962; Calloman, 1963; Westermanm,
1964; Delanoy et al., 1995; Davis et al., 1996; Parent, 1997; Kakabadze, 2004).
Hoploscaphites (Jeletzkytes) spedeni
Numerous scaphitic ammonoids, belonging to the genus Hoploscaphites (Fig. 7), are known from the Late Cretaceous of the Western Interior. A highly variable scaphitic ammonoid,
Hoploscaphites (Jeletzkytes) spedeni was first described by Landman and Waage (1993) and later reassigned by Landman et al. (2010) to its current genus. Like many other scaphites, this heteromorph is variable in shell size (Fig. 6), shape and ornamentation. Diagnostic characteristics of this taxon include:
“Macroconch shape variable, dominantly stout with quadrate to slightly depressed
aperture[,] body chamber with strong, locally clavate ventrolateral tubercles,
usually with bullae below umbilical shoulder[,] scattered flank tubercles/bullae on
hook [and] phragmocone with up to four rows of tubercles or none. Last whorl of
microconch more loosely coiled, commonly with mid flank row of tubercles on
body chamber; phragmocone variously tuberculate” (Landman & Waage, 1993).
Landman and Waage (1993) demonstrate the variability of this taxon through numerous images
(Fig. 8), qualitative descriptions and size and shape morphometrics. They recognized three major forms, which they dubbed typical, rotund and compressed. Seemingly demonstrating Buckman’s
Law of Covariation (Westermann, 1966), the rotund forms tend to be heavily ornate while the compressed forms have sparse ornamentation. The typical form, the most common morph of H. spedeni, has features that grade between the rotund and compressed forms.
Hoploscaphites (Jeletzkytes) nebrascensis
Significantly less variable than Hoploscaphites (Jeletzkytes) spedeni, Hoploscaphites 10
(Jeletzkytes) nebrascensis (Landman & Waage, 1993; Landman et al., 2010) is hypothesized to be the descendant, sister taxon to H. spedeni (Landman & Waage, 1993; Landman et al., 2013).
Diagnostic characteristics of this taxon include:
“Macroconch compressed, last whorl markedly higher than wide, phragmocone
conspicuously tuberculate with strong ventrolateral tubercles and five to seven
rows of weaker flank tubercles, which commonly extend onto part of the body
chamber. Microconch variably tuberculate, commonly with double row of
umbilical tubercles/bullae” (Landman & Waage, 1993).
In a description of scaphitic ammonoid specimens recovered from the Late Cretaceous of the Western Interior, Landman and Waage (1993) state that most H. nebrascensis ammonoids have relatively constrained characters in shell shape, size and ornamentation
(Fig. 9). In fact, the shape of the macroconchs remains consistent among specimens regardless of variations in shell size, and tubercles are found on almost all ribs. Most notably, a researcher can distinguish H. nebrascensis from its hypothesized sister taxon,
H. spedeni, by the extensive tuberculation of the shell through the full ontogeny (Fig. 10).
However, recent discoveries of an intermediate specimen assigned to H. spedeni (Fig. 11)
(Landman et al., 2013) shows that a qualitative assessment of ornamentation (i.e., tubercles) may be ineffective in species identification due to the possible overlap in the variability spectra between H. spedeni and H. nebrascensis.
Biostratigraphy
Stratigraphically, H. spedeni occurs in the topmost beds of the Elk Butte Member of the
Pierre Shale and the overlying Trail City Member of the lower Fox Hills Formation (Fig. 12). H. nebrascensis first appears at a transitional concretion layer between the Trail City Member and 11 the Timber Lake Member before diminishing at the Iron Lightning Member of the upper Fox
Hills Formation (Fig. 12). The type section of the Fox Hills Formation, which represents the
Upper Maastrichtian of the Late Cretaceous, is located in the Black Hills of South Dakota, USA
(Waage, 1964, 1968; Speden, 1970). Biostratigraphers have traditionally used heteromorphic ammonoids, such as baculites and scaphites, for biostratigraphic zoning of these Late Cretaceous formations within the Western Interior (most recently, Landman et al., 2010 and Landman et al.,
2013). H. spedeni is one of the few scaphite ammonoids restricted solely to the Hoploscaphites nicolletii Range Zone (Waage, 1964, 1968; Speden, 1970; Landman & Waage, 1993), which spans the boundary between the older Pierre Shale and younger Fox Hills Formations (Waage,
1964, 1968; Speden, 1970). As such, H. spedeni is an important index fossil for recognizing this range zone. The distribution of the different morphs of this taxon among the assemblage zones comprising the H. nicolletii Range Zone are described by Landman and Waage (1993):
“...In both LNAZ and UNAZ the distribution is similar with the typical form
being most numerous, followed by the compressed form, and then the relatively
rare rotund form. In LGAZ, the rotund form dominates and in POAZ, the rotund
form is second to the typical form” (see Figure 12 for acronyms of assemblage
zones).
The H. nicolletii Range Zone terminates at the “transition concretions” where faunal elements from the H. nicolletii Range Zone and faunal elements from the overlying range zone briefly co- occur. The first appearance of H. nebrascensis is within the “transition concretions,” and it marks the beginning of the Hoploscaphites nebrascensis Range Zone (Fig. 12) (formerly, Jeletzkytes nebrascensis Range Zone) (Waage, 1964, 1968; Speden, 1970; Landman & Waage, 1993;
Landman et al., 2010; Landman et al., 2013). Many “rotund” H. spedeni specimens bear close 12 resemblance to H. nebrascensis or represent transitional forms (Fig. 11); these specimens are especially common near the youngest rocks of the H. nicolletii Range Zone before the first occurrence of H. nebrascensis (Landman & Waage, 1993; Landman et al., 2013).
While numerous morphometric comparisons have been made with regard to size, shape and whorl number to measure variability in scaphitic taxa throughout ontogeny (Landman &
Waage, 1993; Landman et al., 2008; Landman et al., 2010), no researcher has quantified the variability of ornamentation with respect to shell shape through ontogeny in H. spedeni and its stratigraphically younger sister taxon, H. nebrascensis. With several examples of transitional forms between these two closely related species, a new methodology, such as GIS-based analyses, to quantify the spectrum of variability in morphological characters (i.e., shell shape and ornamentation) is necessary for species identification and evolutionary analysis.
Research Objectives
The two, closely related, heteromorphic ammonoids H. spedeni and H. nebrascensis seemingly express two different, but potentially overlapping, ranges of morphological variability in shell shape and ornamentation between juvenile and adult forms and between the sexual dimorphs. Unlike the extreme variability observed in both dimorphs of H. spedeni, H. nebrascensis macroconchs appear to express a much more narrow degree of variability in conch morphology. Microconchs of H. nebrascensis, on the other hand, seem to exhibit a fairly large morphological range much like the microconchs of their ancestor, H. spedeni.
My first objective is to utilize GIS-based spatial analyses to quantify ammonoid characters. Geospatial analyses are well developed and have been used in many interdisciplinary fields, so this methodology may present a way in which paleontologists can quantify complex morphologies. My second objective is to determine the target of evolutionary change between 13 the juvenile and adult forms of H. spedeni and its sister taxon, H. nebrascensis, utilizing the results obtained from the GIS-based analyses. I assume that the aspect of shell form showing the most variability will provide the most “raw material” for natural selection. I therefore hypothesize that the target for evolutionary change most likely resides in the macroconch, specifically during the development of the inflated adult body chamber, which forms during the morphogenetic countdown to full sexual maturity and seems to exhibit higher variability in the arrangement of ornamentation.
In order to achieve these objectives, I first applied a 3D GIS-based methodology to the flank surface of each ammonoid to globally quantify and visualize morphological variability, focusing on the surface roughness or “topography” of the shell. Second, I applied a 2D GIS- based methodology to quantify more specifically the distributions of ornament features over the full ammonoid shell as well as aspects of shell shape. Third, I synthesized all these data and compared the juvenile phragmocone to the adult body chamber between the macroconchs and microconchs of both Hoploscaphites species to determine where the target for evolutionary change most likely resides, as evidenced by the largest spread of morphological variation. Last, I evaluated the GIS-based technique for its usefulness in quantifying ornament and shell shape variability. 14
CHAPTER II. METHODOLOGY
Acquisition of Digital Models
At the Yale Peabody Museum of Natural History (YPM), I observed 55 macroconchs and
43 microconchs of H. spedeni and five macroconchs and five microconchs of H. nebrascensis. In order to quantify variability using GIS techniques, I created digital models of each specimen.
Traditionally, researchers use 3D laser scanners to acquire 3D models of specimens; however, the cost of the hardware and software for such technology was beyond the bounds of my budget, and
YPM did not have a 3D laser scanner available at the time I conducted the analysis. Instead of obtaining digital models by laser scanning specimens, I used photographs, photogrammetric software and 3D rendering and editing software (Falkingham, 2012) to reconstruct each fossil specimen in a 3D format. Falkingham (2012) has shown that this methodology for acquiring 3D models of fossils creates a higher resolution model than if one used a 0.3 mm laser scan to generate 3D reconstructions. For each specimen observed, I captured 140 to 200 photographs, depending on its size, of the best preserved lateral surface of the shell. To ensure that full coverage of the lateral surface of the shell was properly photographed, I placed each shell on a white, bean-bag pillow contained within a nine inch, round cake pan with every 10° marked off along the cake pan’s edge ranging from 0° to 350° (Fig. 13). The bean-bag pillow allowed for each fossil specimen to be positioned so that the dorso-ventral plane was perfectly level, which is convenient for GIS analyses; I used a bubble level to verify the leveling around the shell. I captured three to four photographs varying in angle between parallel with the dorso-ventral plane and perpendicular to the dorso-ventral plane at every 10° mark. Three to four photographs captured at every 10° around the specimen is necessary because three key points must be recognized in each photograph for proper reconstruction (Falkingham, 2012); however, this 15 method does not require that photographs be taken at specified angles from the horizontal plane.
Once these photographs were captured, I took additional random photographs around the specimen at varying angles to ensure photographs overlapped in their coverage of the ammonoids. I used a Nikon 16.0 MPX camera and two 32GB SD cards to take photographs of the specimen at 6.0 MPX; a low pixel resolution allowed for more storage capacity without sacrificing image quality at the small scale of the camera’s view. In addition to photographing the specimens, I measured standard ammonoid shell parameters (Fig. 14) using digital calipers; these measurements allowed for verification of scaling methods within a GIS environment (see below).
Two free photogrammetric software programs (VisualSFM and 123D Catch) were tested for 3D reconstruction of ammonoid models; however, 123D Catch (Autodesk, Inc., 2013) was more efficient at creating accurate models than VisualSFM (Wu, 2011) and was ultimately used for analysis. First, following similar methods to Falkingham (2012), I imported the photographs into VisualSFM, free, open-source photogrammetric software developed by Wu (2011), to produce a sparse-point cloud, that is, XYZ scaleless coordinates of the exterior morphology of an object. Within the VisualSFM software, I executed Furukawa and Ponce’s (2010) Clustering
Views for Multi-View Stereo (CMVS) and Patch-based Multi-View Stereo (PMVS v2) software applications to generate a dense-point cloud, a high resolution XYZ coordinate model, in a file format titled .PLY that laser scanning software can commonly export. Unfortunately, this program is incredibly time-consuming as it operates on the user’s central processing unit (CPU) unless the user has a specific graphics card that allows the software to operate on the graphics processing unit (GPU). Additionally, this program is not capable of producing a mesh surface over the dense-point cloud to generate a continuous model. In order to do so, I would need to 16 import each .PLY file into MATLAB (The MathWorks, Inc., 2007a) and apply the “Griddata” function. The dense-point clouds produced by VisualSFM also tend to have excess points representing background scatter directly on top of the ammonoids’ points, making editing out extraneous points tedious and potentially error-inducing.
Because of these complications, instead of using the VisualSFM photogrammetric software, I opted to use 123D Catch. This program runs in much the same way as VisualSFM except instead of using the user’s CPU or GPU, the dense-point cloud, mesh surface, and camera angles are all constructed on the Autodesk, Inc. server. As such, the program is significantly more time-efficient, automatically produces a high-resolution mesh over the dense-point cloud, and allows for files to be saved in Computer Aided Design (CAD) file formats, which are readable in a GIS environment. Additionally, post 3D reconstruction editing is minimal since this particular program included fewer extraneous points on top of the actual ammonoid, reducing the possibility of erroneous deletion of true points. However, background scatter is still present around the ammonoid that must be deleted, as is true of the models produced by VisualSFM or any other photogrammetric software (Fig. 15).
To delete extra background points, I exported both the mesh and dense-point cloud of the
3D models produced by 123D Catch as .OBJ files (see below for further explanation) and then imported them into MeshLab (Cignoni et al., 2011), an open source program for processing and editing unstructured 3D triangular meshes. Once imported to MeshLab, I removed the mesh surface over the 3D dense-point cloud and removed extraneous background points that were positioned around the ammonoid model; the reason for removing each dense-point cloud’s mesh surface is discussed below. Additionally, I digitally cut each ammonoid along the dorso-ventral plane so that only the lateral surface facing up was assessed, as opposed to the keel and sections 17 of the shell in direct contact with the bean-bag pillow (Fig. 16). I visually approximated the position of the dorso-ventral plane along the ammonoid’s keel; because I verified the orientation of the dorso-ventral plane of each ammonoid with a bubble level when photographing them, the error in the visual approximation of the dorso-ventral plane is unlikely to be a major factor. After editing the ammonoid 3D models in MeshLab, I extracted the vertices of the .OBJ files (the dense-point cloud having between 100,000 to 200,000 points) and saved them separately as a text file to be uploaded directly into GIS software. For unknown reasons, MeshLab occasionally inverted ammonoid models. In other words, some ammonoid models were negative relief instead of positive relief. To revert these few, erroneous models into properly oriented models, the vertices of the .OBJ files were multiplied by -1; while this conversion caused the models to no longer be impressions, it unfortunately caused the ammonoid to be mirror-imaged (i.e., a left lateral surface ammonoid appeared to be a right lateral surface) (see Appendix B). Mirror- imaging of these few ammonoid models had no effect on the results obtained from the GIS analyses (see below).
Digital Models Imported into a GIS Environment
As stated above, 123D Catch is capable of generating models in a CAD file format (i.e.,
.DWG file), readable by many GIS software programs. Unfortunately, I discovered that the GIS software package used for analyses, ArcGIS 10 (ESRI, Inc., 2011), is incapable of reading the old version of the .DWG CAD files generated by 123D Catch. Because of the incompatibility of
ArcGIS 10 and 123D Catch’s .DWG files, I exported each ammonoid model as an .OBJ file in the previous steps so that the dense-point cloud vertices could be uploaded directly into ArcGIS
10, much like a set of GPS coordinates.
To reconstruct the ammonoid surface within ArcGIS 10, I created a triangulated irregular 18 network (TIN), defined by Price (2012) as “a data model for storing surfaces as triangular facets with varying orientations,” from the imported XYZ vertices of the dense-point cloud. Although the TIN might not be exactly the same as the original mesh generated by 123D Catch, the lateral surface of the ammonoid was reconstructed at such a high resolution, due to the large number of vertices making up the point cloud, that the TIN surface is equally visually accurate (Fig. 17). I left the scale of each ammonoid TIN model at its default setting since the GIS analyses of 3D models I utilized are scaleless.
3D GIS Methodology
Orientation Patch Count
Various researchers who utilize GIS to quantify morphology have used surface aspect, or slope direction, on vertebrate teeth (Ungar & Williamson, 2000; Evans et al., 2007; Wilson et al.,
2012; Salazar-Ciudad & Marin-Riera, 2013). Specifically, these researchers used the number of discrete patches of different aspects on the surface of the tooth as an index for dental complexity.
Evans et al. (2007) and Wilson et al. (2012) call this metric “orientation patch count (OPC),” and currently a database is growing that lists the OPC of various mammalian teeth. If the tooth surface is complex (i.e., having numerous ridges or cusps), then the OPC will be higher; likewise, if the tooth surface is fairly simple (i.e., being only a single ridge), then the OPC will be lower.
I applied this OPC index to my ammonoid TIN surfaces within ArcGIS 10 (Fig. 18). H. spedeni and H. nebrascensis both have ribs and tubercles. Ribs running along the flank of the shell are equivalent to an elongate ridge, having two main slope aspects (e.g., North and South).
On the other hand, tubercles are circular to ellipsoid in shape, having potentially more than two main slope aspects (e.g., North, South, East and West). By extracting a polygon feature class 19 containing all the discrete aspect patches (North, Northeast, East, Southeast, South, Southwest,
West, and Northwest), I was able to count the total number of polygons of the lateral shell surface. I compared these specimens between both dimorphs, both lefts and rights, and both phragmocones and body chambers. Due to the manner in which ArcGIS 10 reads class break tables assigned to TINs for reclassification of feature classes, North aspect is divided into two feature class polygons: 337.5° to 359.9° and 0° to 22.5° (the default classification). Two adjacent polygons both belonging to North were not merged together if one polygon was west of 0° and the other east of 0°. Additionally, the TIN surface around the margins of the ammonoid models is erroneously reconstructed, either from errors in the original 3D dense-point cloud made in 123D
Catch or from errors in the generation of the TIN surface in ArcGIS 10. These errors resulted in small numerous polygons along the shells’ edges, many of which have seemingly been assigned a random aspect. These erroneous polygons around the perimeter of each shell coupled with the two North aspect feature classes caused inflation of the OPC. Despite this inflation, all specimens were subject to the same type of conditions or errors, eliminating any bias that might be present in the data.
Surface-to-Planimetric Area Ratio
In addition to OPC as a measure of surface complexity, some vertebrate paleontologists and biologists use a relief index derived from the ratio of the 3D surface area to 2D planimetric area of fossil tooth crowns (Ungar & Williamson, 2000). Rashid (2010) demonstrated that the 3D to 2D area ratio of a topographic surface is a good index for topographic roughness and is independent of scale.
In order to calculate the surface-to-planimetric area ratio, I converted the TIN of each ammonoid to a digital elevation model (DEM) raster, defined by Price (2012) as “a data set 20 composed of an array of numeric values (e.g., elevation data), each of which represents a condition in a square element of ground” within ArcGIS 10. The cells of the raster remained at the default cell size since reducing the cell size had little effect on the resolution of the lateral surface of the ammonoid DEM. After creation of the DEM raster, I used the 3D Analyst Tools within ArcGIS 10 to create a percent slope raster, where the slope of each cell was calculated and displayed. Using the Raster Calculator tool within the Spatial Analyst Tools, I applied the following calculation provided by Berry (2007):
3D to 2D Area Index = 1/(COS(ATAN(Percent Slope/100)))
Because this calculation derives an index independent of size and shape much like the OPC index, I was able to compare left lateral shells to right lateral shells, macroconchs to microconchs, and phragmocones to body chambers of both H. spedeni and H. nebrascensis (Fig.
19).
Generating Coordinate Systems for 2D Methodology
I used ammonoids exhibiting the left lateral surface for additional two dimensional analyses with ArcGIS 10 that require proper scaling of feature classes. For each left ammonoid, I selected one photograph exhibiting the perpendicular view of the lateral surface. Parallax is unlikely to be a huge error-inducing factor due to the centering of the ammonoid within the photograph’s view.
After importing each .JPEG image as a layer into ArcGIS 10, I created a properly scaled coordinate system for each photograph. After scaling each photograph utilizing the number of pixels per centimeter with respect to the scale bar within the image (Fig. 13), the coordinates of each corner of the photograph were entered as ground control points (Price, 2012) within the
Georeferencing Tool menu in ArcGIS 10. After the rectification of each coordinate system, I 21 verified that scaled measurements of shell dimensions were consistent with real-life measurements collected for each ammonoid.
2D GIS Methodology
Once properly scaled within the GIS environment, I used these images to digitize by hand ammonoid morphological features as feature class vectors (i.e., points, polylines and polygons) in shape files (Fig. 20). Points represented tubercle centers for flank tubercles and tubercle base centers for ventrolateral tubercles. Polylines represented shell ribs, including primary ribs and secondary ribs, which branch off the primaries. Polygons capture each shell’s lateral profile (i.e., shell shape), which was further separated into phragmocone and body chamber polygons. I digitized additional feature classes (Fig. 21) consisting of spiral polyline transects positioned at approximately 25%, 50% and 75% of whorl height for rib analyses. I also created 1 mm and 2 mm buffers around points using the Buffer Tool in ArcGIS 10 for tubercle analyses. The size of these buffers was based on the reported average rib density for Hoploscaphites of five to 10 ribs per centimeter (Landman & Waage, 1993; see below).
Rib and Tubercle Spacing
I conducted an Average Nearest Neighbor analysis in ArcGIS 10 to determine the pattern of rib and tubercle spacing (Mitchell, 2005). This type of spatial analysis calculates a nearest neighbor index (NNI) based on the mean distance of each feature to its closest neighboring feature. Average Nearest Neighbor analysis uses the centroids of the polylines or the locations of points to calculate Euclidean distances. Here, the technique was used to derive the average rib spacing approximately halfway up the shell’s flank and the average distance between neighboring tubercles. NNI measures how similar the observed mean distance is to the expected mean distance in the form of a ratio. The null hypothesis states that these features are randomly 22 distributed with the distances of the nearest neighbors having a normal distribution. Features are considered clustered, or in discrete groups, when the NNI is less than 1; features are considered dispersed, or evenly spaced, when the NNI is greater than 1 (Fig. 22). The significance of the results of this spatial analysis can be tested by calculating a Z-score. Unfortunately Average
Nearest Neighbor analysis has been shown to be potentially erroneous depending on how close feature classes are to the margins of the mapping area (Mitchell, 2005). While the possibility of error in this analysis is present, especially for tubercle points given their closeness to the outer margins of the mapping area, Average Nearest Neighbor analysis still has the potential to provide information regarding ornamentation characters on ammonoid shells.
Rib Density
I used the Spatial Join Tool within ArcGIS 10 to generate several indices related to rib density. I calculated rib density by intersecting the rib polyline layers with the spiral transect layers to count the number of ribs intersecting a spiral transect, which was then divided by the length of that specific spiral transect to generate a rib density index. I repeated this process to generate indices for rib density at 25%, 50% and 75% of the shell’s flank height (Fig. 23). While this method provides an average rib density at these three subdivisions over the full ontogeny of the shell, it does not provide information regarding specifically where in ontogenetic development ribs are most dense. To more closely examine rib density, I generated rib density maps from the rib polyline vectors via the Line Density tool in ArcGIS 10 (Fig. 23). These maps allow for comparisons across sexual dimorphs and between phragmocones and body chambers.
Tubercle-Rib Spatial Correlation
I used the Spatial Join Tool in ArcGIS 10 again to generate indices of tubercle-rib spatial correlations. After intersecting the 1 mm and 2 mm buffers and the rib polylines, I calculated the 23 percentage of tubercle buffers with intersecting rib polylines. According to Landman and Waage
(1993), Hoploscaphites specimens have on average five to 10 ribs per centimeter on the phragmocone and seven to 14 ribs per centimeter on the body chamber. As such, I decided that a
1 mm and 2 mm distance around each tubercle point was appropriate to determine if tubercles and ribs are spatially correlated (Fig. 24). However, I do note that ribs and tubercles obviously have variable widths, and polylines and points do not have measurable widths, suggesting that vector data may be incapable of fully representing the true morphology of ornamentation.
Although this represents a simplification of rib and tubercle form, these buffers are appropriate for tubercle-rib correlation given the small size of most of the tubercles (less than 0.5 cm) and the slenderness of the ribs (approximately 0.1 cm) in these taxa.
Shell Coiling
To assess shell shape variability, specifically shell coiling (i.e., how tightly or loosely coiled the body chamber is around the phragmocone), I extracted the planimetric areas of the body chamber and of the entire shell (combined exposed phragmocone and body chamber) polygons with the Utility Toolbox in ArcGIS 10. A ratio of these two areas provides an index for shell coiling (Fig. 25). A large value (i.e., greater than 0.5) indicates that the total planimetric area of the whole shell consists of mostly body chamber area with the body chamber presumably coiled tightly around the phragmocone. On the other hand, a value near 0.5 indicates that the body chamber area constitutes approximately half of the whole shell area, suggesting that much of the phragmocone is exposed and that the body chamber is loosely coiled around the phragmocone. While requiring a scale for the actual calculation of area, the ratios themselves are scaleless and allow for specimens of H. spedeni and H. nebrascensis to be compared directly between macroconch and microconch. 24
Body Chamber Shape Distribution
In both H. spedeni and H. nebrascensis, many changes in morphology occur throughout the development of the adult body chamber, which deviates from the planispiral form observed in the juvenile phragmocone. To understand developmental plasticity in the formation of the body chamber, I defined two sets of landmarks on the body chamber on which to overlay polygons to see morphological differences: the ventral and dorsal end points of the ultimate septum and the dorsal end point of the ultimate septum paired with the dorsal lip of the aperture.
These landmarks were easily found on each specimen, with the former representing the onset of sexual maturity during development and the latter representing the start and end of the body chamber. Utilizing these landmarks allowed for specimens to be scaled up or down to these two coordinates for comparisons without true size influencing the results. Using the Merge Tool within ArcGIS 10, I overlaid and merged specimens according to species and sexual dimorph, oriented and re-scaled to these defined landmark pairs. This method allowed for a composite polygon to be created that defined the possible morphospace that any specimen’s body chamber could occupy (Manship, 2004; Yacobucci & Manship, 2011) (Fig. 26). To further assess the body chamber shape distribution of both H. spedeni and H. nebrascensis, I developed an index, the ratio of the area of each body chamber to its respective composite polygon morphospace area, so that specimens could be compared.
Statistical Analyses
I compared data and calculated statistics using PAleontological STatistics (PAST)
(Hammer & Harper, 2013). I tested all data for normality using a Shapiro-Wilk test at a 95% confidence limit for statistical significance. I compared data that were normally distributed using 25
Student’s T and F tests. For data that were non-normally distributed, I made comparisons using non-parametric analyses, such as the Mann-Whitney U test, which do not assume normality
(Hammer & Harper, 2006). The applied statistical analyses had a 95% confidence limit for statistical significance in comparisons.
Sample Size for Statistical Analyses
Due to poor preservation or inaccuracy in 3D reconstruction of H. spedeni, I utilized only the best preserved specimens for 3D analyses. Because of the limited sample size of H. nebrascensis, I used all specimens belonging to this species in the 3D analyses. Forty-nine specimens of Hoploscaphites were used: 19 H. spedeni macroconchs, 20 H. spedeni microconchs, five H. nebrascensis macroconchs and five H. nebrascensis microconchs.
Of the specimens chosen for 2D spatial analyses, I found the rib polylines of H. spedeni specimens to be easier to digitize due to the relatively few tubercles of this species. H. nebrascensis often had significantly more tubercles throughout its ontogeny, which causes the exact location of ribs to be obscured. As such, I excluded all H. nebrascensis from analyses that measured aspects of ribs. However, I retained these specimens in analyses of tubercles. Nineteen
H. spedeni (nine macroconchs and 10 microconchs) were used in rib analyses, and 22 H. spedeni
(11 macroconchs and 11 microconchs) and eight H. nebrascensis (four macroconchs and four microconchs) were used in tubercle analyses. All specimens used in 2D analyses exhibited the left lateral shell surface for consistency in digitizing features. 26
CHAPTER III: RESULTS
3D GIS Analyses
Results of OPC Analysis
The number of discrete aspect patches counted on each ammonoid TIN is shown in Table
A1. To test the correlation of OPC values versus shell size (LMAX, mm), a reduced major axis regression was conducted (Fig. 27). A positive correlation (r = 0.55, p = 0.00) was present between OPC value and shell size.
Because heteromorphic ammonoids are known to have some asymmetry between the left and right flanks of the shell (Bucher et al., 1996), a comparison of the OPC values of whole shells between lefts and rights of each dimorph was conducted. In macroconchs, left H. spedeni and right H. spedeni were shown to have equal means and equal variances in OPC (t = -0.61, p =
0.55; F = 1.90, p = 0.41). However in microconchs, left H. spedeni and right H. spedeni were shown to have equal means but unequal variances in OPC (t = -2.15, p = 0.052; F = 4.55, p =
0.03). The right H. spedeni microconch specimens had both a larger mean (though not statistically significant at the 95% confidence limit) and larger variance in comparison to the mean and variance of the left H. spedeni microconchs. Although these statistical analyses showed that there were significant differences in the variances of OPC between left and right sides of microconchs of H. spedeni, the differences were most likely attributable to my sampling only the best preserved specimens. I therefore concluded that pooling both left and right H. spedeni was suitable to increase the sample size for OPC analyses. Due to the small sample size of H. nebrascensis specimens, a comparison between left and right lateral shell models was not appropriate. 27
When I compared all complete macroconch and microconch specimens belonging to H. spedeni, statistically unequal means but equal variances were detected. Whole H. spedeni macroconchs had statistically bigger mean OPC than the whole H. spedeni microconchs. Figure
28a displays a frequency histogram of OPC from both whole shell macroconchs and microconchs of H. spedeni demonstrating the overlapping variances but the higher mean OPC in the macroconchs.
When the body chamber and phragmocone of H. spedeni were assessed separately for
OPC, the results were different than the results obtained from the shells’ entirety. No significant differences were determined between macroconchs and microconchs of H. spedeni body chambers or phragmocones. However, macroconchs did have consistently larger means and variances in comparison to the microconchs’ relatively smaller means and variances.
Nevertheless, the statistical analyses detected no differences in these distributions, which are shown in Figures 28b and 28c.
I conducted the same statistical tests on OPC of H. nebrascensis whole shells before analyzing body chambers and phragmocones separately as in the tests conducted on OPC of H. spedeni. In all the H. nebrascensis OPC analyses, the macroconchs tended to have larger means but smaller variances relative to the microconchs. However, statistical analyses demonstrated that these differences were not significant. Figures 28d through 28f exhibit the distributions of H. nebrascensis dimorphs.
To compare the OPC between the Hoploscaphites species, I conducted several additional tests comparing the whole shells, separated body chambers and phragmocones. All analyses between the two species respective dimorphs exhibited no statistically significant differences
(Fig. 29). Despite the lack of statistical differences, I do note that the H. nebrascensis specimens 28 constantly had larger means in OPC than the H. spedeni specimens, particularly in the separated body chambers (Fig. 29b and Fig. 29e). Additionally, phragmocones between the two species and dimorphs had the most similar means (Fig. 29c and Fig. 29f).
Results of Surface-to-Planimetric Area Ratio Analysis
The same 49 specimens used for the OPC analysis were also used in the surface-to- planimetric area ratio analysis. Tables A2 – A4 contain the mean, standard deviation and median of all cell values within each rasterized ammonoid model. Unlike in the OPC analysis, lefts and rights of H. spedeni were not compared for consistency due to each ammonoid’s raster cell values having a right-tailed distribution. The skewness of the raster cell values occurred due to the sudden change in slope around the ammonoid shells’ margins (i.e., the keel and umbilicus of the shell having a near vertical slope). I applied a Mann-Whitney U test, a non-parametric test that compares the medians of two non-normal distributions, to the two Hoploscaphites species.
Because the Mann-Whitney U test can only do pairwise comparisons of univariate data, I created separate frequency histograms of the median cell values of whole shell macroconchs and microconchs of H. spedeni using only three bins (Fig. 30a and Fig. 30b). Utilizing three bins allowed for specimens to be grouped into the lowest, the middle, or the highest ends of the spectrum of variability. These three bins of whole shell median values reflected the three morphotypes described by Landman and Waage (1993): a smooth compressed type, an intermediate typical type, and an ornate rotund type. I visually selected a representative from each morphotype bin of both macroconchs and microconchs for the analyses. While the selection of each representative was potentially subjective, binning the H. spedeni specimens into three groups narrowed down the possible choices for each morphotype, which allowed for comparisons to be more appropriately made. Due to the low sample size and presumably more 29 constrained morphology, I selected the H. nebrascensis with the highest and the lowest whole shell medians for comparison.
Given the large number of raster cells for each ammonoid, the whole shells medians were not used for the pair-wise comparisons. Instead, analyses were divided into body chamber comparisons and phragmocone comparisons of the same representative ammonoids between the two species of Hoploscaphites. I predicted that the median cell values of the separated body chambers and phragmocones would reflect their respective whole shell morphotype groups; however, that was not the case for the H. spedeni macroconchs. For the body chambers, the compressed specimens had the lowest median, the rotund specimens had the intermediate median, and the typical had the highest median. Conversely for the phragmocones, the typical specimens had the lowest median, the compressed specimens had the intermediate median, and the rotund specimens had the highest median. Separated body chambers and phragmocones of the microconchs of H. spedeni did in fact have the lowest median for the compressed forms and the highest median for the rotund forms.
Nonetheless, the three morphotypes (again, originally determined by whole shell median value) of H. spedeni were divided into pairs (i.e., typical versus rotund (ornate) and typical versus compressed (smooth)) for both dimorphs (see CHAPTER IV: DISCUSSION). All U* values derived from the Mann-Whitney U test equation were compared against a t distribution table with infinite degrees of freedom. The test statistic for each comparison revealed statistically significant differences in medians for all comparisons made. Table 1 presents these results.
To test the consistency between the surface-to-planimetric area ratio results to the OPC results, I performed a reduced major axis regression (Fig. 31). The results of these two independent tests of surface rugosity/shell complexity were statistically uncorrelated (r = 0.02, p 30
= 0.87). Likewise, a reduced major axis regression of the medians from 3D to 2D ratio analysis to shell size were also uncorrelated (r = -0.08, p = 0.59) (Fig. 32). This lack of correlation between the 3D to 2D ratio to shell size was in direct contrast to the positive correlation of OPC values to shell size (Fig. 27).
2D GIS Analyses
Results of Rib and Tubercle Spacing Analysis
Results of Average Nearest Neighbor analysis of rib and tubercle spacing are presented in
Table 2 (also see Tables A5 and A6). This analysis of rib spacing revealed that 100% of H. spedeni macroconchs and microconchs analyzed had statistically clustered ribs; however, the same analysis applied to tubercle spacing presented mixed results. Individual specimens of H. spedeni had a dispersed, random, or clustered tubercle distribution. Microconchs of H. spedeni were more likely to be evenly spaced, with 63.6% of specimens exhibiting a dispersed distribution of tubercles and 36.4% of specimens exhibiting a random distribution. In contrast, macroconchs of H. spedeni were more likely to be randomly distributed, with 63.6% of specimens exhibiting no detectable pattern to tubercle spacing, 9.09% exhibiting a clustered distribution, and 27.3% having a dispersed distribution. Tubercles on H. nebrascensis macroconchs and microconchs both tended to be statistically dispersed, but H. nebrascensis macroconchs had tubercles that were more strongly dispersed (75% of specimens) than the microconchs (50% of specimens).
Results of Rib Density Analysis
Frequency histograms exhibiting the distribution of rib density at 25%, 50% and 75% whorl height are presented in Figure 33a and Figure 33b. Raw data are presented in Table A7. A higher rib density was present at 25% than 75% of whorl height in macroconchs of H. spedeni. 31
Moreover, macroconchs were more variable in the density of ribs at 25% whorl height than 50% and 75% whorl height. Conversely, microconchs of H. spedeni had a higher rib density difference across the flank of the ammonoid shell than macroconchs, but the microconchs were less variable across the flank than macroconchs. Figures 33a and 33b show that both dimorphs are non-normally distributed, with the macroconchs being unimodal and the microconchs being bimodal. Due to the non-normality of the data, a Mann-Whitney U test for equality of medians and a Kolmogorov-Smirnov test for equality of distributions were conducted to compare the rib density at the different transects between macroconchs and microconchs of H. spedeni. The
Mann-Whitney U test comparing the macroconchs and microconchs of H. spedeni at the 25%,
50% and 75% transects detected no difference in the medians. The Kolmogorov-Smirnov test also detected no difference in the distributions of ribs at the 25%, 50% and 75% transects between macroconchs and microconchs of H. spedeni.
Rib density maps created for the 19 H. spedeni specimens used in this analysis allowed for visualization of the variation in rib density throughout the ontogeny of this taxon (Fig. 23). In some macroconchs, ribs appeared most dense on the dorsal side of adapertural end of the body chamber, but in others, rib density appeared consistent throughout ontogeny (Fig. 34a and Fig.
34b). Microconchs of H. spedeni appeared to have relatively low variability in the density of ribs except for near the ventral portion of the adapertural end of the body chamber, particularly on the recurved hook. Similarly to the macroconchs of this species, some microconch specimens also appeared to have consistent rib density throughout ontogeny (Fig. 34c and Fig. 34d).
Results of Tubercle-Rib Spatial Correlation Analysis
In both dimorphs of H. spedeni, 87% or more of the total number of tubercles observed on each ammonoid correlated with ribs when a 2 mm buffer around each tubercle was 32 used. Because the 2 mm buffers were fairly large given the slenderness and close approximation of the ribs, I present the results of the 1 mm point buffer analysis only (Fig. 24, see also Table
A8). In H. spedeni, the mean percentage of tubercle-rib intersections for macroconchs was
79.3% and for microconchs was 87.2%. The variance of the percentages in macroconchs was
127.5 and the variance in microconchs was 81.9. Although the macroconchs had a smaller mean percentage but larger variance in tubercle-rib spatial correlation, the means and variances were not statistically different between macroconchs and microconchs of H. spedeni (t = -1.68, p =
0.11; F = 1.56, p = 0.52), indicating similar spatial correlations between ribs and tubercles in both dimorphs.
Results of Shell Coiling Analysis
Table A9 displays the body chamber to whole shell area ratios of all specimens analyzed.
Frequency histograms displaying the ratio of body chamber to whole shell area in H. spedeni and
H. nebrascensis macroconchs and microconchs are shown in Figure 35. The two dimorphs of H. spedeni had unequal means but equal variances (Fig. 35a). Specifically, the macroconchs of H. spedeni had a larger mean in body chamber to whole shell area ratio. Conversely, the two dimorphs of H. nebrascensis had equal means but unequal variances (Fig. 35b). Of the two dimorphs of H. nebrascensis, the microconchs had the larger variance in comparison to the smaller variance of the macroconchs.
Comparisons across species were also made. In an assessment of the body chamber to the whole shell area between macroconchs of H. spedeni and H. nebrascensis, equal means but unequal variances were determined (Fig. 35c). In this particular comparison, macroconchs of H. spedeni had a larger variance than the H. nebrascensis macroconchs. On the other hand, microconchs of H. spedeni and H. nebrascensis had both equal means and equal variances. 33
Although not statistically different, H. nebrascensis did have a larger mean and variance relative to H. spedeni (Fig. 35d).
Results of Body Chamber Shape Distribution Analysis
The body chambers of H. spedeni dimorphs were separately scaled to the first landmark pair: the dorsal and ventral sides of the ultimate septum. The ratio of each body chamber’s area to the area of its respective composite polygon morphospace was compared (Table A10). Figure
36 displays frequency histograms of the body chamber area to the area of the composite polygon morphospace for these taxa. Statistical tests revealed equality of means and equality of variances between the macroconchs and microconchs of H. spedeni (Fig. 36a). Equal means and equal variances between dimorphs of H. nebrascensis were also determined (Fig. 36b). However, I note that in both of these comparisons, the microconch specimens had the larger mean, and in H. nebrascensis specifically, a larger variance was also present.
When comparing the body chamber area to the area of the composite polygon morphospace between both species, the means and variances were determined to both be equal for the macroconchs, with the H. spedeni specimens having the higher mean and higher variance
(Fig. 36c). Likewise, the microconchs also displayed equality of means and equality of variances; however, the H. nebrascensis specimens had a slightly higher mean (Fig. 36d).
The second set of landmarks used in this analysis, the dorsal end of the ultimate septum and the dorsal lip of the aperture, revealed similar results as the first landmark pair (Table A11).
Figure 37 displays a frequency histogram exhibiting the distributions of these data. The H. spedeni microconchs had a larger mean; however, the statistical analyses detected no difference in means or variances between both dimorphs of H. spedeni (Fig. 37a). Likewise, dimorphs of H. 34 nebrascensis also had no statistically significant differences in the means and variances despite the larger values in the microconchs (Fig. 37b).
As in the first landmark pair analysis of body chamber area to the area of the composite polygon morphospace, macroconchs of H. spedeni and H. nebrascensis were compared. In the macroconchs, H. nebrascensis had a slightly larger mean but smaller variance in comparison to
H. spedeni. Any differences between the means and variances between both species’ macroconchs were not determined by statistical analyses (Fig. 37c). Microconchs of H. spedeni had a slightly larger mean than the H. nebrascensis microconchs, but both were determined to have equal means and equal variances (Fig. 37d). 35
CHAPTER IV: DISCUSSION
3D GIS Discussion
Interpretation of OPC Analysis
As previously stated, each TIN ammonoid surface used to export aspect patches (Fig. 18) contained extraneous, seemingly randomly oriented polygons around its margins, particularly on the upward sloping ventrolateral tubercles around the body chamber. Additionally, North aspect was divided into two features classes, causing an inflation of the OPC values. Despite these issues causing potentially erroneous results, the OPC values of Hoploscaphites positively correlated with shell size (LMAX, mm) (Fig. 27). Through my own observations and the morphological descriptions documented by Landman and Waage (1993), the largest shells tended to also be the most ornate (i.e., tuberculate). A positive correlation between OPC value and shell size, as expressed by my data, is therefore predicted. Hence, OPC analyses were unlikely to be largely affected by the problems encountered with the aspect TINs.
Of all the comparisons of the OPC values, the statistical analyses detected only a difference between the mean OPC values of whole shells of the macroconchs and microconchs of H. spedeni. Specifically, macroconchs had a larger mean OPC than microconchs of H. spedeni. The different means but same variances suggests that H. spedeni dimorphs have equal spectra of variability (i.e., equally plastic morphologies), but the average forms for each dimorph have distinctly different morphological centers, or in this case, different ornamental features (Fig.
28a). The similarity in the range of variability for both dimorphs of H. spedeni fits Landman and
Waage’s (1993) description that both the macroconchs and microconchs had complimentary morphotypes, but the different means suggests that differences are present between the two dimorphs despite the similarities in their variability. The different means may be attributable to 36 the inflation of the macroconch body chamber, which exhibits more space for larger tubercles and more ribs to form during development, potentially increasing the mean OPC in macroconchs of H. spedeni.
The comparison of OPC values of only the body chambers of H. spedeni macroconchs and microconchs, on the other hand, did not demonstrate the differences expressed by the whole shell analysis. The statistical analyses revealed equal means and variances of body chamber OPC
(Fig. 29b); however, visually, the frequency histograms of body chambers look very similar to the frequency histograms of the whole shells of H. spedeni. Additionally, the p-value for statistical significance for the Student’s T test for equality of means was equal to 0.08, a value very close but not reaching, the 95% confidence limit. The unexpected equality of body chamber
OPC between dimorphs of H. spedeni may have been due to the sampling bias of utilizing only the best preserved specimens for 3D analyses. Of course, the extraneous polygons around the
TINs’ margins and the two North aspect feature classes may have affected the data by altering the OPC values as previously stated.
As expected, the phragmocones of both dimorphs of H. spedeni had equal means and variances for OPC (Fig. 28c). Given that much of the morphological change between the dimorphs, most dramatically observed in macroconchs, occurs during body chamber development, it is reasonable to predict that the phragmocones would be more morphologically similar to each other than the body chambers. My data supports this prediction given the equality of the means and variances between both dimorphs. The phragmocones of both macroconchs and microconchs of H. spedeni have relatively widely spaced ribs with small tubercles on every rib to every other rib. I observed that this morphology is consistent among specimens regardless of which dimorph is examined. The similarity of the OPC values among the phragmocones 37 demonstrates that even though this species contains a high degree of intraspecific variation, some degree of morphological constraint is present in the juvenile stages when the phragmocone forms, and major changes only occur during the latest stages of ontogeny.
When the H. nebrascensis macroconch and microconch specimens were compared for
OPC values, the statistical analyses detected no differences. Visually, the separated body chamber frequency histogram (Fig. 28e) is very similar to the whole shell frequency histogram for H. nebrascensis (Fig. 28d); this same visual pattern is presented with H. spedeni body chambers and whole shells and further supports the interpretation that both species’ OPC distributions were mostly swayed by the body chamber OPC values. Much like the OPC value distributions for the dimorphs’ phragmocones of H. spedeni, the OPC values of H. nebrascensis phragmocones were very similar between dimorphs. This similarity again suggests a heavily constrained juvenile morphology with much of the variation in ammonoid ontogeny occurring during the formation of the body chamber.
Figure 29 demonstrates that the range of OPC values for H. spedeni and H. nebrascensis overlapped. Specifically in the macroconch whole shells and body chambers, H. nebrascensis specimens fell within the range of variability but close to the higher end of the spectrum of H. spedeni. Additionally, the variances of the macroconch body chambers of H. spedeni and H. nebrascensis were near, but did not exceed, the 95% confidence limit for statistical significance
(p = 0.08). Although the variances were determined to be the same, H. nebrascensis macroconchs had much more limited OPC values (i.e., more constrained ornamentation) than H. spedeni macroconchs (Fig. 29a and Fig. 29b). Equal variances may be attributable to the small sample size of H. nebrascensis compared to H. spedeni in this analysis; more samples of either taxa, especially H. nebrascensis, may allow for statistical analyses to determine whether 38 statistically significant differences in variances are actually present. If sample size is increased and statistically significantly different variances are determined, the narrow morphological variability in H. nebrascensis macroconchs compared to H. spedeni macroconchs would fit
Landman and Waage’s (1993) description of these taxa. The phragmocones of the two species macroconchs had nearly identical OPC distributions, which further demonstrates the developmental constraint in juvenile Hoploscaphites.
In the microconchs of both species, the whole shell and body chamber comparisons were consistent with the macroconch comparisons previously made; microconchs of H. nebrascensis tended to fall toward the higher end of the variability of H. spedeni microconchs. However, the microconchs of H. nebrascensis had a slightly broader variance, though not significantly different from microconchs belonging to H. spedeni. Demonstrating a constrained morphology, the microconch phragmocones of both species reflected similar distributions. According to these data, differences in the degree of ornamentation between the two species of Hoploscaphites are mostly not present until the later ontogenetic stages, when the body chambers begin to form sex- specific ornamentation and shell shapes.
The TIN aspect models not only allow for statistical analyses to be applied to univariate, tabulated data, but also allow for the different degrees of ornamentation between the two
Hoploscaphites species to be visually compared. For example, ammonoid ornamentation comprised of ribs with multiple rows of appended tubercles, like that observed on the phragmocone of the rotund H. spedeni, exhibited all aspect classes along each tubercle-bearing rib. However, if no tubercles were present on certain regions of the shell, such as on the H. spedeni body chamber adapical of the recurved hook, then only two aspect classes were observed representing the two different sloping surfaces of individual ribs. In H. nebrascensis, the 39 tuberculate morphology was maintained into full sexual maturity, and all aspect classes were present along these heavily tuberculate ribs. Figure 18 demonstrates how these ornament types can be visualized by viewing surface aspect within GIS. Visualizing the change in slope direction may even help to verify species identification within genera where species are only distinguished by dissimilarities in ornamentation across the lateral surface of the shell. Visualizing the global patterns of surface aspect, that is, observing the changes in ornamentation over an entire shell’s surface, is one advantage of utilizing a geospatial approach to compare specimens.
Interpretation of Surface-to-Planimetric Area Ratio Analysis
While all comparisons were determined to be statistically different, the 3D:2D area ratio analysis and the OPC analysis had similarities. In H. spedeni macroconchs, the typical and compressed morphotypes had relatively low U* values (relatively more similar median cell values). Given that the typical H. spedeni macroconch often has very few flank tubercles on the phragmocone and none on the body chamber except for scattered tubercles on the recurved hook, and that the compressed morphotype has almost no tubercles on the exposed phragmocone and body chamber, I determined that the low U* values are attributable to very similar ornamentation and shape of these two morphotypes. Although the U* values of the phragmocone comparisons of H. spedeni macroconchs were statistically significant at 95% confidence limit (U* = 4.91), the relatively low U* values indicates that, at least to some degree, the phragmocones are more similar than the body chambers between the typical and compressed morphotypes. Likewise the large U* value (200.95) between typical and rotund H. spedeni macroconch body chambers and the relatively small U* value (30.81) between typical and rotund H. spedeni phragmocones further supports the interpretation of a constrained juvenile morphology with differences expressed mostly in the late ontogeny during body chamber development. 40
In H. nebrascensis macroconch body chambers, a comparison of the highest median value and the lowest median value revealed a large U* value (204.08). However, when the macroconch phragmocones were compared, a relatively low U* value (9.70) was determined.
These statistical analyses demonstrate that H. nebrascensis macroconchs are more constrained in their earlier ontogeny during phragmocone formation. Additionally, microconchs of both H. nebrascensis and H. spedeni seemed to have fairly similar U* values between body chambers and phragmocones, exhibiting the slightly more constrained early ontogeny of the microconchs between both species.
A comparison of the macroconch body chamber of the rotund H. spedeni and lowest and highest median value H. nebrascensis specimens revealed the most similar results to the OPC analysis. The rotund H. spedeni specimen in both comparisons of body chamber and phragmocone had a more similar median to the highest median value H. nebrascensis specimen than to the lowest median H. nebrascensis specimen. I infer that most of the H. nebrascensis macroconch specimens fall within the spectrum of variability of H. spedeni macroconchs, with the highest end members of both H. spedeni and H. nebrascensis having potentially similar morphologies (Fig. 11).
Interestingly, when the 3D:2D area ratio rasters were divided into body chambers and phragmocones and analyzed separately for ease of calculations, the representative H. spedeni macroconchs did not fit the expected order (Fig. 30). In other words, the rotund H. spedeni body chamber did not necessarily have the highest median value even though this same rotund specimen did when it was assessed in its entirety. This observation suggests that large changes in ornamentation and shell shape occur during body chamber formation. The timing and duration in the expression of characters through ontogeny seems to be a major factor that needs more 41 attention in order to successfully describe, visualize and quantify the morphotypes within and across Hoploscaphites species (Yacobucci, 2012).
The reduced major axis regression shown in Figure 31 shows that the results between the whole shell 3D:2D area ratio median values and the OPC values did not correlate, suggesting that these two analyses are quantifying two different characteristics of the shell morphology. I suspect the lack of correlation between these two independent surface rugosity analyses has to do with a number of factors. Transforming the TIN surface into a DEM raster caused some of the surface resolution to be lost. For example, features of low relief (e.g., ribs) were not always discernible after TIN surfaces were converted to DEM rasters (Fig. 19). Adjusting the raster cell size to be smaller had no profound effect on the resolution of the ammonoid DEM. Additionally, the flanks of these ammonoids are relatively flat, even more so in H. nebrascensis specimens.
Given that not all ornamental features are preserved in the conversion from TIN to DEM raster, the 3D:2D area ratio data are most likely demonstrating variations in larger scale features, such as the shape of the flanks (i.e., how flat or steeply curved the lateral surfaces are).
The marginal cells around each DEM exhibited substantially higher 3D:2D area ratio values (e.g., greater than 10) due to the near vertical slope of the keel. On the other hand, the values of the horizontal flank surface were all consistently small (e.g., between 1 and 2). The manner in which the 3D:2D area ratio values change over the ammonoid DEM surface supports that this metric is assessing characteristics of shell shape and not necessarily ornamentation. The reduced major axis regression in Figure 32 also shows that the 3D:2D median values did not correlate with shell size (LMAX, mm), which further supports my inference that the OPC metric and the 3D:2D area ratio metric are quantifying two different features of the shell morphology.
For instance, the larger taxon, H. nebrascensis, often has the most ornate morphology; however, 42 these large specimens do not necessarily have the most steeply curved flanks. In other words, their compressed shell shape, and not ornamentation, was detected by the 3D:2D area ratio analysis.
The 3D:2D area ratio rasters presented in Figure 19 illustrate how the overall curvature of the shells’ flank surfaces weighed more heavily than ornamentation in this particular rugosity analysis. Visually, most of the lateral surface of the shells comprised of the lowest raster class values, indicated by the white to light purple. Near the ventrolateral shoulder where the curvature of the flanks steepen, the raster class values became higher, indicated by warmer colors (i.e., closer to red). I predict that the rotund H. spedeni macroconchs would have the least flat, or compressed, flanks near the ventrolateral shoulder, such that the warmer colors (i.e., a darker purple) would be visible along the shoulder. In contrast, H. nebrascensis specimens should have cooler colors near the ventrolateral shoulder (i.e., more white to light purple across the entire flank) due to their more compressed shell morphology. My prediction is confirmed by visual inspection of the 3D:2D area ratio rasters, demonstrating that this global geospatial rugosity analysis allows for both quantitative and visual classification of shell shape, specifically, shell compression.
2D GIS Discussion
Interpretation of Rib and Tubercle Spacing Analysis
According to the Average Nearest Neighbor analysis, ribs were 100% clustered in H. spedeni regardless of the dimorph. Considering that this analysis assessed the center of each rib and its approximation to the next rib center, the results indicat that ribs are always formed in pairs. Pairing of ribs may imply either the presence of secondaries, particularly secondaries branching off near the midpoint of each primary rib, or that primary ribs form in closely 43 approximated pairs with space between each set of paired ribs. Given the description of
Landman & Waage (1993), the former implication is more likely than the latter since H. spedeni has secondaries branching off of primaries approximately halfway between the umbilicus and ventrolateral shoulder. The consistency in the clustering of the ribs in all specimens used for the analysis demonstrates that the morphogenetic process of rib formation is fairly conservative in H. spedeni, with the formation of a primary rib almost always inducing the development of a secondary rib in both macroconchs and microconchs throughout ontogeny.
The mixed results for the tubercles demonstrate that tubercle spacing and development are much more variable and plastic than rib spacing. Macroconchs of H. spedeni seemed to have the most variably distributed tubercles, which were either dispersed, random or clustered (Fig.
22). Most of the macroconch specimens (63.6%) had randomly distributed tubercles, some exhibited dispersed tubercles (27.3%) and even fewer exhibited clustered tubercles (9.1%). In H. spedeni macroconchs, tubercles often appear evenly spaced on every rib to every other rib on the phragmocone, diminish in number and spacing on the body chamber, and reappear again on the recurved hook in a scattered distribution (Fig. 20 and Fig. 24). Hence, these percentages of dispersed, random and clustered tubercle distributions are consistent with my own observations and with the description made by Landman & Waage (1993). The scattered tubercles on the recurved hook of the body chamber in some macroconchs most likely skewed the results towards a more random distribution despite the visually dispersed tubercles on the phragmocones. Unlike the H. spedeni macroconchs, the H. spedeni microconchs lack the scattered flank tubercles on the recurved hook, often sustaining one to two rows of evenly spaced tubercles until the adapertural end of the body chamber, so a dispersed distribution of tubercles is expected and confirmed by my data. The results obtained for the H. spedeni microconchs reinforce my interpretation of the 44 macroconchs in that most of the microconch specimens (63.6%) had a dispersed tubercle distribution due to the lack of scattered tubercles on the recurved hook.
In H. nebrascensis macroconchs, tubercles were nearly all dispersed (75%) with only one macroconch exhibiting a random distribution (25%). Visually, H. nebrascensis have more plentiful, relatively evenly spaced tubercles aligned in five to seven rows throughout ontogeny except for near the adapertural end of the body chamber (Fig. 20). The Average Nearest
Neighbor analysis was able to detect the near uniform spacing of the tubercles in the macroconchs. On the other hand, microconchs of H. nebrascensis were more variable in that only half the specimens had dispersed tubercles, and the other half had either clustered and random tubercles. This high variation in microconch tubercle distribution reflects the slightly higher variance in OPC for H. nebrascensis microconchs than macroconchs.
Knowing that the phragmocones’ tubercles in H. spedeni specimens appeared to be evenly dispersed, I interpreted these results to imply that H. nebrascensis adults have similar genetic expressions as H. spedeni juveniles. It should be noted, though, that the relatively low sample size for H. spedeni and even smaller sample size for H. nebrascensis specimens could affect these results. Additionally, Average Nearest Neighbor analysis is known to sometimes produce erroneous results when the features are near the edges of the mapping area.
Ventrolateral tubercles of both Hoploscaphites taxa were near the mapping edges, which means interpreting these data must proceed with caution, especially since some of the H. spedeni specimens only have ventrolateral tubercles.
Interpretation of Rib Density Analysis
The Mann-Whitney U and Kolmogorov-Smirnov tests applied to the spiral transects identified no differences in rib density across the shells’ flanks. These results suggest that within 45 each dimorph of H. spedeni, ribs, most likely secondaries, are inserted during growth to fill the added shell area, which allow for similar results for the rib density at 25%, 50% and 75% whorl height.
While there were similarities in the central tendency of rib density, differences in the variance were present across the shells’ flanks. For example, macroconchs had a significantly higher variance (8.08) in rib density at 25% whorl height; conversely, microconchs had a fairly low variance (1.67) in rib density at 25% whorl height. Macroconchs are easily distinguished from microconchs because of their inflated body chamber shafts, which variably have bulging umbilical shoulders to straight umbilical shoulders. The different macroconch body chamber morphology allows for more ribs to be inserted in some specimens (Fig. 34a) but not others (Fig.
34b), which explains some of the high variance at 25% whorl height. Additionally, some macroconchs of H. spedeni also seem to have very closely approximated ribs near the dorsal side of the recurved hook, which most likely explains the rest of the variance at 25% whorl height.
Microconchs were much less variable in rib density across the flanks, but what little variance was present was concentrated at the 75% whorl height near the ventrolateral shoulder. A closer examination of the rib density maps revealed the higher rib density on the ventral side of the flank of the recurved hook in some, but not all, microconch specimens of H. spedeni (Fig. 34c and Fig. 34d). Unexpectedly, some microconch specimens even had closely approximated ribs at the base of the body chamber, suggesting that at least in some microconchs, shell accretion may slow before development of the body chamber resumes (Fig. 34d).
The more concentrated rib morphology at the adapertural end of the body chamber in both dimorphs suggests that rib formation is more tightly constrained than other features such as shell shape. The presumed retardation of shell growth during the formation of the recurved hook 46 as full sexual maturity drew near had little effect on the expression of the genes responsible for rib formation. In other words, the ribs appear to form at a constant rate despite the lower accretion rate of the shell, causing ribs to be closely approximated at the adapertural end of the body chamber.
Interpretation of Tubercle-Rib Spatial Correlation Analysis
As stated throughout this work, tubercles appear to be found on every rib to every other rib in H. spedeni. The tubercle-rib spatial correlation analysis confirmed that tubercles were more likely to be observed on ribs. In the correlation analysis utilizing a 1 mm buffer, both macroconchs and microconchs had most tubercle buffers intersected by at least one rib polyline; the mean tubercle-rib intersections were 79.3% and 87.2% tubercle-rib intersections, respectively. If tubercles always form on ribs, then I would expect the tubercle spacing to reflect the rib spacing and 100% of tubercles to correlate with ribs. However, tubercles are not found on every rib, and only a few specimens had clustered tubercles in macroconchs and none in the microconchs, even though the ribs of both dimorphs were 100% clustered (as determined by the
Average Nearest Neighbor). The higher variance of the tubercle-rib spatial correlation in the macroconchs did reflect the tubercle spacing analysis, which showed mixed dispersed, random and clustered tubercles. These distributions are mostly due to the fact that tubercles are not necessarily found on every rib. Sometimes tubercles are present on every primary or every other primary with strong secondary ribs in between. I interpret these results to mean that the formation of a rib (i.e., the expression of the genes for rib development) may activate or induce the genes downstream in the regulatory pathway responsible for forming tubercles; however, tubercle formation does not necessarily induce expression of ribs. Given that ribs are so prevalent in both Hoploscaphites taxa (recall that ribs were not quantified in H. nebrascensis 47 because they were obscured by tubercles, not because they were absent) and that tubercles disappear and reappear throughout the exposed shell, particularly on the body chamber of H. spedeni, variations in the genetic program related to how strongly rib formation induced tubercle formation seem the most likely mechanism in which large scale morphological variation in ornamentation becomes present in H. spedeni.
Interpretation of Shell Coiling Analysis
In the shell coiling analysis, H. spedeni macroconchs had a similar degree of morphological variation as H. spedeni microconchs, but the central tendencies (i.e., the means) of their morphologies were significantly different; the macroconchs had a larger mean and tighter coiling. Moreover, macroconchs and microconchs of H. nebrascensis had similar means but quite different variances; the microconchs had a subtantially larger variance than the macroconchs. These results suggest that macroconchs of H. spedeni have a broader range of shell coiling than macroconchs of H. nebrascensis, with some H. spedeni specimens exhibiting a tightly coiled shell and others exhibiting a loosely coiled shell. The larger mean but more narrow variance in the H. nebrascensis macroconchs suggests that this descendant taxon coiled its shell much like the most tightly coiled H. spedeni macroconchs. Additionally, microconchs of H. nebrascensis were not only more variable than the macroconchs, but they were also more variable than H. spedeni microconchs, suggesting that morphological variation is maintained in microconchs, if not increased, during the evolutionary divergence of H. spedeni and H. nebrascensis. Since microconchs do not have an inflated body chamber like the macroconchs, fewer developmental constraints on body chamber shape and coiling are presumably present in both species, which may have allowed for the variation in shell coiling to persist. 48
According to the First Buckman Law of Covariance, shells that are more heavily ornamented tend to be more evolute (i.e., loosely coiled) and rotund than less ornate shells.
However, the more tuberculate H. nebrascensis had an equally involute shell to some of the tightly coiled H. spedeni, particularly in the macroconchs. These data seem to violate Buckman’s
Law at first glance, since one would expect H. nebrascensis to have a more evolute shell due to its heavy ornamentation. A closer examination of the shell morphology revealed how H. nebrascensis specimens are able to maintain a tightly coiled, compressed shell that is more ornate than those shells belonging to H. spedeni. At least in the macroconchs of H. nebrascensis, relative whorl height appears to be larger throughout the exposed phragmocone, which allowed for more rows of flank tubercles to be inserted onto the shell. In addition to increased whorl height of the phragmocone, the size of the flank tubercles is reduced in comparison to the slightly broader, more widely spaced flank tubercles of H. spedeni (Fig. 8 – 11).
Interpretation of Body Chamber Shape Distribution Analysis
No significant differences were detected when the body chamber polygons were scaled to the same landmarks and ratioed to the merged Hoploscaphites body chamber polygon, which represented the total morphospace the body chambers could fill. The first landmark pair, the dorsal and ventral sides of the ultimate septum (i.e., the base of the body chamber) (Fig. 26a and
Fig. 26b), demonstrate that a high range of variability in the placement of the ultimate septum is present between macroconchs and microconchs of both species, since many body chambers were rotated to be placed and scaled at the first landmark pair. Despite the lack of significant differences, macroconchs of H. nebrascensis had a more narrow range of variability in the placement of the ultimate septum than the H. spedeni macroconchs. The less variable morphology reflects the results obtained from previous analyses of a more constrained 49 morphology in H. nebrascensis. Statistical significance may not have been detected either because sample size for H. nebrascensis is too low, or because the variances in the placement of the ultimate septum are close enough to not be detected at the 95% confidence limit. In the microconchs, H. spedeni and H. nebrascensis specimens had similarly broad variances, which suggest that the high degree of morphological variation in H. spedeni microconch body chamber shape is conserved in the descendant H. nebrascensis microconchs.
For the second landmark pair, the dorsal side of the ultimate septum and the dorsal lip of the aperture (Fig. 26c and Fig. 26d), similar results were obtained as from the first landmark pair.
No statistically significant differences were detected between the H. spedeni and H. nebrascensis macroconchs, even though the latter taxon had a more narrow variance relative to the former.
However, the variances were near the 95% confidence limit (p = 0.09), suggesting that the low sample size of H. nebrascensis may be affecting the results. Between the microconchs of both species, the distributions of body chamber area to composite Hoploscaphites body chamber morphospace area are nearly identical visually (Fig. 37d). This suggests that similarly large ranges of variance are present between the microconchs of both H. spedeni and H. nebrascensis; on the other hand, H. nebrascensis macroconchs lose this broad range of body chamber shape while their respective microconchs retain the ancestral broader morphological range.
The second landmark pair also showed that some degree of isometry is present in the two
Hoploscaphites taxa. Larger specimens often were equally proportionate to smaller specimens when scaled to the landmarks. This proportionality is suprising given that H. nebrascensis macroconch specimens appear to have disproportionately large whorl heights in the phragmocones compared to phragmocones belonging to H. spedeni macroconchs. A reasonable prediction would be that the macroconch body chamber whorl heights of H. nebrascensis would 50 also be disproportionately large compared to the H. spedeni macroconch body chamber whorl heights, but this prediction is not borne out by my results. The observation that the phragmocone is more compressed (flattened) in H. nebrascensis macroconchs might explain why the phragmocones appear to have larger whorl heights in both the shell coiling analysis and the body chamber shape distribution analysis, following Buckman’s Law.
Evolutionary Implications
The results obtained from these GIS-based spatial analyses are consistent with the interpretation that H. nebrascensis is a paedomorphic descendant of H. spedeni (Landman &
Waage, 1993; Landman et al., 2013). Several studies have shown that heterochronic processes, especially paedomorphosis, are some of the primary mechanisms for creating large scale variation and rapid speciation (Landman & Geyssant, 1993; Landman & Waage, 1993;
Yacobucci, 1999). Specifically, a prolonged and up-regulated expression of the H. spedeni phragmocones’ ornamentation (i.e., tubercles) into the development of the body chambers allows for the rotund H. spedeni morphotype to form. This same process may have been the mechanism by which some H. spedeni ammonoids develop the strong tuberculate morphology to give rise to the H. nebrascensis taxon. Given the consistently compressed but ornate morphology of H. nebrascensis, the question of which morphotype of H. spedeni, the compressed, smooth form or the rotund, ornate form, is the direct ancestor of H. nebrascensis still remains. Did the compressed H. spedeni persist in its expression of tubercles from earlier ontogenetic stages
(possibly from earlier whorls developed prior to those formed found on the exposed phragmocone), which resulted in the splitting of H. nebrascensis from H. spedeni? In contrast, did the rotund morphotype of H. spedeni become more compressed while reducing the size and 51 spacing of tubercles and increasing the number of tubercle rows to split off into the subsequent
H. nebrascensis?
Recently recovered intermediate forms discovered just below the “transition concretions”
(Fig. 11 and Fig. 12) may provide an answer to these questions. Landman et al. (2013) document that some Hoploscaphites specimens recovered from the highest strata of the H. nicolletii Range
Zone boundary exhibit ornamentation (i.e., tubercles) similar to that observed on rotund H. spedeni specimens. The large and bullate shape of the widely spaced flank tubercles, which are aligned in approximately three rows, caused Landman et al. (2013) to classify the specimen as H. spedeni. However, the high degree of tuberculation is very reminiscent of H. nebrascensis specimens, which are stratigraphically observed just above the “transition concretions” that mark the boundary between the H. nicolletii Range Zone and the H. nebrascensis Range Zone (Fig.
12). In Landman and Waage (1993), the authors describe the top assemblage zone, POAZ, of the
H. nicolletii Range Zone as having the “…rotund form…second to the typical form…”
Additionally, they claim that H. nebrascensis is most closely related to more compressed variants of H. spedeni due to the relative flatness of the flanks observed in both species.
However, Yacobucci’s (2004) research on acanthoceratid ammonoids from the Cretaceous
Western Interior Seaway concluded that shell shape, in this group at least, is much more variable than ornamentation, and that ornamentation may have been controlled by the genetic- developmental program of these cephalopods while shell shape may have been a function of the physical environment. Hence, I find the scenario in which rotund H. spedeni changes its shell shape while retaining its ornamentation to be more likely.
This idea is borne out when considering environmental changes occurring with this evolutionary transition. The regression of the Western Interior Seaway during the Maastrichtian, 52 as observed from the lithofacies transition of deep sea deposits of the Pierre Shale to the shallow, coastal deposits of the Fox Hills Formation, would have favored more compressed shell morphologies. Compressed shells with smaller ornamentation have been shown to reduce drag
(Jacobs et al., 1994; Jacobs & Chamberlain, 1996). Therefore, one would expect that ammonoids living in higher energy environments, such as shallow seas, would have more compressed shells.
If the degree of shell compression easily fluctuated to reflect the physical environment, then the rotund, ornate H. spedeni (and not the compressed, smooth H. spedeni) most likely is the ancestor of H. nebrascensis. This hypothesis also fits the stratigraphic distribution of the dominant morphotypes of H. spedeni in the youngest rocks of the H. nicolletii Range Zone.
In general, ornament seems to be relatively more constrained than shell shape between these two taxa; however, ribs are even more constrained than tubercles. As suggested above, each ammonoid specimen consistently has widely spaced ribs on the phragmocone and closely approximated ribs on the body chamber, especially near the aperture. These ribs are always measured and observed in pairs (clustered). These constrained morphological features are in stark contrast to the shells’ tubercles, which are often associated with ribs, but not necessarily connected to every rib. Again assuming that ornamentation is controlled internally, as opposed to externally, “switching on” the genes responsible for rib formation most likely activated the developmental regulatory genes or the constructional genes for tubercle formation. Prolonging the expression of “juvenile” tubercles well into the development of the body chamber may have been a primary mechanism to increase variation within H. spedeni. Physical conditions of the environment may have only served to fine-tune the ornamentation and shell shape in the subsequent H. nebrascensis before locking-in and reducing variations in genetic expression through development that caused the large-scale variation in its predecessor. 53
The GIS-based spatial analyses demonstrate using quantitative methods that the macroconch body chamber is the target for evolutionary change from the ancestral H. spedeni to its sister taxon and descendant, H. nebrascensis, as evinced by the loss of overall morphological variation and the retention of juvenile traits into body chamber formation. Microconchs of H. nebrascensis appear to retain their ancestral developmental plasticity in shell characters. The microconchs may have had fewer constructional constraints due to the lack of an inflated body chamber. The interplay between the morphogenetic expression of tubercles and ribs may be an important evolutionary mechanism for all Hoploscaphites taxa and the closely related genus,
Discoscaphites, which have large tubercles similarly associated with ribs, such as those observed on H. spedeni and H. nebrascensis specimens (Landman & Waage, 1993).
GIS as a Morphometric Tool
GIS was successful in quantifying the morphology of the two taxa in question. Of course, this methodology could be improved or refined for future studies. Specifically, a few problems and limitations made themselves apparent while working with the GIS-based methodology.
For example in using 3D geospatial analyses, photogrammetric software used in creating
3D digital models of the Hoploscaphites specimens appeared to give visually accurate results, but after closer examination, some of the digital models had minor errors in how their lateral surfaces were reconstructed. Highly tuberculate specimens of H. spedeni had tubercles that did not appear correctly modeled. The photogrammetric software package, 123D Catch, was able to
“understand” that rough surfaces were present where tubercles were observed, but slight variations in the shape of individual tubercules were often not properly represented (Fig. 15 and
Fig. 16). I hypothesize that the outer nacreous layer of these ammonoids may appear as different colors depending on the camera angle and light source, making the same tubercles seem like 54 different features. A simple solution to avoid this problem in future studies may be to powder coat the specimens so that colors and light refractions from the nacreous layer are not influential in the calculations made by the photogrammetric software to produce accurate morphological surfaces.
Additionally, this same software package reconstructed a few ammonoid shells with some of the white bean bag pillow attached to the keel of the shell, which only affected the lateral surface directly adjacent to ventrolateral tubercles. A white bean bag pillow was opted over a black one so that the margins of the shells could be easily seen in photographs for vector digitizing in 2D geospatial analyses. Simply using a black pillow (along with powder coating the ammonoid specimens) may have allowed for the photogrammetric software to distinguish differences between the shell and its immediate surrounding environment (the pillow).
Accounting for these different conditions may allow for more accurate models to be generated. To test the ability of the photogrammetric software to produce accurate 3D digital representations, multiple trials with different photographs should be made and the results visually compared so that the best possible model for 3D geospatial analyses is selected. If access to specimens is time limited, a future researcher wanting to employ 3D geospatial analyses could merely take numerous photographs (400 or more, for example) and conduct multiple trials utilizing a smaller, random subset of the total number of photographs. This method for testing the accuracy of the photogrammetric software would be faster than performing five or more independent trials with different sets of 100 to 200 photographs per specimen.
The 3D geospatial analyses used in GIS software could also be easily adjusted and refined. The TIN model exhibiting aspect had North divided into two feature classes: east of 0° 55 and west of 0°. Using the Merge Tool within ArcGIS 10 would have combined all the North aspect polygons into one composite polygon instead of only merging adjacent North aspect polygons while maintaining the true number of North patches. However, if computer script could be designed to merge these adjacent North aspect polygons autonomously, then the default setting of the aspect class breaks can be easily eluded. I plan to use computer script in future studies when time constraints are less pressing.
Besides OPC and 3D:2D area ratio analyses, which are relatively scaleless, other scale- based 3D GIS-based spatial analyses could be explored in future work to expand the toolkit of this method as a morphometric tool. For instance, the standard deviation of elevation, slope variability, and various other methods to generate terrain roughness indices are all options available to quantify surface roughness given that the 3D models are properly scaled. Cooley’s
(2013) website provides an extensive list of various sources on how to calculate surface rugosity metrics using different methods in ArcGIS software.
Methods utilizing 2D GIS-based analyses assumed that the ammonoid photographs were perfectly georeferenced and that parallax had little effect on the spatial statistics obtained from the vectorized features. Parallax is a potential problem since the corners of the photographs were used as ground control points. However, I had confirmed measurements after georeferencing the ammonoid images and the scaling appeared adequate. To better refine the 2D GIS techniques, photographs should be taken from the same location above each specimen, and this location should be selected to reduce parallax so that specimens would not be digitized inaccurately.
Another problem with utilizing 2D GIS-based analyses was that errors introduced while vectorizing ammonoid features in 2D could lead to inaccuracy in the results. Ammonoids are obviously 3D forms, and digitizing features as 2D vectors may not be adequate for most 56 ammonoids. Luckily, both taxa used in this analysis have relatively flat flanks, especially in the more compressed H. nebrascensis, which allow for 2D analyses to give fairly accurate results given the sample sizes used.
Every fossil specimen was once a living organism that had a continuous, 3D morphology.
Traditional morphometrics (e.g., lengths, widths, etc.) are incapable of holistically quantifying this morphology and the changes that occur through ontogeny. Additionally, methods such as landmark geometric morphometrics can only assess differences in the distributions of homologous points, if such points can even be defined. Geospatial-based analyses, on the other hand, can be used to quantify differences over the entire shell surface. With refinement of GIS- based spatial analyses, most ammonoid morphologies can be quantified and visualized globally using this new technique.
The versatility of GIS-based methods allows for many different kinds of ammonoids, or other types of fossil organisms, with complex morphologies to be quantified and visualized through their full ontogeny. This new GIS-based technique may even allow for characters and character states previously undiscovered to be defined and coded for phylogenetic analyses. The power of geospatial analyses to address questions in paleobiology and morphology has only been explored by a mere few, and more researchers should investigate the potential of this technique in quantifying complex morphologies. 57
CHAPTER V. CONCLUSIONS
According to GIS-based spatial analyses, H. nebrascensis specimens have a morphology that falls within the extremely large variability spectrum observed for H. spedeni. The 3D geospatial analyses employed here, such as OPC analysis and 3D:2D area ratio analysis, allow for not only quantification but also visualization of the range of ornamentation and shell compression through late ontogeny between specimens. The 2D geospatial analyses allow for quantification and visualization of more specific shell features, such as the spacing of tubercles, the density of ribs, the variations in body chamber shape, etc. While 2D geospatial methods may oversimplify the ammonoid morphology, the results of these methods are sufficient for addressing paleobiological questions such as intraspecific variation and developmental plasticity.
The target for evolutionary change between the ancestral H. spedeni and the descendant
H. nebrascensis resides in the macroconch body chamber. The body chambers of the macroconchs are the most plastic, particularly in the degree of tuberculation. The genes responsible for tubercle formation may have become up-regulated with a sustained expression through ontogeny, causing more ornate specimens belonging to H. spedeni to appear in the fossil record. The ornate, rotund H. spedeni macroconchs most likely became more compressed during the regression of the Western Interior Seaway since shell shape, particularly compression, has been shown to be more plastic than ornamentation and is influenced by the physical environment in which the ammonoid species lives. Unlike the macroconchs, microconchs of H. nebrascensis maintain their variability through their full ontogeny. Therefore, the target for evolutionary change resides in the macroconchs of H. spedeni, resulting in the speciation of H. nebrascensis.
The macroconchs of the descendant taxon exhibit a relatively more constrained morphology during the morphogenetic countdown to full sexual maturity, suggesting that the developmental 58 pathway became canalized in only the macroconchs, possibly in response to the strong selective pressure exerted by the physical environment.
GIS-based spatial analyses are an excellent new tool for paleontologists to exploit for morphometrics. Utilizing geospatial analysis permits researchers to quantify and visualize fossil form in not only 2D, but also 3D, without the requirement of one-to-one homology that is necessary for traditional and landmark-based morphometric methods. One of the greatest advantages of using a GIS framework to analyze specimens is that the fossil form can be incorporated and analyzed globally to allow complex morphologies to be studied in a different way. Although refinement of the techniques used is still necessary, this study demonstrates the versatility of GIS to quantify and visualize the variability of complex forms to address paleobiological questions. In future work, I plan to not only refine these methods, but also to expand the toolkit for GIS-based morphometrics. 59
CHAPTER VI. FIGURES AND TABLES
Figure 1. Example of ammonoid intraspecific variation. Representatives of the Triassic Svalbardiceras spitzbergensis from a single concretionary layer in Siberia exhibiting a broad spectrum of intraspecific variability in different morphological characters (modified from Weitschat, 2008). Note the range of variability between specimen A and specimen L, and increasing ornamentation with increasing coiling radius (First Buckman Law of Covariance). 60
Figure 2. Heterochrony as an evolutionary mechanism. Illustration of how heterochrony can create different forms through paedomorphosis (progenesis and neoteny) and peramorphosis (hypermorphosis and acceleration) (modified from Landman & Geyssant, 1993). Small black arrow along keel of ammonoid shell represents the last septum, and therefore, the base of the body chamber.
61
Figure 3. Raup’s parameters. Measurements used to calculate Raup's (1966, 1967) parameters illustrated on an ammonoid shell (modified from Gerber et al., 2007). Shell diameter (D), shell radius (R) from the coiling center, whorl height (Wh) and umbilicus diameter (U) are all labeled. Numbers 1-4 represent placement of landmarks used for measurements.
62
Figure 4. Example of 3D GIS-based analyses on fossil form. Digital 3D models and Digital Elevation Models (DEMs) of Gorilla gorilla molar crowns. A, Digital models of molar crowns obtained through 3D laser scanning that were used to generate DEMs; B, Slope data obtained in GIS software from DEMs (higher frequency colors represent steeper slopes); C, Topographic aspect data obtained in GIS software from DEMs (various colors represent different modal aspect for each cusp); and D, Surface area data obtained in GIS software from DEMs (red contour lines to show relief and variations in surface area) (modified from Ungar & Williamson, 2000). For more detailed information, see Ungar and Williamson (2000).
63
Figure 5. Generalized scaphitid morphology. Illustration of Scaphites whitfieldi showing the adult body chamber shaft and juvenile planispiral phragmocone (modified from Landman et al., 2012).
64
Figure 6. Ammonoid dimorph sizes. Size distribution of Hoploscaphites spedeni showing the bimodal distribution and variation in size (LMAX, mm—see Figure 14) between the macroconch (M) and microconch (m). Note the overlap in sizes between the two forms (modified from Landman & Waage, 1993).
65
Figure 7. Scaphitid shell features. Lateral schematic view of a scaphite belonging to the genus Hoploscaphites exhibiting its basic body plan: tightly coiled phragmocone, shaft-like body chamber terminating in a recurved hook and various ornamental features including ribs and tubercles (modified from Landman et al., 2012).
66
Figure 8. Dimorphs of Hoploscaphites spedeni. Examples of both macroconch and microconch forms of H. spedeni not shown to scale (LMAX (mm) range: 73.0 – 131.7). Note the broad range of variation in ornamentation and shape in this taxon. Black arrow indicates base of body chamber. A-B, macroconch (YPM 27156). A, right lateral; B, apertural. C-D, macroconch (YPM 23122). C, posteroventral; D, left lateral. E-F, microconch (YPM 23699). E, right lateral; F, posteroventral. G-H, microconch (YPM 23718). G, right lateral; H, apertural. (Modified from Landman & Waage, 1993).
67
Figure 9. Dimorphs of Hoploscaphites nebrascensis. Examples of both macroconch and microconch forms of H. nebrascensis not shown to scale (LMAX (mm) range: 71.7 – 150.2). Note the more constrained morphology in this taxon. Black arrow indicates base of body chamber. A-B, macroconch (YPM 23146). A, right lateral; B, apertural. C-D, macroconch (YPM 23147). C, posteroventral; D, left lateral. E-F, microconch (YPM 23697). E, apertural; F, left lateral. G-H, microconch (YPM 23195). G, apertural; H, left lateral. (Modified from Landman & Waage, 1993).
68
Figure 10. Dimorphs of both Hoploscaphites species. Comparison of H. spedeni and H. nebrascensis macroconchs and microconchs. Black arrow indicates base of body chamber. A-B, H. spedeni macroconch (YPM 23124). A, apertural; B, left lateral. C-D, H. spedeni macroconch (USNM 468854). C, apertural; D, left lateral. E-F, H. spedeni microconch (YPM 23710). E, right lateral; F, apertural. G-H, H. spedeni microconch (YPM 23718). G, right lateral; H, apertural. I-J, H. nebrascensis macroconch (YPM 23145). I, right lateral; J, apertural. K-L, H. nebrascensis macroconch (YPM 23151). K, apertural; L, left lateral. M-N, H. nebrascensis microconch (YPM 23687). M, right lateral; N, apertural. O-P, H. nebrascensis microconch (YPM 23735). O, left lateral; P, apertural. (Modified from Landman & Waage, 1993).
69
Figure 11. Intermediate morphology. Comparison of a rotund H. spedeni macroconch (A and B) (modified from Landman & Waage, 1993), to an intermediate H. spedeni macroconch from the very top of the H. nicolletii Range Zone (C and D) (modified from Landman et al., 2013), and to a typical H. nebrascensis macroconch (E and F) (modified from Landman & Waage, 1993). Black arrow indicates base of body chamber. A-B, H. spedeni macroconch (YPM 23122). A, posteroventral; B, left lateral. C-D, H. spedeni macroconch (AMNH 64406). C, posteroventral; D, left lateral. E-F, H. nebrascensis macroconch (YPM 23147). E, left lateral; F, posteroventral.
70
Figure 12. Biostratigraphy of the Hoploscaphites taxa. The Hoploscaphites nicolletii Range Zone spans the Upper Elk Butte Member of the Pierre Shale and the Trail City Member of the Fox Hills Formation in South Dakota. Assemblage zones within the H. nicolletii Range Zone containing fossiliferous concretionary layers are listed as Lower nicolletii A.Z. (LNAZ), Limopsis-Gervillia A.Z. (LGAZ), Upper nicolletii A.Z. (UNAZ) and Protocardia-Oxytoma A.Z. (POAZ). H. spedeni is restricted to this range zone. H. nicolletii Range Zone terminates at the transition concretions between the Trail City Member and the Timber Lake Member of the Fox Hills Formation. At the transition concretions begins the H. nebrascensis Range Zone spanning the Timber Lake Member to the Iron Lightning Member (not shown) of the Fox Hills Formation. (Modified from Landman & Waage, 1993).
71
Figure 13. Materials. Materials used to pedestal ammonoid specimens for photography and photogrammetric 3D modeling. Note the black hash-lines around the circumference of the nine inch cake pan; these hash-lines were placed every 10° around the cake pan to ensure full, overlapping coverage of the specimen in the photographs for proper 3D modeling. A rotund macroconch of Hoploscaphites spedeni, YPM 23122, is shown here.
72
Figure 14. Modified Raup parameters for scaphitids. A schematic of standard measurements for adult scaphitic ammonoids, specifically the genus Hoploscaphites. These measurements are based on Raup’s (1966, 1967) parameters but modified to reflect the heteromorphic body plan. The bold arrow indicates the base of the body chamber. Abbreviations: LMAX = maximum length (shell diameter), WUS = whorl width at the ultimate septum, HUS = whorl height at the ultimate septum, WAPT = whorl width at the aperture, HAPT = whorl height at the aperture, UD = umbilical diameter, A = apertural angle. (Modified from Landman & Waage, 1993.)
73
Figure 15. Photogrammetric accuracy. Photograph of Hoploscaphites spedeni (YPM 23118) macroconch (Top) and its respective high resolution digital model (Bottom) derived from 179 photographs and constructed by 123D Catch (Autodesk, Inc., 2013) and displayed in MeshLab (Cignoni et al., 2011). Note the background scatter around the digital model from the nine inch cake pan and pillow being partially reconstructed along with the ammonoid specimen.
74
Figure 16. Edited digital model example. High resolution digital model of Hoploscaphites spedeni (YPM 23118) macroconch after being edited within MeshLab (Cignoni et al., 2011). Background scatter and color were removed, and the ammonoid digital model was cut along the dorso-ventral plane.
75
Figure 17. TIN accuracy. Photograph of Hoploscaphites spedeni (YPM 23122) macroconch (Top) and its respective triangular irregular network (TIN) model displayed in ArcGIS 10 (ESRI, Inc., 2011) (Bottom).
76
Figure 18. TIN model exhibiting aspect. Aspect (slope direction) on Hoploscaphites TIN models used in the Orientation Patch Count (OPC) analysis. A color wheel is provided as a legend for aspect orientation. A, H. spedeni macroconch (YPM 23122); B, H. spedeni microconch (YPM 23199); C, H. nebrascensis macroconch (YPM 23144); D, H. nebrascensis microconch (YPM 23195).
77
Figure 19. 3D:2D area ratio rasters. Examples of surface-to-planimetric area ratio index rasters of Hoploscaphites. Light purple cells cells are near 1 (i.e., near equal surface area and planimetric area); red cells are furthest from 1 (i.e., very steep slopes resulting in a high surface area to planimetric area ratio). A, H. spedeni macroconch (YPM 23122); B, H. spedeni microconch (YPM 23199); C, H. nebrascensis macroconch (YPM 23144); D, H. nebrascensis microconch (YPM 23195).
78
Figure 20. Vectorization of ammonoid features. Vector data of Hoploscaphites features digitized by hand. Tubercle centers (or tubercle bases for ventrolateral tubercles) were digitized as points, ribs were digitized as polylines, and body chambers (Green) and phragmocones (Blue) were digitized separately as polygons. Rib polylines were digitized only for H. spedeni specimens since the highly tuberculate morphology of H. nebrascensis specimens made digitizing rib polylines potentially erroneous. A, H. spedeni macroconch (YPM 27164); B, H. spedeni microconch (YPM 23730); C, H. nebrascensis macroconch (YPM 23144); D, H. nebrascensis microconch (YPM 23687).
79
Figure 21.Transect and buffer vectors. Additional vector data collected on Hoploscaphites specimens displayed in relation to rib polylines and tubercle points. The 25% transect (Green), the 50% transect (Red) and the 75% transect (Blue) were digitized by hand for H. spedeni dimorphs. The 1 mm buffers (Yellow) for tubercle points were generated with the Buffer Tool in ArcGIS 10 (ESRI, Inc., 2011) and were created for both H. spedeni and H. nebrascensis specimens. The additional 2 mm buffers that were created with the Buffer Tool are not shown (see Figure 24). This macroconch specimen belongs to H. spedeni (YPM 27160).
80
Figure 22. Dispersed, random and clustered features. Examples of dispersed, random and clustered feature classes (i.e., points) (modified from ESRI, Inc., 2011). An Average Nearest Neighbor analysis calculates a Nearest Neighbor Index (NNI), which measures how similar the observed mean distance is to the expected mean distance in the form of a ratio. The null hypothesis states that these features are randomly distributed with the distances of the nearest neighbors having a normal distribution. Features are considered dispersed, or evenly spaced, when the NNI is greater than 1. Features are considered clustered, or in discrete groups, when the NNI is less than 1. An Average Nearest Neighbor analysis was conducted on rib polylines and tubercle points of select Hoploscaphites specimens.
81
Figure 23. Rib density transect and map. Methods used for calculating rib density as shown on a Hoploscaphites spedeni macroconch (YPM 23118). Rib polylines were joined with spiral transects representing 25% (Green), 50% (Red) and 75% (Blue) of whorl height (Left). Rib density map generated from rib polylines to visualize the change in rib density through ontogeny; white areas contained the most dense rib locations on the shell (Right).
82
Figure 24. Tubercle buffers and intersecting ribs. Examples of the 1 mm buffers (Left) and 2 mm buffers (Right) digitized around tubercle points in a Hoploscaphites spedeni microconch (YPM 23704). Note that the 2 mm buffers (Right) intersected ribs 100% of the time, making these buffers impractical to use for determining how tightly tubercles and ribs correlate. The 1 mm buffers were retained for tubercle-rib correlation analysis.
83
Figure 25. Separated body chamber and whole shell polygons. Planimetric area polygon of the whole shell (Yellow) and of only the body chamber (Red) of Hoploscaphites used in the shell coiling analysis. An index for shell coiling was generated by ratioing polygonal area of the body chamber to the whole shell. A relatively large index for the body chamber to whole shell area indicates that much of the phragmocone is not exposed due to the tight coiling of the body chamber around it. Conversely, a relatively small index indicates that much of the phragmocone is exposed due to the loose coiling of the body chamber around the phragmocone. A, Whole shell polygon of H. nebrascensis macroconch (YPM 23147); B, Body chamber polygon of H. nebrascensis macroconch (YPM 23147); C, Whole shell polygon of H. nebrascensis microconch (YPM 23697); D, Body chamber polygon of H. nebrascensis microconch (YPM 23697).
84
Figure 26. Landmark-scaled body chambers. Body chamber morphospace of the Hoploscaphites species (combined) composed of merged polygons scaled to two, biologically meaningful landmark pairs: the dorsal and ventral ends of the ultimate septum (A, macroconchs; B, microconchs), and the dorsal end of the ultimate septum paired with the dorsal lip of the aperture (C, macroconchs; D, microconchs). All polygons merged to create the morphospace are shown for comparison. Note: H. nebrascensis specimens are within the body chamber morphospace of H. spedeni with respect to the two different landmark pairs used.
85
Figure 27. OPC vs. LMAX. Reduced major axis regression of Orientation Patch Count (OPC) versus shell size (LMAX, mm) in both species and dimorphs of Hoploscaphites. Pearson’s r = 0.55, p = 5.0E-05. The biggest shells, which are also the most ornate, tend to have the largest OPC values.
86
Figure 28. OPC results 1. Frequency histograms of OPC of Hoploscaphites spedeni and H. nebrascensis compared separately between macroconchs (Blue) and microconchs (Red). The mean, variance and results of statistical analyses are presented with each histogram. Statistically significant differences are highlighted in yellow and marked with asterisks. 87
Figure 29. OPC results 2. Frequency histograms of OPC of Hoploscaphites spedeni (Blue) and H. nebrascensis (Red) compared together by dimorph. The mean, variance and results of statistical analyses are presented with each histogram. No statistically significant differences are present. 88
Figure 30. Morphotype bins. Frequency histograms of H. spedeni median cell values binned into three groups, which reflect Landman & Waage’s (1993) description of the three morphotypes. Note: Due to a broken shell, one microconch specimen had an unusually large 3D:2D median value; this specimen was placed into a fourth bin, allowing the distributions of the microconchs to reflect their morphotype groups.
89
Table 1. Results of Mann-Whitney U Test comparisons for surface-to-planimetric area ratio medians.
Specimen1 Specimen 1 Median Specimen 2 Specimen 2 Median U * p H. spedeni macro body chamber: typical vs rotund (ornate) YPM23124 1.105 YPM23118 1.101 200.95 < 0.01 typical vs compressed (smooth) YPM23124 1.105 YPM23116 1.056 25.33 < 0.01
H. spedeni macro phragmocone: typical vs rotund (ornate) YPM23124 1.091 YPM23118 1.141 30.81 < 0.01 typical vs compressed (smooth) YPM23124 1.091 YPM23116 1.092 4.91 < 0.01
H. spedeni and H. nebrascensis macro body chamber: rotund H. spedeni vs lowest H. nebrascensis YPM23118 1.101 YPM23145 1.054 44.61 < 0.01 rotund H. spedeni vs highest H. nebrascensis YPM23118 1.101 YPM23147 1.089 13.15 < 0.01
H. spedeni and H. nebrascensis macro phragmocone: rotund H. spedeni vs lowest H. nebrascensis YPM23118 1.141 YPM23145 1.070 48.43 < 0.01 rotund H. spedeni vs highest H. nebrascensis YPM23118 1.141 YPM23147 1.109 23.17 < 0.01
H. nebrascensis macro body chamber: lowest vs highest YPM23145 1.054 YPM23147 1.089 204.08 < 0.01
H. nebrascensis macro phragmocone: lowest vs highest YPM23145 1.070 YPM23147 1.109 9.70 < 0.01
H. spedeni micro body chamber: typical vs rotund (ornate) YPM23714 1.086 YPM23702 1.139 32.68 < 0.01 typical vs compressed (smooth) YPM23714 1.086 YPM23199 1.059 28.65 < 0.01
H. spedeni micro phragmocone: typical vs rotund (ornate) YPM23714 1.120 YPM23702 1.129 9.56 < 0.01 typical vs compressed (smooth) YPM23714 1.120 YPM23199 1.092 21.27 < 0.01
H. spedeni and H. nebrascensis micro body chamber: rotund H. spedeni vs lowest H. nebrascensis YPM23702 1.139 YPM23198 1.046 66.41 < 0.01 rotund H. spedeni vs highest H. nebrascensis YPM23702 1.139 YPM23697 1.100 26.30 < 0.01
H. spedeni and H. nebrascensis micro phragmocone: rotund H. spedeni vs lowest H. nebrascensis YPM23702 1.129 YPM23198 1.073 42.12 < 0.01 rotund H. spedeni vs highest H. nebrascensis YPM23702 1.129 YPM23697 1.127 7.15 < 0.01
H. nebrascensis micro body chamber: lowest vs highest YPM23198 1.046 YPM23697 1.100 41.50 < 0.01
H. nebrascensis micro phragmocone: lowest vs highest YPM23198 1.073 YPM23697 1.127 33.16 < 0.01 90
Figure 31. 3D:2D area median vs. OPC. Reduced major axis regression of whole shell 3D:2D Median Value versus Orientation Patch Count (OPC) in both species and dimorphs of Hoploscaphites. Pearson’s r = 0.02, p = 0.87. No correlation is present between these two independent surface rugosity analyses.
91
Figure 32. 3D:2D area median vs. LMAX. Reduced major axis regression of whole shell 3D:2D Median Value versus shell size (LMAX, mm) in both species and dimorphs of Hoploscaphites. Pearson’s r = -0.08, p = 0.59. No correlation is present between these two metrics.
92
Table 2. Results of Average Nearest Neighbor analysis for rib spacing and tubercle spacing. % Specimens Showing N Dispersed Random Clustered Rib spacing
H. spedeni Macroconch 9 0 0 100
H. spedeni Microconch 10 0 0 100
Tubercle spacing
H. spedeni Macroconch 11 27.3 63.6 9.1
H. spedeni Microconch 11 63.6 36.4 0
H. nebrascensis Macroconch 4 75 25 0
H. nebrascensis Microconch 4 50 25 25
93
Figure 33. Rib density spiral transect results. Comparison of the spiral transects located at 25% whorl height (Green), 50% whorl height (Red) and 75% whorl height (Blue) on H. spedeni macroconchs (A) and microconchs (B). The Mann-Whitney U test comparing the two dimorphs at 25%, 50% and 75% transects (U* = 35, p = 0.43; U* = 45, p = 0.97; and U* = 35, p = 0.44, respectively) determined no difference in the medians. The Kolmogorov-Smirnov test comparing the dimorphs at 25% (D = 0.39, p = 0.38), 50% (D = 0.29, p = 0.66) and 75% transects (D = 0.39, p = 0.34) also detected no difference in the distributions.
94
Figure 34. Rib density map results. Rib density maps showing the different rib distributions through ontogeny of H. spedeni. A, macroconch (YPM 23118) exhibiting increased rib density on the dorsal side of the adapertural end of the body chamber; B, macroconch (YPM 23124) exhibiting consistent rib density throughout ontogeny; C, microconch (YPM 23199) exhibiting increased rib density on the ventral side of the adapertural end of the body chamber; D, microconch (YPM 23706) exhibiting relatively consistent rib density throughout ontogeny, though the base of the body chamber appears to have some ribs closely approximated. 95
Figure 35. Shell coiling analysis results. A, H. spedeni macroconchs (Blue) compared to H. spedeni microconchs (Red); B, H. nebrascensis macroconchs (Blue) compared to H. nebrascensis microconchs (Red); C, macroconchs of both H. spedeni (Blue) and H. nebrascensis (Red); D, microconchs of both H. spedeni (Blue) and H. nebrascensis (Red). Statistically significant differences are highlighted in yellow and marked with asterisks.
96
Figure 36. Landmark-scaled body chamber results 1. Body chamber shape distribution analysis with respect to the first set of landmark pairs (dorsal and ventral sides of the ultimate septum). A, H. spedeni macroconchs (Blue) compared to H. spedeni microconchs (Red); B, H. nebrascensis macroconchs (Blue) compared to H. nebrascensis microconchs (Red); C, macroconchs of both H. spedeni (Blue) and H. nebrascensis (Red); D, microconchs of both H. spedeni (Blue) and H. nebrascensis (Red). No statistically significant differences are present.
97
Figure 37. Landmark-scaled body chamber results 2. Body chamber shape distribution analysis with respect to the second set of landmark pairs (dorsal end of ultimate septum and the dorsal lip of the aperture). A, H. spedeni macroconchs (Blue) compared to H. spedeni microconchs (Red); B, H. nebrascensis macroconchs (Blue) compared to H. nebrascensis microconchs (Red); C, macroconchs of both H. spedeni (Blue) and H. nebrascensis (Red); D, microconchs of both H. spedeni (Blue) and H. nebrascensis (Red). No statistically significant differences are present.
98
REFERENCES
Alberch, P., Gould, S.J., Oster, G.F., Wake, D.B., 1979. Size and shape in ontogeny and phylogeny. Paleobiology, 5(3), pp. 296-317.
Autodesk, Inc., 2013. Autodesk 123D Catch. http://www.123dapp.com/catch (accessed May 4, 2013).
Berry, J.K., 2013. Beyond Mapping III: Procedures and applications in GIS modeling. Berry & Associates, Basis Press, Fort Collins, Colorado: http://www.innovativegis.com/basis/MapAnalysis/Default.htm.
Bookstein, F.L., Ward, P.D., 2013. A modified Procrustes analysis for bilaterally symmetrical outlines, with an application to microevolution in Baculites. Paleobiology, 39(2), pp. 214-234.
Bucher, H., Landman, N.H., Klofak, S.M., Guex, J., 1996. Mode and rate of growth in ammonoids. In: Landman, N.H., Tanabe, K., Davis, R.A. (Eds.). Ammonoid Paleobiology, Volume 13: Topics in Geobiology, Plenum Press, New York, New York, pp. 407-461.
Bocxlaer, B.V., Schultheiss, R., 2010. Comparison of morphometric techniques for shapes with few homologous landmarks based on machine-learning approaches to biological discrimination. Paleobiology, 36(3), pp. 497-515.
Calloman, J.H., 1963. Sexual dimorphism in Jurassic ammonites. Transactions of the Leicester Literary and Philosophical Society, 57, pp. 21-56.
Checa, A., Company, M., Sandoval, J., Weitschat, W., 1996. Covariation of morphological characters in the Triassic ammonoid Czekanowskites rieberi. Lethaia, 29(3), pp. 225-235.
Cignoni, P., Ranzuglia, G., Callieri, M., Corsini, M., Ganovelli, F., Pietroni, N., Tarini, M., 2011. MeshLab. Visual Computing Lab, Italian National Research Council.
Clarkson, E.N.K., 1986. Invertebrate Palaeontology and Evolution, 2nd ed. Allen & Unwin Ltd, London, United Kingdom, 382 pp.
Cooley, S.W., 2013. GIS 4 geomorphology: Geomorphology of mountain landscapes & upland watersheds: http://gis4geomorphology.com/ (accessed May 4, 2013).
Dagys, A.S., 2001. The ammonoid family Arctohungaritidae from Boreal Lower-Middle Anisian (Triassic) of Arctic Asia. Revue Paleobiology Geneve, 20, pp. 543-641.
Dagys, A.S., Bucher, H., Weitschat, W., 1999. Intraspecific variation of Parasibirites kolymensis Bychkov (Ammonoidea) from the Lower Triassic (Spathian) of Arctic Asia. Mitteilungen aus dem Geologisch-Paläontologischen der Institut Universität Hamburg, 83, pp. 163- 178. 99
Dagys, A.S., Weitschat, W., 1993. Extensive intraspecific variation in a Triassic ammonoid from Siberia. Lethaia, 26(2), pp. 113-121.
Davis, R.A., Landman, N.H., Dommergues, J.L., Marchand, D., Bucher, H., 1996. Mature modifications and dimorphism in ammonoid cephalopods. In: Landman, N.H., Tanabe, K., Davis, R.A. (Eds.). Ammonoid Paleobiology, Volume 13: Topics in Geobiology, Plenum Press, New York, New York, pp. 463-539.
De Baets, K., Klug, C., Korn, D., Landman, N.H., 2012. Early evolutionary trends in ammonoid embryonic development. Evolution, 66(6), pp. 1788-1806.
De Baets, K., Klug, C., Monet, C., 2013. Intraspecific variability through ontogeny in early ammonoids. Paleobiology, 39(1), pp. 75-94.
Delanoy, G., Ropolo, P., Magnin, A., Autran, G., Poupon, A., Gonnet, R., 1995. Sur le dimorphisme chez les Ancyloceratina (Ammonoidea) de Crétacé inférieur. Comptes Rendus de l’Académie des Sciences de Paris, série 2a, 321, pp. 537-543.
Eronen, J. T., Evans, A. R., Fortelius, M., Jernvall, J., 2009. The impact of regional climate on the evolution of mammals: A case study using fossil horses. Evolution, 64(2), pp. 398- 408.
ESRI, Inc., 2011. ArcGIS Desktop: Release 10. Environmental Systems Research Institute, Redlands, California.
Evans, A.R., Wilson, G.P., Fortelius, M., Jernvall, J., 2007. High-level similarity of dentitions in carnivorans and rodents. Nature, 445, pp. 78-81.
Falkingham, P. L., 2012. Acquisition of high resolution 3D models using free, open-source, photogrammetric software. Palaeontologia Electronica, 15(1), 15 pp.
Furukawa, Y., Ponce, J., 2010. Accurate, dense, and robust multi-view stereopsis. IEEE Transactions on Pattern Analysis and Machine Intelligence, 32(8), pp. 1362-1376.
Gangopadhyay, T.K., Bardhan, S., 2007. Ornamental polymorphism in Placenticeras kaffrarium (Ammonoidea; Upper Cretaceous of India): Evolutionary implications. In: Landman, N.H., Davis, R.A., Mapes, R.H., (Eds). Cephalopods Present and Past: New Insights and Fresh Perspectives. Springer, Netherlands, pp. 97-120.
Gerber, S., Neige, P., Eble, G.J., 2007. Combining ontogenetic and evolutionary scales of morphological disparity: a study of early Jurassic ammonites. Evolution & Development, 9(5), pp. 472-482.
Gould, S.J., 1977. Ontogeny and Phylogeny, The Belknap Press of Harvard University Press. Cambridge, Massachusetts, 501 pp. 100
Guex, J., Koch, A., O'Dogherty, L., Bucher, H., 2003. A morphogenetic explanation of Buckman's law of covariation. Bulletin de la Société Géologique de France, 174(6), pp. 603-606.
Hammer, Ø., Bucher, H., 2005. Buckman's first law of covariation – a case of proportionality. Lethaia, 38(1), pp. 67-72.
Hammer, Ø., Bucher, H., 2006. Generalized ammonoid hydrostatics modelling, with application to Intornites and intraspecific variation in Amaltheus. Paleontological Research, 10(1), pp. 91-96.
Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: PAleontological STatistics software package for education and data analysis. Palaeontologia Electronica, 4(1), 9 pp.: http://palaeo-electronica.org/2001_1/past/issue1_01.htm.
Hammer, Ø., Harper, D.A.T., 2006. Paleontological Data Analysis, Blackwell Publishing Ltd, Oxford, United Kingdom, 351 pp.
Harada, K., Tanabe, K., 2005. Paedomorphosis in Turonian (Late Cretaceous) collignoniceratine ammonite lineage from the north Pacific region. Lethaia, 38(1), pp. 47-57.
Hariri, K.E., Bachnou, A., 2004. Describing ammonite shape using Fourier analysis. Journal of African Earth Sciences, 39, pp. 347-352.
Hohenegger, J., Tatzreiter, F., 1992. Morphometric methods in determination on ammonite species, exemplified through Balatonites shells (Middle Triassic). Journal of Paleontology, 66(5), pp. 801-816.
Howarth, M.K., 1978. The stratigraphy and ammonite fauna of the Upper Lias of Northamptonshire. Bulletin of the British Museum (Natural History), 29(3), pp. 235-288.
Jacobs, D.K., Chamberlain, Jr., J.A., 1996. Buoyancy and hydrodynamics in ammonoids. In: Landman, N.H., Tanabe, K., Davis, R.A. (Eds.). Ammonoid Paleobiology, Volume 13: Topics in Geobiology, Plenum Press, New York, New York, pp. 169-224.
Jacobs, D.K., Landman, N.H., Chamberlain, Jr., J.A., 1994. Ammonite shell shape covaries with facies and hydrodynamics: Iterative evolution as a response to changes in basinal environment. Geology, 22(10), pp. 905-908.
Kakabadze, M.V., 2004. Intraspecific and intrageneric variabilities and their implications for the systematics of Cretaceous heteromorph ammonites; a review. Scripta Geologica, 128, 17- 37.
Kennedy, W.J., Cobban, W.A., 1976. Aspects of ammonite biology, biogeography, and biostratigraphy. Special Papers in Palaeontology, 17, pp. 1-94. 101
Knauss, M.J., 2012. Quantifying morphological variability in ammonoids using GIS spatial analyses. Geological Society of America Abstracts with Programs, 44(7), 442 pp.
Korn, D., 2012. Quantification of ontogenetic allometry in ammonoids. Evolution and Development, 14(6), pp. 501-514.
Landman, N.H., 1987. Ontogeny of Upper Cretaceous (Turonian-Santonian) scaphitid ammonites from the Western Interior of North America: Systematics, developmental patterns, and life history. Bulletin of the American Museum of Natural History, 185(2), pp. 117-241.
Landman, N.H., 1989. Iterative progenesis in Upper Cretaceous ammonites. Paleobiology, 15(2), pp. 95-117.
Landman, N.H., Cobban, W.A., Larson, N.L., 2012. Mode of life and habitat of scaphitid ammonites. Geobios, 559, pp. 87-98.
Landman, N.H., Dommergues, J.L., Marchand, D., 1991. The complex nature of progenetic species—examples from Mesozoic ammonites. Lethaia, 24(4), pp. 409-421.
Landman, N.H., Geyssant, J.R., 1993. Heterochrony and ecology in Jurassic and Cretaceous ammonites. Geobios, M.S., 15, pp. 247-255.
Landman, N. H., Kennedy, W. J., Cobban, W. A., Larson, N. L., 2010. Scaphites of the “Nodosus Group” from the Upper Cretaceous (Campanian) of the Western Interior of North America, Bulletin of the American Museum of Natural History, 342, 242 pp.
Landman, N.H., Klofak, S.M., Sarg, K.B., 2008. Variations in adult size of scaphitid ammonites from the Upper Cretaceous Pierre Shale and Fox Hills Formation. In: Harries, P.J. (Eds.). High-Resolution Approaches in Stratigraphic Paleontology, Kluwer Academic Press, Boston, Massachusetts, pp. 149-194.
Landman, N.H., Remin, Z., Garb, M.P., Chamberlain Jr., J.A., 2013. Cephalopods from the Badlands National Park area, South Dakota: Reassessment of the position of the Cretaceous/Paleogene boundary. Cretaceous Research, 42, pp. 1-27.
Landman, N.H., Waage, K.M., 1993. Scaphitid ammonites of the Upper Cretaceous (Maastrichtian) Fox Hills Formation in South Dakota and Wyoming. Bulletin of the American Museum of Natural History, 215, 257 pp.
Makowski, H., 1962. Problem of sexual dimorphism in ammonites. Acta Palaeontologia Polonica, 12, pp. 1-92.
Manship, L.L., 2004. Pattern matching: Classification of ammonitic sutures using GIS. Palaeontologia Electronica, 7(2), 15 pp. 102
The MathWorks, Inc., 2007a. MATLAB and Statistics Toolbox. Natick, Massachusetts.
Mitchell, A., 2005. The ESRI Guide to GIS Analysis, Volume 2: Spatial Measurements & Statistics, ESRI Press, Redlands, California, pp. 71-146.
Monnet, C., Bucher, H., 2005. New Middle and Late Anisian (Middle Triassic) ammonoid faunas from northwestern Nevada (USA): taxonomy and biochronology. Fossils and Strata, 52, pp. 1-121.
Monnet, C., Bucher, H., Wasmer, W., Guex, J., 2010. Revision of the genus Acrochordiceras Hyatt, 1877 (Ammonoidea, Middle Triassic): morphology, biometry, biostratigraphy and intraspecific variability. Palaeontology, 53(5), pp. 961-996.
Morard, A., Guex, J., 2003. Ontogeny and covariation in the Toarcian genus Osperlioceras (Ammonoidea). Bulletin de la Société Géologique de France, 174, pp. 607-615.
Neige, P., 1999. The use of landmarks to describe ammonite shape: Examples from the Harpoceratinae. In: Oloriz, F. & Rodriguez-Tovar, F.J., (Eds). Advancing Research on Living and Fossil Cephalopods, Kluwer Academic/Plenum Publishers, New York, New York, pp. 263-272.
Nishimura, T., Maeda, H., Tanaka, G., Ohno, T., 2010. Taxonomic evaluation of various morphological characters in the Late Cretaceous desmoceratine polyphyletic genus “Damesites” from the Yezo Group in Hokkaido and Sakhalin. Paleontological Research, 14(1), pp. 33-55.
Parent, H., 1997. Ontogeny and sexual dimorphism of Eurycephalites gottschei (Tornquist) (Ammonoidea) of the Andean Lower Callovian (Argentine-Chile). Geobios, 30, pp. 407- 419.
Price, M., 2012. Mastering ArcGIS, Fifth Edition. McGraw-Hill, New York, New York, pp. 295- 326.
Rashid, H., 2010. 3-D Surface-area computation of the state of Jammu & Kashmir using Shuttle Radar Topographic Mission (SRTM) data in Geographical Information System (GIS). Journal of Geomatics, 4(2), pp. 77-82.
Raup, D.M., 1966. Geometric analysis of shell coiling: general problems. Journal of Paleontology, 40(5), pp. 1178-1190.
Raup, D.M., 1967. Geometric analysis of shell coiling: coiling in ammonoids. Journal of Paleontology, 41(1), pp. 43-65.
Reeside Jr., J.B., Cobban, W.A., 1960. Studies of the Mowry Shale (Cretaceous) and Contempory Formations in the United States and Canada. Geological Survey 103
Professional Paper 335, Washington D.C., 126 pp.
Salazar-Ciudad, I., Marin-Riera, M., 2013. Adaptive dynamics under development-based genotype-phenotype maps. Nature, 497, pp. 361-365.
Sheffield, S.L., Zachos, L.G., Lewis, R.D., 2012. A morphometric study of Erisocrinus (Crinoidea) using ArcGIS. Geological Society of America Abstracts with Programs, 44(7), 232 pp.
Speden, I. G., 1970. The type Fox Hills Formation, Cretaceous (Maastrichtian), South Dakota: Part 2, Systematics of the Bivalvia. Peabody Museum of Natural History Bulletin, 33, 222 pp.
Stone, J.R., 1998. Landmark-based thin-plate spline relative warp analysis on gastropod shells. Systematic Biology, 47(2), pp. 254-263.
Ubukata, T., Tanabe, K., Shigeta, Y., Maeda, H., Mapes, R.H., 2008. Piggyback whorls: A new theoretical morphologic model reveals constructional linkages among morphological characters in ammonoids. Acta Palaeontologica Polonica, 53(1), pp. 113-128.
Ungar P., Williamson, M., 2000. Exploring the effects of toothwear on functional morphology: A preliminary study using dental topographic analysis. Palaeontologia Electronica, 3(1), 18 pp.
Waage, K. M., 1964. Origin of repeated fossiliferous concretion layers in the Fox Hills Formation. Kansas Geological Survey Bulletin, 169, pp. 541-563.
Waage, K. M., 1968. The type Fox Hills Formation, Cretaceous (Maestrichtian), South Dakota. Part I, stratigraphy and paleoenvironments. Peabody Museum of Natural History Bulletin, 27, 175 pp.
Waggoner, K.J., Manship, L.L., 2004. Sutural variation in ammonite ontogeny: applying GIS for paleontologic analyses. Geological Society of America Abstracts with Programs, 36(5), 442 pp.
Weitschat, W., 2008. Intraspecific variation of Svalbardiceras spitzbergensis (Frebold) from Early Triassic (Spathian) of Spitsbergen. Polar Research, 26, pp. 292-297.
Westermann, G.E.G., 1964. Sexual-Dimorphism bei Ammonoideen und seine Bedeutung für Taxonomie der Otoitidae (einschliesslich Sphaeroceratinae; Ammonitina, M. Jura). Palaeontographica A, 124, pp. 1-3, 33-73.
Westermann, G.E.G., 1966. Covariation and taxonomy of the Jurassic ammonite Sonninia adicra (Waagen). Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen, 124(3), pp. 289-312.
104
Wiedmann, J., 1969. The heteromorphs and ammonoid extinction. Biological Reviews, 44, pp. 563-602.
Wilson, G.P., Evans, A.R., Corfe, I.J., Smits, P.D., Fortelius, M., and Jernvall, J., 2012. Adaptive radiation of multituberculate mammals before the extinction of dinosaurs. Nature, 483, pp. 457-460.
Wu, C., 2011. VisualSFM-A visual structure from motion system. Version 0.5. http://www.cs.washington.edu/homes/ccwu/vsfm/ (accessed May 4, 2013).
Yacobucci, M.M., 1999. Plasticity of developmental timing as the underlying cause of high speciation rates in ammonoids: An example from the Cenomanian Western Interior Seaway of North America. In Oloriz, F. & Rodriguez-Tovar, F.J. (Eds). Advancing Research in Living and Fossil Cephalopods, Proceedings, IV International Symposium Cephalopods – Present and Past. Plenum Press, New York, New York, pp. 59-76.
Yacobucci, M.M., 2004. Buckman's Paradox: variability and constraints on ammonoid ornament and shell shape. Lethaia, 37(1), pp. 57-69.
Yacobucci, M.M., 2012. Meta-analysis of character utility and phylogenetic information content in cladistic studies of ammonoids. Geobios, 45, pp. 139-143.
Yacobucci, M.M., Manship, L.L., 2011. Ammonoid septal formation and suture asymmetry explored with a geographic information systems approach. Palaeontologia Electronica, 14(1), 17 pp.
Zachos, L.G., 2012. Morphometric analysis of fossils using heads-up digitizing and geographic information system (GIS) software. Geological Society of America Abstracts with Programs, 44(4), 18 pp.
105
APPENDIX A. RAW DATA TABLES
Table A1. Orientation patch count (OPC) for whole shells, body chambers and phragmocones of both Hoploscaphites species.
Specimen # Species Dimorph Lateral Surface OPC (Whole Shell) OPC (Body Chamber) OPC (Phragmocone) Shell Size (LMAX, mm) YPM-23110 H. spedeni macro right 41537 24219 18294 109 YPM-23116 H. spedeni macro right 42746 33350 10843 123 YPM-23117 H. spedeni macro right 53869 42558 12312 102 YPM-23118 H. spedeni macro left 72865 59829 13953 113 YPM-23120 H. spedeni macro left 52438 36310 17423 95 YPM-23122 H. spedeni macro left 50917 36148 15677 123 YPM-23124 H. spedeni macro left 31364 21425 10981 101 YPM-23129 H. spedeni macro left 36211 24374 12721 84 YPM-23132 H. spedeni macro right 59558 36829 23722 85 YPM-23138 H. spedeni macro right 56694 46083 11447 86 YPM-23144 H. nebrascensis macro left 60174 44387 18197 118 YPM-23145 H. nebrascensis macro left 48581 34505 14942 80 YPM-23146 H. nebrascensis macro right 42855 36233 8197 92 YPM-23147 H. nebrascensis macro left 57759 43218 15551 133 YPM-23151 H. nebrascensis macro left 58668 38115 21716 122 YPM-23195 H. nebrascensis micro left 36429 27877 9298 71 YPM-23198 H. nebrascensis micro left 36406 27022 10091 69 YPM-23199 H. spedeni micro left 27722 19067 9421 69 YPM-23200 H. spedeni micro left 19148 13453 6338 63 YPM-23687 H. nebrascensis micro left 66201 46246 21051 102 YPM-23694 H. spedeni micro left 52890 37383 16200 70 YPM-23697 H. nebrascensis micro left 48796 32570 17176 100 YPM-23699 H. spedeni micro right 40568 24820 16381 95 YPM-23700 H. spedeni micro right 37306 28077 9922 67 YPM-23702 H. spedeni micro right 39512 27020 13128 76 YPM-23704 H. spedeni micro left 38758 20272 19180 90 YPM-23706 H. spedeni micro left 32749 21341 12275 78 YPM-23710 H. spedeni micro right 44736 31445 14079 82 YPM-23713 H. spedeni micro right 38834 28711 10776 62 YPM-23714 H. spedeni micro left 30250 19275 11559 85 YPM-23715 H. spedeni micro right 48951 39310 10348 78 YPM-23719 H. spedeni micro left 38687 23828 15646 83 YPM-23723 H. spedeni micro right 39706 26600 13936 82 YPM-23724 H. spedeni micro right 52849 35176 18453 85 YPM-23726 H. spedeni micro right 49440 36952 13470 84 YPM-23727 H. spedeni micro left 52740 33787 19699 86 YPM-23730 H. spedeni micro left 32725 19559 13761 71 YPM-23732 H. spedeni micro left 21752 15146 7104 70 YPM-23735 H. nebrascensis micro right 45675 36343 10190 76 YPM-23776 H. spedeni macro right 37817 29049 9776 90 YPM-23779 H. spedeni macro right 34798 21309 14817 111 YPM-23783 H. spedeni macro right 59169 37157 22857 108 YPM-24817 H. spedeni macro left 23691 13033 11364 75 YPM-24818 H. spedeni macro left 41271 23214 19137 94 YPM-27160 H. spedeni macro left 41375 26293 15955 106 YPM-27161 H. spedeni macro left 45365 30475 15916 81 YPM-27164 H. spedeni macro left 60659 45804 15761 130 YPM-44402 H. spedeni macro left 35678 23122 13511 105 YPM-44569 H. spedeni micro right 41208 30713 11084 64
106
Table A2. Whole shell surface-to-planimetric (3D:2D) area ratio mean, standard deviation (STD) and median cell values.
Specimen # Species Dimorph Lateral Surface Mean 3D:2D Area Ratio STD 3D:2D Area Ratio Median 3D:2D Area Ratio Cell Number YPM-23110 H. spedeni macro right 1.241 0.763 1.067 37568 YPM-23116 H. spedeni macro right 1.281 1.000 1.065 39584 YPM-23117 H. spedeni macro right 1.257 0.991 1.063 39529 YPM-23118 H. spedeni macro left 1.395 1.310 1.106 36776 YPM-23120 H. spedeni macro left 1.306 0.882 1.067 38791 YPM-23122 H. spedeni macro left 1.311 0.814 1.100 38902 YPM-23124 H. spedeni macro left 1.309 0.659 1.099 36607 YPM-23129 H. spedeni macro left 1.332 1.365 1.068 38652 YPM-23132 H. spedeni macro right 1.264 0.973 1.050 40074 YPM-23138 H. spedeni macro right 1.314 1.350 1.079 41959 YPM-23144 H. nebrascensis macro left 1.227 0.617 1.086 36287 YPM-23145 H. nebrascensis macro left 1.303 1.321 1.055 38217 YPM-23146 H. nebrascensis macro right 1.231 0.698 1.085 38451 YPM-23147 H. nebrascensis macro left 1.222 0.652 1.091 39303 YPM-23151 H. nebrascensis macro left 1.241 0.893 1.062 37467 YPM-23195 H. nebrascensis micro left 1.489 2.236 1.067 38708 YPM-23198 H. nebrascensis micro left 1.332 1.320 1.050 38158 YPM-23199 H. spedeni micro left 1.203 0.527 1.067 36965 YPM-23200 H. spedeni micro left 1.499 2.087 1.098 37648 YPM-23687 H. nebrascensis micro left 1.312 1.016 1.089 37810 YPM-23694 H. spedeni micro left 1.432 1.603 1.088 35834 YPM-23697 H. nebrascensis micro left 1.317 1.111 1.105 37565 YPM-23699 H. spedeni micro right 1.288 0.718 1.088 36570 YPM-23700 H. spedeni micro right 1.514 1.579 1.119 38728 YPM-23702 H. spedeni micro right 1.541 1.864 1.129 36650 YPM-23704 H. spedeni micro left 1.276 0.636 1.118 35844 YPM-23706 H. spedeni micro left 1.225 0.542 1.084 36257 YPM-23710 H. spedeni micro right 1.347 0.850 1.091 36876 YPM-23713 H. spedeni micro right 1.492 2.209 1.087 36705 YPM-23714 H. spedeni micro left 1.323 0.903 1.094 37731 YPM-23715 H. spedeni micro right 1.451 1.540 1.120 38097 YPM-23719 H. spedeni micro left 1.396 1.302 1.104 37695 YPM-23723 H. spedeni micro right 1.368 1.398 1.071 36188 YPM-23724 H. spedeni micro right 1.456 1.265 1.169 35055 YPM-23726 H. spedeni micro right 1.413 1.199 1.116 37146 YPM-23727 H. spedeni micro left 1.372 1.286 1.088 36534 YPM-23730 H. spedeni micro left 1.477 1.703 1.109 37431 YPM-23732 H. spedeni micro left 1.245 0.751 1.070 37081 YPM-23735 H. nebrascensis micro right 1.392 1.648 1.078 37453 YPM-23776 H. spedeni macro right 1.266 0.963 1.075 37268 YPM-23779 H. spedeni macro right 1.326 1.111 1.106 38749 YPM-23783 H. spedeni macro right 1.418 1.250 1.124 37806 YPM-24817 H. spedeni macro left 1.294 0.659 1.086 38743 YPM-24818 H. spedeni macro left 1.258 1.069 1.049 37583 YPM-27160 H. spedeni macro left 1.391 0.879 1.104 38566 YPM-27161 H. spedeni macro left 1.348 1.133 1.083 40410 YPM-27164 H. spedeni macro left 1.242 0.613 1.060 38533 YPM-44402 H. spedeni macro left 1.277 1.139 1.090 34126 YPM-44569 H. spedeni micro right 1.524 2.279 1.108 36510
107
Table A3. Body Chamber surface-to-planimetric (3D:2D) area ratio mean, standard deviation (STD) and median cell values.
Specimen # Species Dimorph Lateral Surface Mean 3D:2D Area Ratio STD 3D:2D Area Ratio Median 3D:2D Area Ratio Cell Number YPM-23110 H. spedeni macro right 1.247 0.890 1.061 26017 YPM-23116 H. spedeni macro right 1.289 1.174 1.056 28503 YPM-23117 H. spedeni macro right 1.259 1.003 1.062 29207 YPM-23118 H. spedeni macro left 1.373 1.216 1.101 27471 YPM-23120 H. spedeni macro left 1.363 1.174 1.070 26506 YPM-23122 H. spedeni macro left 1.345 0.980 1.099 28227 YPM-23124 H. spedeni macro left 1.325 0.713 1.105 26771 YPM-23129 H. spedeni macro left 1.373 1.732 1.066 29819 YPM-23132 H. spedeni macro right 1.324 1.277 1.055 27177 YPM-23138 H. spedeni macro right 1.346 1.528 1.079 33011 YPM-23144 H. nebrascensis macro left 1.271 0.762 1.093 26663 YPM-23145 H. nebrascensis macro left 1.340 1.596 1.054 27509 YPM-23146 H. nebrascensis macro right 1.257 0.874 1.084 29304 YPM-23147 H. nebrascensis macro left 1.248 0.882 1.089 28027 YPM-23151 H. nebrascensis macro left 1.219 0.846 1.059 26643 YPM-23195 H. nebrascensis micro left 1.577 2.596 1.063 27769 YPM-23198 H. nebrascensis micro left 1.380 1.579 1.046 27909 YPM-23199 H. spedeni micro left 1.197 0.614 1.059 29181 YPM-23200 H. spedeni micro left 1.639 2.695 1.106 27248 YPM-23687 H. nebrascensis micro left 1.309 0.944 1.089 26772 YPM-23694 H. spedeni micro left 1.489 2.029 1.076 25033 YPM-23697 H. nebrascensis micro left 1.311 1.033 1.100 25602 YPM-23699 H. spedeni micro right 1.311 0.829 1.090 27247 YPM-23700 H. spedeni micro right 1.578 1.691 1.126 27789 YPM-23702 H. spedeni micro right 1.717 2.546 1.139 28717 YPM-23704 H. spedeni micro left 1.279 0.777 1.112 26889 YPM-23706 H. spedeni micro left 1.214 0.480 1.086 25237 YPM-23710 H. spedeni micro right 1.402 1.058 1.091 26283 YPM-23713 H. spedeni micro right 1.551 2.508 1.084 25889 YPM-23714 H. spedeni micro left 1.378 1.177 1.086 27379 YPM-23715 H. spedeni micro right 1.478 1.485 1.128 27753 YPM-23719 H. spedeni micro left 1.407 1.288 1.109 25899 YPM-23723 H. spedeni micro right 1.429 1.637 1.075 25739 YPM-23724 H. spedeni micro right 1.525 1.690 1.174 26394 YPM-23726 H. spedeni micro right 1.455 1.405 1.114 25169 YPM-23727 H. spedeni micro left 1.444 1.626 1.090 24387 YPM-23730 H. spedeni micro left 1.651 2.358 1.117 25063 YPM-23732 H. spedeni micro left 1.262 0.775 1.080 27646 YPM-23735 H. nebrascensis micro right 1.426 1.855 1.071 29258 YPM-23776 H. spedeni macro right 1.267 1.023 1.072 27022 YPM-23779 H. spedeni macro right 1.326 1.196 1.099 28357 YPM-23783 H. spedeni macro right 1.368 1.129 1.117 26269 YPM-24817 H. spedeni macro left 1.344 0.821 1.096 29994 YPM-24818 H. spedeni macro left 1.268 1.144 1.049 30091 YPM-27160 H. spedeni macro left 1.436 1.090 1.105 26314 YPM-27161 H. spedeni macro left 1.450 1.478 1.103 28922 YPM-27164 H. spedeni macro left 1.248 0.692 1.050 28246 YPM-44402 H. spedeni macro left 1.266 0.965 1.099 23436 YPM-44569 H. spedeni micro right 1.578 2.655 1.100 26877
108
Table A4. Phragmocone surface-to-planimetric (3D:2D) area ratio mean, standard deviation (STD) and median cell values.
Specimen # Species Dimorph Lateral Surface Mean 3D:2D Area Ratio STD 3D:2D Area Ratio Median 3D:2D Area Ratio Cell Number YPM-23110 H. spedeni macro right 1.403 1.458 1.083 33575 YPM-23116 H. spedeni macro right 1.439 1.526 1.092 27383 YPM-23117 H. spedeni macro right 1.457 2.088 1.073 26259 YPM-23118 H. spedeni macro left 1.797 3.220 1.141 24708 YPM-23120 H. spedeni macro left 1.425 1.577 1.071 30316 YPM-23122 H. spedeni macro left 1.404 1.311 1.113 27339 YPM-23124 H. spedeni macro left 1.348 0.959 1.091 26778 YPM-23129 H. spedeni macro left 1.052 0.069 1.027 30470 YPM-23132 H. spedeni macro right 1.346 1.405 1.047 31835 YPM-23138 H. spedeni macro right 1.663 3.060 1.094 25766 YPM-23144 H. nebrascensis macro left 1.240 0.829 1.084 24643 YPM-23145 H. nebrascensis macro left 1.525 2.586 1.070 25641 YPM-23146 H. nebrascensis macro right 1.402 1.598 1.100 27195 YPM-23147 H. nebrascensis macro left 1.337 1.174 1.109 25480 YPM-23151 H. nebrascensis macro left 1.446 1.945 1.082 27634 YPM-23195 H. nebrascensis micro left 1.783 3.691 1.089 29162 YPM-23198 H. nebrascensis micro left 1.633 2.782 1.073 27934 YPM-23199 H. spedeni micro left 1.306 0.843 1.092 38690 YPM-23200 H. spedeni micro left 1.851 3.797 1.090 33207 YPM-23687 H. nebrascensis micro left 1.554 2.231 1.104 28188 YPM-23694 H. spedeni micro left 1.744 2.848 1.143 30060 YPM-23697 H. nebrascensis micro left 1.574 2.509 1.127 28685 YPM-23699 H. spedeni micro right 1.348 1.003 1.091 36083 YPM-23700 H. spedeni micro right 1.743 3.089 1.115 32048 YPM-23702 H. spedeni micro right 1.685 2.768 1.129 38152 YPM-23704 H. spedeni micro left 1.366 0.862 1.143 38978 YPM-23706 H. spedeni micro left 1.275 0.728 1.089 30776 YPM-23710 H. spedeni micro right 1.358 0.998 1.102 32248 YPM-23713 H. spedeni micro right 1.851 3.876 1.109 32094 YPM-23714 H. spedeni micro left 1.411 1.219 1.120 35382 YPM-23715 H. spedeni micro right 1.812 3.642 1.108 32716 YPM-23719 H. spedeni micro left 1.620 2.459 1.100 33843 YPM-23723 H. spedeni micro right 1.613 2.797 1.070 34490 YPM-23724 H. spedeni micro right 1.670 2.293 1.170 37245 YPM-23726 H. spedeni micro right 1.529 1.880 1.129 32276 YPM-23727 H. spedeni micro left 1.587 2.431 1.099 32056 YPM-23730 H. spedeni micro left 1.701 2.898 1.109 36194 YPM-23732 H. spedeni micro left 1.452 1.879 1.060 35418 YPM-23735 H. nebrascensis micro right 1.742 3.033 1.140 27969 YPM-23776 H. spedeni macro right 1.491 2.019 1.089 25537 YPM-23779 H. spedeni macro right 1.602 2.567 1.134 32484 YPM-23783 H. spedeni macro right 1.814 2.692 1.155 31076 YPM-24817 H. spedeni macro left 1.346 1.038 1.077 35707 YPM-24818 H. spedeni macro left 1.468 2.170 1.062 33149 YPM-27160 H. spedeni macro left 1.481 1.513 1.111 29041 YPM-27161 H. spedeni macro left 1.499 2.341 1.064 31329 YPM-27164 H. spedeni macro left 1.315 0.842 1.106 27967 YPM-44402 H. spedeni macro left 1.546 2.581 1.086 28716 YPM-44569 H. spedeni micro right 2.084 4.553 1.162 28522
109
Table A5. Results of Average Nearest Neighbor analysis of rib spacing on the left lateral surface of ammonoids.
Specimen # Species Dimorph NN Observed NN Expected NN Index (Ratio) NN Z-Score p-Value Result YPM-23118 H. spedeni macro 0.300 0.381 0.788 -5.043 0.000 clustered YPM-23120 H. spedeni macro 0.236 0.311 0.760 -6.034 0.000 clustered YPM-23122 H. spedeni macro 0.304 0.390 0.778 -5.895 0.000 clustered YPM-23124 H. spedeni macro 0.255 0.318 0.801 -4.891 0.000 clustered YPM-23129 H. spedeni macro 0.211 0.286 0.738 -6.466 0.000 clustered YPM-23199 H. spedeni micro 0.181 0.225 0.806 -4.942 0.000 clustered YPM-23200 H. spedeni micro 0.220 0.256 0.860 -2.502 0.012 clustered YPM-23694 H. spedeni micro 0.208 0.243 0.857 -3.323 0.001 clustered YPM-23704 H. spedeni micro 0.269 0.374 0.718 -5.776 0.000 clustered YPM-23706 H. spedeni micro 0.246 0.294 0.836 -3.718 0.000 clustered YPM-23712 H. spedeni micro 0.219 0.251 0.872 -3.202 0.001 clustered YPM-23714 H. spedeni micro 0.272 0.310 0.877 -2.965 0.003 clustered YPM-23727 H. spedeni micro 0.211 0.270 0.780 -5.961 0.000 clustered YPM-23730 H. spedeni micro 0.186 0.235 0.792 -5.164 0.000 clustered YPM-23732 H. spedeni micro 0.213 0.237 0.900 -2.302 0.021 clustered YPM-27160 H. spedeni macro 0.269 0.324 0.830 -4.687 0.000 clustered YPM-27161 H. spedeni macro 0.179 0.219 0.817 -5.276 0.000 clustered YPM-27164 H. spedeni macro 0.309 0.375 0.825 -5.339 0.000 clustered YPM-44402 H. spedeni macro 0.292 0.345 0.844 -3.859 0.000 clustered
110
Table A6. Results of Average Nearest Neighbor analysis of tubercle spacing on the left lateral surface of ammonoids.
Specimen # Species Dimorph NN Observed NN Expected NN Index (Ratio) NN Z-Score p-Value Result YPM-23118 H. spedeni macro 0.696 0.590 1.179 3.301 0.001 dispersed YPM-23120 H. spedeni macro 0.761 0.742 1.025 0.307 0.759 random YPM-23122 H. spedeni macro 0.587 0.518 1.134 2.919 0.004 dispersed YPM-23124 H. spedeni macro 0.456 0.551 0.828 -2.889 0.004 clustered YPM-23129 H. spedeni macro 0.537 0.598 0.897 -1.346 0.178 random YPM-23144 H. nebrascensis macro 0.362 0.314 1.152 5.362 0.000 dispersed YPM-23145 H. nebrascensis macro 0.317 0.272 1.166 4.602 0.000 dispersed YPM-23147 H. nebrascensis macro 0.496 0.395 1.257 8.053 0.000 dispersed YPM-23151 H. nebrascensis macro 0.398 0.424 0.938 -1.589 0.112 random YPM-23195 H. nebrascensis micro 0.377 0.403 0.936 -0.881 0.378 random YPM-23198 H. nebrascensis micro 0.281 0.352 0.798 -3.121 0.002 clustered YPM-23199 H. spedeni micro 0.407 0.411 0.990 -0.156 0.876 random YPM-23200 H. spedeni micro 0.372 0.285 1.303 6.033 0.000 dispersed YPM-23687 H. nebrascensis micro 0.402 0.326 1.234 6.240 0.000 dispersed YPM-23694 H. spedeni micro 0.414 0.327 1.266 5.011 0.000 dispersed YPM-23697 H. nebrascensis micro 0.392 0.303 1.294 7.599 0.000 dispersed YPM-23704 H. spedeni micro 0.509 0.421 1.210 4.077 0.000 dispersed YPM-23706 H. spedeni micro 0.508 0.414 1.227 4.049 0.000 dispersed YPM-23712 H. spedeni micro 0.661 0.578 1.143 1.472 0.141 random YPM-23714 H. spedeni micro 0.651 0.521 1.248 3.773 0.000 dispersed YPM-23719 H. spedeni micro 0.539 0.477 1.130 2.028 0.043 dispersed YPM-23727 H. spedeni micro 0.598 0.509 1.175 2.725 0.006 dispersed YPM-23730 H. spedeni micro 0.552 0.521 1.061 0.724 0.469 random YPM-23732 H. spedeni micro 0.621 0.523 1.187 1.926 0.054 random YPM-24817 H. spedeni macro 0.473 0.530 0.893 -1.448 0.148 random YPM-24818 H. spedeni macro 0.635 0.743 0.854 -1.694 0.090 random YPM-27160 H. spedeni macro 0.763 0.663 1.150 2.279 0.023 dispersed YPM-27161 H. spedeni macro 0.667 0.644 1.036 0.291 0.771 random YPM-27164 H. spedeni macro 0.555 0.582 0.954 -1.003 0.316 random YPM-44402 H. spedeni macro 0.753 0.734 1.025 0.328 0.743 random
111
Table A7. Rib density calculated by the number of ribs that intersect spiral transects at 25%, 50% and 75% whorl height on the left lateral surface of ammonoids.
25% Transect 50% Transect 75% Transect 25% Transect 50% Transect 75% Transect 25% Transect 50% Transect 75% Transect Specimen Species Dimorph Length (cm) Length (cm) Length (cm) Rib Count Rib Count Rib Count Ribs/Length (cm) Ribs/Length (cm) Ribs/Length (cm) YPM-23118 H. spedeni macro 13.00 20.05 26.94 78 117 140 6.00 5.84 5.20 YPM-23120 H. spedeni macro 10.86 16.41 22.27 62 108 156 5.71 6.58 7.01 YPM-23122 H. spedeni macro 13.48 20.09 27.17 51 86 141 3.78 4.28 5.19 YPM-23124 H. spedeni macro 10.92 16.63 22.44 65 104 147 5.95 6.25 6.55 YPM-23129 H. spedeni macro 9.08 13.84 19.02 66 86 107 7.27 6.21 5.63 YPM-23199 H. spedeni micro 8.58 12.08 15.52 68 103 142 7.93 8.53 9.15 YPM-23200 H. spedeni micro 7.79 11.49 14.73 36 65 70 4.62 5.66 4.75 YPM-23694 H. spedeni micro 9.29 12.86 16.44 51 89 115 5.49 6.92 7.00 YPM-23704 H. spedeni micro 11.98 16.70 21.32 45 70 94 3.75 4.19 4.41 YPM-23706 H. spedeni micro 10.08 14.38 18.67 56 83 112 5.56 5.77 6.00 YPM-23712 H. spedeni micro 9.01 13.34 17.41 57 84 127 6.33 6.30 7.30 YPM-23714 H. spedeni micro 11.27 16.00 20.46 47 91 123 4.17 5.69 6.01 YPM-23727 H. spedeni micro 10.48 15.16 19.84 65 112 163 6.20 7.39 8.21 YPM-23730 H. spedeni micro 9.51 13.40 17.30 62 100 134 6.52 7.46 7.75 YPM-23732 H. spedeni micro 7.67 11.63 15.66 53 84 128 6.91 7.22 8.18 YPM-27160 H. spedeni macro 11.70 18.35 24.93 74 135 158 6.33 7.36 6.34 YPM-27161 H. spedeni macro 7.98 12.54 17.45 112 150 191 14.03 11.96 10.94 YPM-27164 H. spedeni macro 14.69 22.79 30.78 102 157 191 6.94 6.89 6.21 YPM-44402 H. spedeni macro 11.91 18.15 24.19 73 117 132 6.13 6.44 5.46
112
Table A8. Percentage of tubercles with intersecting ribs at 1 mm and 2 mm buffers on the left lateral surface of ammonoids.
Specimen # Species Dimorph Percentage of Tubercles with Intersecting Percentage of Tubercles with Intersecting Ribs (1 mm Buffer) Ribs (2 mm Buffer) YPM-23118 H. spedeni macro 58.06 89.25 YPM-23120 H. spedeni macro 87.50 95.00 YPM-23122 H. spedeni macro 64.34 91.47 YPM-23124 H. spedeni macro 85.71 94.81 YPM-23129 H. spedeni macro 74.47 87.23 YPM-23199 H. spedeni micro 93.85 100.00 YPM-23200 H. spedeni micro 83.33 97.22 YPM-23694 H. spedeni micro 87.63 98.97 YPM-23704 H. spedeni micro 65.05 91.26 YPM-23706 H. spedeni micro 85.06 97.70 YPM-23712 H. spedeni micro 89.66 100.00 YPM-23714 H. spedeni micro 84.13 96.83 YPM-23727 H. spedeni micro 92.42 100.00 YPM-23730 H. spedeni micro 97.44 100.00 YPM-23732 H. spedeni micro 93.10 100.00 YPM-27160 H. spedeni macro 88.89 98.41 YPM-27161 H. spedeni macro 83.33 94.44 YPM-27164 H. spedeni macro 82.31 96.15 YPM-44402 H. spedeni macro 89.13 97.83
113
Table A9. Body chamber to whole shell area ratios used in the shell coiling analysis on the left lateral surface of ammonoids.
Specimen # Species Dimorph Whole Shell (cm^2) Body Chamber (cm^2) Phragmocone (cm^2) Body Chamber:Whole Shell (%) YPM-23118 H. spedeni macro 104.54 81.11 23.43 77.59 YPM-23120 H. spedeni macro 69.60 49.04 20.56 70.46 YPM-23122 H. spedeni macro 109.50 81.40 28.09 74.34 YPM-23124 H. spedeni macro 73.80 55.45 18.35 75.13 YPM-23129 H. spedeni macro 54.25 38.86 15.40 71.62 YPM-23144 H. nebrascensis macro 102.81 75.78 27.03 73.71 YPM-23145 H. nebrascensis macro 48.74 36.02 12.72 73.90 YPM-23147 H. nebrascensis macro 133.15 97.28 35.87 73.06 YPM-23151 H. nebrascensis macro 100.98 73.82 27.15 73.11 YPM-23195 H. nebrascensis micro 38.67 28.19 10.48 72.90 YPM-23198 H. nebrascensis micro 32.48 23.14 9.34 71.24 YPM-23199 H. spedeni micro 33.84 23.64 10.20 69.86 YPM-23200 H. spedeni micro 29.17 19.96 9.21 68.42 YPM-23687 H. nebrascensis micro 76.09 50.56 25.53 66.45 YPM-23694 H. spedeni micro 35.97 24.79 11.18 68.92 YPM-23697 H. nebrascensis micro 67.06 45.61 21.45 68.01 YPM-23704 H. spedeni micro 62.11 39.19 22.92 63.10 YPM-23706 H. spedeni micro 47.11 33.93 13.18 72.02 YPM-23712 H. spedeni micro 42.02 28.86 13.16 68.67 YPM-23714 H. spedeni micro 56.93 38.48 18.45 67.59 YPM-23719 H. spedeni micro 48.96 32.97 15.99 67.34 YPM-23727 H. spedeni micro 55.25 37.69 17.56 68.22 YPM-23730 H. spedeni micro 37.27 24.13 13.14 64.75 YPM-23732 H. spedeni micro 32.48 22.36 10.12 68.85 YPM-24817 H. spedeni macro 44.90 31.85 13.05 70.93 YPM-24818 H. spedeni macro 62.63 43.29 19.34 69.12 YPM-27160 H. spedeni macro 87.40 60.99 26.41 69.78 YPM-27161 H. spedeni macro 46.21 31.84 14.37 68.90 YPM-27164 H. spedeni macro 136.22 98.51 37.71 72.32 YPM-44402 H. spedeni macro 82.35 59.45 22.89 72.20
114
Table A10. Body chamber area data from the first landmark pair used in the body chamber shape distribution analysis of the left lateral surface of ammonoids.
Body Chamber Area:Composite Body Specimen # Species Dimorph Scaled Body Chamber Area (cm^2) Chamber Area Ratio YPM-23118 H. spedeni macro 81.11 0.91 YPM-23120 H. spedeni macro 57.43 0.64 YPM-23122 H. spedeni macro 70.56 0.79 YPM-23124 H. spedeni macro 74.24 0.83 YPM-23129 H. spedeni macro 59.05 0.66 YPM-23144 H. nebrascensis macro 58.49 0.66 YPM-23145 H. nebrascensis macro 55.69 0.62 YPM-23147 H. nebrascensis macro 63.92 0.72 YPM-23151 H. nebrascensis macro 56.76 0.64 YPM-23195 H. nebrascensis micro 28.19 0.86 YPM-23198 H. nebrascensis micro 22.24 0.68 YPM-23199 H. spedeni micro 25.22 0.77 YPM-23200 H. spedeni micro 22.53 0.69 YPM-23687 H. nebrascensis micro 21.98 0.67 YPM-23694 H. spedeni micro 24.84 0.76 YPM-23697 H. nebrascensis micro 20.76 0.64 YPM-23704 H. spedeni micro 20.34 0.62 YPM-23706 H. spedeni micro 29.28 0.90 YPM-23712 H. spedeni micro 19.92 0.61 YPM-23714 H. spedeni micro 23.39 0.72 YPM-23719 H. spedeni micro 23.34 0.72 YPM-23727 H. spedeni micro 20.68 0.63 YPM-23730 H. spedeni micro 22.16 0.68 YPM-23732 H. spedeni micro 20.44 0.63 YPM-24817 H. spedeni macro 52.60 0.59 YPM-24818 H. spedeni macro 49.03 0.55 YPM-27160 H. spedeni macro 56.33 0.63 YPM-27161 H. spedeni macro 49.44 0.55 YPM-27164 H. spedeni macro 55.45 0.62 YPM-44402 H. spedeni macro 59.28 0.66
115
Table A11. Body chamber area data from the second landmark pair used in the body chamber shape distribution analysis of the left lateral surface of ammonoids.
Body Chamber Area:Composite Body Specimen # Species Dimorph Scaled Body Chamber Area (cm^2) Chamber Area Ratio YPM-23118 H. spedeni macro 81.11 0.59 YPM-23120 H. spedeni macro 88.29 0.64 YPM-23122 H. spedeni macro 125.62 0.92 YPM-23124 H. spedeni macro 74.70 0.54 YPM-23129 H. spedeni macro 94.98 0.69 YPM-23144 H. nebrascensis macro 93.90 0.68 YPM-23145 H. nebrascensis macro 98.63 0.72 YPM-23147 H. nebrascensis macro 89.82 0.65 YPM-23151 H. nebrascensis macro 101.02 0.74 YPM-23195 H. nebrascensis micro 28.19 0.85 YPM-23198 H. nebrascensis micro 25.00 0.76 YPM-23199 H. spedeni micro 28.09 0.85 YPM-23200 H. spedeni micro 23.74 0.72 YPM-23687 H. nebrascensis micro 22.59 0.68 YPM-23694 H. spedeni micro 22.48 0.68 YPM-23697 H. nebrascensis micro 22.88 0.69 YPM-23704 H. spedeni micro 20.78 0.63 YPM-23706 H. spedeni micro 26.42 0.80 YPM-23712 H. spedeni micro 28.04 0.85 YPM-23714 H. spedeni micro 26.75 0.81 YPM-23719 H. spedeni micro 22.78 0.69 YPM-23727 H. spedeni micro 24.39 0.74 YPM-23730 H. spedeni micro 22.19 0.67 YPM-23732 H. spedeni micro 30.12 0.91 YPM-24817 H. spedeni macro 105.48 0.77 YPM-24818 H. spedeni macro 102.07 0.74 YPM-27160 H. spedeni macro 88.85 0.65 YPM-27161 H. spedeni macro 103.53 0.75 YPM-27164 H. spedeni macro 104.42 0.76 YPM-44402 H. spedeni macro 76.09 0.55
116
APPENDIX B. ADDITIONAL FIGURES
This appendix contains examples of GIS-based images. Only the 29 specimens used in both 3D and 2D geospatial analyses are included. Specimens retained for only 3D or only 2D
GIS-based analyses are not shown here. 117
Figure B1. Hoploscaphites spedeni macroconch, YPM 23118. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon. 118
Figure B2. Hoploscaphites spedeni macroconch, YPM 23120. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
119
Figure B3. Hoploscaphites spedeni macroconch, YPM 23122. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
120
Figure B4. Hoploscaphites spedeni macroconch, YPM 23124. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
121
Figure B5. Hoploscaphites spedeni macroconch, YPM 23129. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
122
Figure B6. Hoploscaphites nebrascensis macroconch, YPM 23144. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
123
Figure B7. Hoploscaphites nebrascensis macroconch, YPM 23145. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
124
Figure B8. Hoploscaphites nebrascensis macroconch, YPM 23147. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
125
Figure B9. Hoploscaphites nebrascensis macroconch, YPM 23151. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
126
Figure B10. Hoploscaphites nebrascensis microconch, YPM 23195. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
127
Figure B11. Hoploscaphites nebrascensis microconch, YPM 23198. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
128
Figure B12. Hoploscaphites spedeni microconch, YPM 23199. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
129
Figure B13. Hoploscaphites spedeni microconch, YPM 23200. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
130
Figure B14. Hoploscaphites nebrascensis microconch, YPM 23687. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
131
Figure B15. Hoploscaphites spedeni microconch, YPM 23694. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
132
Figure B16. Hoploscaphites nebrascensis microconch, YPM 23697. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
133
Figure B17. Hoploscaphites spedeni microconch, YPM 23704. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
134
Figure B18. Hoploscaphites spedeni microconch, YPM 23706. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
135
Figure B19. Hoploscaphites spedeni microconch, YPM 23714. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
136
Figure B20. Hoploscaphites spedeni microconch, YPM 23719. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
137
Figure B21. Hoploscaphites spedeni microconch, YPM 23727. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
138
Figure B22. Hoploscaphites spedeni microconch, YPM 23730. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
139
Figure B23. Hoploscaphites spedeni microconch, YPM 23732. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
140
Figure B24. Hoploscaphites spedeni macroconch, YPM 24817. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
141
Figure B25. Hoploscaphites spedeni macroconch, YPM 24818. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, tubercle points; E, whole shell polygon; F, body chamber polygon.
142
Figure B26. Hoploscaphites spedeni macroconch, YPM 27160. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
143
Figure B27. Hoploscaphites spedeni macroconch, YPM 27161. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
144
Figure B28. Hoploscaphites spedeni macroconch, YPM 27164. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.
145
Figure B29. Hoploscaphites spedeni macroconch, YPM 44402. A, perpendicular view of the lateral surface; B, TIN model exhibiting aspect; C, raster exhibiting surface-to-planimetric area ratio; D, rib polylines; E, whorl height transects; F, rib density map; G, tubercle points; H, tubercle points with 1 mm buffers; I, tubercle points with 2 mm buffers; J, whole shell polygon; K, body chamber polygon.