xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003442
How plastic is vendobionta morphology? a geometric morphometric study of two groups of pteridinium from the latest neoproterozoic
h [electronic resource] /
by Michael Meyer.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Thesis (M.S.)--University of South Florida, 2010.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
ABSTRACT: The analysis and interpretation of Vendobionta morphology is critical to elucidating a range of issues about their ontogeny and evolution, as well as life habits. These analyses, however, are complicated because these organisms are often morphologically enigmatic and defy ready categorization within modern taxonomic schemes. This study delves into the morphology of one of these problematic groups: Pteridinium. Specimens were investigated from two localities, Namibia and North Carolina, using geometric morphometrics. The landmark data, which was analyzed to compare specimens based on locality, taphonomy, and preservation, were subjected to three statistical tests: Principle Components Analysis, Procrustes shape analyses, and Foote's disparity test. All tests revealed no distinct clustering within or by either group due to any of the variables. All variables plotted within the same 95% confidence ellipses, displaying a lack of statistical support for the distinctness of these groups. Therefore, the most parsimonious reason for the lack of differences observed by these two groups stem from them being part of the same morphological group, a conclusion that places into question the validity of the inclusion of two separate species in the genus Pteridinium.
Advisor: Peter Harries, Ph.D.
t USF Electronic Theses and Dissertations.
How Plastic is Vendobionta Morphology? A Geometric Morphometric Study of Two Gr oups of Pteridinium From the Latest Neoproterozoic by Michael B. Meyer A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Geology College of Arts and Sciences University of South Florida Major Professor: Peter Harries, Ph.D. Gregory Herbert, Ph.D. Jonathan Wynn, Ph.D. Date of Approval: August 13, 2009 Keywords: Ediacaran, Nama, Nami bia, North Carolina, Terrane Copyright 2010, Michael B. Meyer
Acknowledgements I would like to thank all of my com mittee members, especially my advisor Peter Harries for helping me put together this thesis. I would also like to thank Jennifer Sliko for being patient in the making and edit ing of this presentati on and its associated thesis, I couldnÂ’t have done it without you; al so to my parents for always supporting me, whatever my goal. Finally, the following indivi duals have helped me significantly (in no particular order): Gregory Herbert, Jona than Wynn, Dolf Seilacher, Gail Gibson, Steve Teeter, Chris Tacker, Patricia Weaver, Patric ia Vickers-Rich, The ErniÂ’s (of Namibia), Roger Portell, the entire administrative sta ff in the department, and my gregarious Lab Mates.
Table of Contents List of Tables ii List of Figures iii Abstract v Introduction 1 Geology and History of the Ediacaran 2 Geology of the Ediacaran 2 Previous work on Pteridinium 8 Description of Pteridinium morphology 8 Pteridinium Species 9 Pteridinium habitat and life habit 12 Geologic setting and Pteridinium Preservation 14 Namibia 14 North Carolina 16 Geometric Morphometrics 18 Landmarks 19 Methods 21 Image and Data Acquisition 22 Statistical Tests 25 Analysis 27 Results of Statistical Tests 29 Discussion 34 Conclusion 39 References 40 i
List of Tables Table 1 Specimen Information 21 Table 2a Pteridinium Eigenvalues 29 Table 2b Piranha Eigenvalues 29 ii
List of Figures Figure 1a. Composite Neoproterozoic carbon isotope record 3 Figure 1b. Correlation of globa l Neoproterozoic successions 4 Figure 2. Mat ecology during the Precambrian 5 Figure 3a. Cyanobacterial mat types 5 Figure 3b. Ediacaran Â“Death MaskÂ” Preservation 7 Figure 4. Quilted Ediacaran fossils 8 Figure 5. Pteridinium simplex 9 Figure 6a. Two different in ferred growth paths for Pteridinium 10 Figure 6b. Linear regression of SeilacherÂ’s original data 11 Figure 6c. Two leading hypothesized Pteridinium life position 13 Figure 7. Map of Namibia 14 Figure 8. Seilacher Slab 15 Figure 9a. North Carolina Map 16 Figure 9b. Column and Map of Stanly County 17 Figure 10. Traditional Morphometri c measurements on a Trilobite 18 Figure 11. Non-homologous width m easurements on two Arthropods 19 Figure 12. Landmark Types 20 Figure 13. Photo setup 23 Figure 14a. Photo setup, position A 24 iii
Figure 14b. Photo setup, position B 24 Figure 15a. Procrustes deformation vectors for Namibian samples 26 Figure 15b. Procrustes deformation v ectors for North Carolina samples 27 Figure 16. Landmark placement on Pteridinium 27 Figure 17a. General Procrustes centroid of all samples 30 Figure 17b. Mean Procrustes centroids of the two groups 30 Figure 17c. North Carolina landmarks 31 Figure 17d. Namibian landmarks 31 Figure 18a. Locality PCA results by specimen 32 Figure 18b. Preservational PCA results 32 Figure 18c. Taphonomic PCA results 33 Figure 19a. Landmarks from two different species of Piranha 35 Figure 19b. Mean Procrustes centroi ds of two species of Piranha 35 Figure 19c. PCA plot for two di fferent species of Piranha 36 Figure 20a. PCA plot with 95% ellipse for two different localities 36 Figure 20b. PCA plot with 95% ellipse for two different taphonomic factors 37 Figure 20c. PCA plot with 95% ellipse for two different preservational factors 37 Figure 20d. PCA plot with 95% ellipse for two different Piranha species 38 iv
How Plastic is Vendobionta Morphology? A Geometric Morphometric Study of Two Gr oups of Pteridinium From the Latest Neoproterozoic Michael B. Meyer ABSTRACT The analysis and interpretation of Vendobionta morphology is critical to elucidating a range of issues about their ontogeny and evolut ion, as well as life habits. These analyses, however, are complicated because these organisms are often morphologically enigmatic and defy ready categorization within modern taxonomic schemes. This study delves into the mor phology of one of these problematic groups: Pteridinium Specimens were investigated from two localities, Namibia and North Carolina, using geometric morphometrics. Th e landmark data, which was analyzed to compare specimens based on locality, taphonomy, and preservation, were subjected to three statistical tests: Principle Component s Analysis, Procrustes shape analyses, and FooteÂ’s disparity test. All tests revealed no di stinct clustering within or by either group due to any of the variables. All variable s plotted within the same 95% confidence ellipses, displaying a lack of statistical support for the distinctness of these groups. Therefore, the most parsimonious reason for the lack of differences observed by these two groups stem from them being part of the same morphological group, a conclusion that places into question the vali dity of the inclusion of two separate species in the genus Pteridinium v
1 Introduction Ediacaran fossils allow us the opportun ity to investigate how the earliest preserved metazoans lived, interacted, and di versified. Such interpretations are complex as these organisms are often morphologically enigmatic and defy ready categorization within modern taxonomic schemes (Jensen et al, 1998). Therefore, researchers have an incomplete view of their biological affinities. Pteridinium (Family Petalonamae) (Grich, 1930) has been the focus of taxonomic studies due to its unique quilted, triple-bladed morphology (Fedonkin, 1992; Ivantsov and Graz hdankin, 1997; Gehling, 1999; Seilacher and Grazhdankin, 2002). While currently there are only two formally recognized species within the genus, P. simplex and P. carolinaensis (Grich 1930, 1933; Richter 1955; Glaessner 1963, Seilacher, 1999), in the past decade discoveries of new quilted/multibladed Neoproterozoic organisms have initia ted a debate as to what characteristics differentiate Pteridinium from other Ediacaran genera, such as Onegia nenoxa or Swartpuntia The primary features that have been used in past research to differentiate these specimens are: wall size, the mode of attachment between blades, individual quilt size, outline shape, and hypothesized life position (Seilacher and Grazhdankin, 2002; Grazhdankin, 2004). Some authors speculate that Pteridinium as classically defined, only exists in Namibia (Grazhdankin, 2004), whereas others contend that all triple-bladed forms, regardless of location, are morphol ogically identical and simply represent specimens deformed during preservati on or ecophenotypy reflecting variable environmental factors (Dro ser, 2002; McCall, 2006) In this study, material will be examined from two localities, Namibia and North Carolina, as the fossils are thought to repres ent two distinct species and, hence, display pronounced morphological differences between them. The morphology of Pteridinium will be investigated using geometric morphometr ics. This analytical technique allows for the statistical quantification of the differences in various attributes present in these two groups utilizing a process th at allows a fuller understan ding of the nature of the differences and what process(es) may underl ie them, including such factors as ecology, taphonomy, and metamorphism. By using ge ometric morphometric analysis to statistically examine the nature of the mo rphological variability displayed by two groups of Pteridinium morphology unknown in extant taxa, a greater understanding can be reached as to how disparate populations of Pteridinium were related to each other. More so, a study of these two species differe nces, within the larger genus of Pteridinium might help us better understand how Ediacaran organisms diversifie d, and how that differs from later, during the Phanerozoic.
2 Geology and History of the Ediacaran Geology of the Ediacaran The Ediacaran Period spans from the end of the Marinoan glaciation (~635 mya) to the Precambrian-Cambrian boundary, curren tly placed at ~542 mya. The beginning of the Ediacaran is marked by rapid global wa rming following the Marinoan glaciation (Fig. 1a, Halverson et al, 2005). This deglacia tion is thought to be expressed by the synchronous appearance of gl obal Â“capÂ” carbonates (Halve rson et al, 2005, McCall, 2006), which overlie glacial diamictites found below them, and is believed to represent the rapid transformation from an Â‘icehouseÂ’ into a Â‘hothouseÂ’ climate. This dramatic change in climate regime is hypothesized to ha ve been initiated by th e large-scale release of trapped greenhouse gases during the late Cryogenian Period as melting of the Marinoan ice sheets commenced (Knoll et al, 2002). Evidence for the massive extent and the bi otic effects of th e Marinoan glaciation are mainly recorded in secula r trends in the carbon isotopic record. Although there is still debate on the Marinoan glacia tionsÂ’ impact on life, there is little doubt that conditions following the late Cryogenian were more favorab le to early metazoan life as the planet was warming and oxygen was increasing (Zhang et al, 1998; Droser et al, 2002). These trends in 13C show positive excursions prior to and after the glaciations, the interpretation of these positiv e then negative trends found during the latter part of Marinoan glaciation is thought to be evidence that there where actually two different smaller glaciations during Marinoan time, rather than one drawn out one (Halverson et al, 2005). If this is true, the glaciations coul d have had more of a Â“one-two punchÂ” (Hoffman, 2001) effect on life, rath er than a Â“starvingÂ” effect. The first body fossils representative of complex organisms initially appeared at ~610 mya; which is correlated to the greate st positive excursion in the carbon isotope record after the Marinoan glaci ation prior to the Gaskiers glaciation. The early Ediacaran metazoan fossil record is dominated by embr yos and acritarchs (e.g., Knoll et al, 2006), whereas the larger forms, such as Pteridinium are found in the la ter third of the Ediacaran (Knoll et al, 2006), an interval of ~30 Ma (figure 1b) following the Gaskiers glaciation. Ediacaran fossils have long been cat egorized into three major Ediacaran Â‘assemblagesÂ’: the Avalon, the White Sea/Ed iacaran, and the Nama (Grazhdankin, 2004). These assemblages were originally defined as three distinct temporal assemblages due to the view that there was a prevalence of sim ilar group of fossils present during distinct temporal intervals (Narbonne, 1999). However, recent studies from the White Sea region have found all three assemblages in roughly co eval strata reflecting fluctuations in depositional environments from continental shel f, near shore, inter-t idal, to estuarine and
3 back; a specific set of foss ils is associated with each environment (Grazhdankin, 2004). The explanation for the earlier view that cert ain assemblages were i ndicative of distinct temporal intervals is an artifact of the rela tively few horizons that ha ve been collected (in relation to the Phanerozoic) and the fe w (~25) localities found worldwide. Pteridinium fossils have been found on four c ontinents: North America, Africa, Europe, and Australia. At the sites where it is found, the fossils are usually preserved as lone individuals, or partial specimens. The most complete and best preserved of those fossils, such as those of North Carolina ( NA), southern Namibia, and the White Sea region of Russia, have been collected from out crops that are believed to have been laid down within deltaic environments. Because th ese sites are also the only locations where Pteridinium fossils are found in groups, these are thought to be the primary habitat of these organisms. Figure 1a, Composite Neoproterozoic carbon isot ope record, gray area encompasses the Ediacaran Period, modified from Halverson et al, 2004.
4 Figure 1b, Correlation of global Neoprotero zoic successions, from Knoll et al, 2006. In regards to preservation, the Ediacaran fossils seem to have lived in a unique taphonomic window, a time when there were few to no sediment burrowers, a factor that many researchers believe allowed their Â‘softÂ’ body traces to be pr eserved in otherwise coarse sediment. A further oddity of the Ed iacaran fossil record is the 3-dimensional preservation found at many localities. Very few organisms with hard skeletal parts existed in the Ediacaran; therefore, the biot ic record is dominated by trace fossils and impressions. A critical factor in their preservation is be lieved to be the ubiquitous cyanobacterial mats that were the primary sediment cover during the Neoproterozoic. These mats acted as both a food source and tomb for the majority of the Ediacaran biota (Fig. 2; Gehling, 1999; Grazhdankin and Seil acher, 2002; Droser et al, 2002, McCall, 2006), as a large percentage of them are hypot hesized to be Â‘mat suckersÂ’ (Seilacher, 1998, 1999) or tied to the mats in some way for their survival (Xiao and Knoll, 2000). These mats existed in many environments, whic h created numerous ma t textures (see Fig. 3a). An advanced or aged mat usually ha s preserved surface patterns, whereas younger, less developed mats are flat and generally only fossils are noticeable. These surface patterns are the result of th e current action influencing the top of the mat. Although advanced mat textures are not necessarily associated with Pteridinium fossils in Namibia or North Carolina (though this may be an artifact of the small availability of outcrops in NC), they are found together in the White Sea region of Russia and Australia (Gehling, 1999; Droser et al, 2002; McCall, 2006).
5 Figure 2, Mat ecology during the Precambrian, Modified from Seilacher, 1984 Figure 3a, Cyanobacterial mat types, modified from Gehling, 1999
6 When these organisms where covered with sediment during a burial event it is hypothesized that the oxidati on and mineralization of organic material along the matÂ’s surface was what preserved its impression on th e mat or sediment (based on whether the organismÂ’s inferred life habits, see figur e 3b); this is evidenced by pyrite layers associated with mat surfaces (Gehling, 1999). This has been dubbed Â‘death maskÂ’ preservation (Fig. 3b; Gehling, 1999), and onl y preserves the top of the organism; many Ediacaran organisms have no record of th eir mat-facing side (Gehling, 1999; Droser, 2006). The only preservation of qui lted organisms, such as Pteridinium is thought to have occurred in this manner. Pteridinium is a genus in the Vendobionta (Sei lacher, 1984), an extinct taxon of Â‘quiltedÂ’ organisms (Grich 1930, 1933; Ri chter 1955; Glaessner 1963, Seilacher, 1999). A quilted organismÂ’s body structure is dominated by a compartmentalized inflationary system, and this quilted body plan is a hallmark of Ediacaran organisms. Pteridinium is one of the best examples of this structure. In a quilted organismÂ’s body, the outer, tougher integument was supported by hydrostatic pressure from the body fluids contained within each individual section of the organism. Th ese individual units (vanes) are thought to structurally resemble a cnidarian polyp although they apparently lacked mouths, tentacles, and cnidae (Se ilacher and Pflug, 1982; Seil acher, 1984). This simple inflationary system was then repeated thr ough the modular addition of these vanes that created a Â“quiltedÂ” effect along the organismÂ’s central seam (Seilacher, 1984, 1988, 1992, 2007). A similar body structure is present in virt ually all Ediacaran fossils, and these can be subdivided into four groups based on the arrangement of the Â‘quiltsÂ’. These subdivisions include: serially segmenting, frac tal segmenting, and stalked versions of the previous two (Fig. 4). Stalke d Ediacaran fossils are some of the most complex of Ediacaran fossils and have been hypothesized to have had a life habit ranging from detrivores to chemotrophs (Gould, 1989; Jenkins, 1992; Jens en et al, 1998; Fedonkin et al, 2007). Although they are quilted, none of the stalked Ediacarans are thought to be as closely related to Pteridinium morphologically as non-stalked organisms. Dickinsonia Spriggina and Pteridinium are a few of the better-know n serially divided Vendobionts. Ernietta Phylozoon and Fractofusus are thought to be closely related to Pteridinium. Ernietta has similarities in the central seam and vane arrangements (see Fig. 4). Phylozoon resembles a two-walled Pteridinium, whereas Fractofusus is potentially more similar in that it is hypothes ized to have a third wall. Rangea Glaessneria Charnia and Paracharnia are some of the most studied fractally segmented Vendobionts (Jenkins, 1992; Brasier; 1999, 2001, 2003). Numerous other Ediacaran fossils, most notably those within the Triloboza exhibited tri-radial symmetry (McCall, 2006); whereas this symmetry is absent in all extant taxa tri-radial symmetry was common in the Neoproterozoic. Despite its abund ance in the Neoproterozoic, Pteridinium is the only known quilted organism to display tri-radial symmetry (Jensen et al, 1998; Jenkins, 2005; McCall, 2006) with the potential exception of Fractofusus
7 Figure 3b, Ediacaran Â“Death MaskÂ” Preser vation, modified from Gehling, 1999
8 Previous Work on Pteridinium Description of Pteridinium morphology For numerous decades the incomplete knowledge of the morphology and processes of preservation of Ediacarans pr evented a fuller understanding of not only Pteridinium but other quilted organisms as well (Grich, 1930, 1933; Richter 1955; Glaessner, 1963; Seilacher, 1999). The earlie st study of the taxon by Pflug (1970), on specimens collected from what is now know n as Aar Farm in southern Namibia, documented the unique construction of Pteridiniu m. He showed that it consists of a collection of three Â‘sheetsÂ’ or walls (referre d to by Pflug, 1970 as Â‘petaloidsÂ’) connected by a central seam to form the petalodium, wh ich results in a canoe-shaped body (Fig. 5). Figure 4, Quilted Ediacaran fossils with possible life habits, Pteridinium pictured is P. simplex from Namibia (modified from Seilacher, 2003). In his reconstruction, each of Pteridinium Â’s three sheets consists of a series of tubeshaped components (vanes), which appear sq uare in cross section. They initiate at a central seam, growing outward to terminat e in funnel-shaped apertures (Grich, 1930,
9 1933, Jenkins, 1995; Grazhdankin and Seilacher 2002). Pflug (1970) reconstructed the modular nature of these fossils as a simple r state between a colonial and a multi-cellular individual as that displayed by Modern siphonophores or chondr ophores (Grazhdankin and Seilacher, 2002; Seilacher and Pflug, 1984), a position that is still under examination (Fedonkin et al, 2007). However, it remains th e most widely accepted hypothesis for how most Ediacarans were constructed. This type of biological construc tion implies that each vane in Pteridinium belongs to a separate individual, w ith all of the zooids living together (in Pteridinium Â’s case, attached at the central seam), but lacking specialized functions as seen in various truly colonial groups. This t ype of construction was considered unique at the time, and Pflug (1970) erected a ne w order, the Petalonamae, to house Pteridinium and other Â‘petaledÂ’, quilted organi sms that would be found later. Pteridinium displays two primary outline shapes: ovate and flame-shaped. The ovate outline is the most common and is preferred in r econstructions of Pteridinium whereas the flame-shaped outline is found exclus ively in North America, specifically in the Slate Belt of North Carolina. Vane curvat ure can vary from almost none to curving back on the organism, while also affecting the thickness of the vane from equal to thinning thickness along its length. Figure 5, Pteridinium simplex unaltered canoe shaped body with middle median wall (top left) and folded specimenÂ’s possibly s howing intergrowth into the canoe shaped specimen, from Grazhdankin and Seilacher, 2002. Pteridinium Species In the past decade, morphological fact ors have aided in the debate on the characteristics that define Pteridinium species. The two currently recognized species, P. simplex and P. carolinaensis are differentiated in relatively simple terms. The type
10 Pteridinium fossil, as classically defined by Pfl ug (1970), is an organism having three walls joined along a central axis. These walls ar e comprised of an alte rnating insertion of tubular chambers (vanes), which are pot entially open at th e end. This classic Pteridinium specimen is considered by many to be P. simplex which is the type specimen, specifically based on those from Namibia. In these specimens, the vanes have little curvature and are usually only rigid near th e seam and commonly display a reduction in relief distally to the seam. The entire organism has an ovate outline when viewed from above, which is reconstructed as a more canoe-shape organisms when alive. In P. carolinaensis the vane relief is more pronounced w ith little flattening away from the seam. This species tends to have a flame-shaped or tear-drop outline with the vanes narrowing and curving back toward the center of the organism the further the vane is situated from the central seam. While these descriptions are the standa rd by which most researchers classify Pteridinium species, there is often sufficient va riation among specimens at the same locality that it becomes difficult to dete rmine where the boundaries between species should be positioned (Waggoner et al, 2001). Despite examples of both species being found at the major Nama Assemblage sites, in most cases the North Carolina and Namibian fossils are listed as different specie s due only to their diffe rences in shape in addition to their present-day locations. There is further confusion ar ising from differing body types generally that are pres ent at the same locality (though rarely in the same unit), Figure 6a, Two different inferred growth paths for Pteridinium with the addition of P. carolinaensis modified from Grazhdankin and Seilacher, 2002.
11 although the North Carolina specimens appear to display a single morphotype. Therefore, a more rigorous, statistical a pproach to analyzing the shape of these organisms is necessary to better understa nd their taxonomic and environmental relationships. Grazhdankin and Seilacher (2002) were the fi rst to examine a multitude of specimens, which were separated into two groups based on vane width as compared to total seam length (which they estimated was equivalent to total body le ngth) (Fig. 6a, b). This was the first attempt at a statistical test of Pteridinium Â’s morphology to document differences in its populations. Though th e reasons for the existence of the two groups are still unknown, the simple nature of the length versus width measurements is too general to be useful. It often ignores the various taphonomic or deformatio nal effects that the fossil could have been subjected too. More recent analyses of Pteridinium morphology have taken into consideration numerous other factor s, which include: outline, vane curvature, and specific preservational effects, such as warping or deforma tion during burial. The documentation of these features tends to be mo re problematic with those fossils preserved in three dimensions, whereas the features of flattened specimens are less ambiguous; they, however, may depart significantly from the Â‘livingÂ’ morphology. Preservational effects have contributed to artifacts in Pteridinium morphological comparisons. The central seam is an area of greater resistance to collapse or deformation during burial due to the height ened stability resulting from the convergence of the three walls. The existence of specimens composed of internal vane molds are rare, show limited detail of the vanesÂ’ interior, a nd are usually found in winnowed or scoured sediments (Fedonkin et al, 2007) resulting in poorly preserved fossils. These poorly preserved molds have been the s ources of questionable additions ( Pteridinium nenoxa (= Onegia nenoxa of Grazhdankin, 2004) and Pteridinium latum ) to Pteridinium (Grazhdankin, 2004; Fedonkin et al, 2007). Furt hermore, the evaluation of the effects driven by deformation have revealed a further obstacle to the study of these fossils as it has been thought by many aut hors that the sole reason for how a specimen appears is due to deformation from metamorphism or to preservational artifacts (Jenkins, 1995, Gehling and Narbonne, 1999). They cite the Â‘shearingÂ’ e ffects seen in some fossils which appear to be preferentially dist orted towards one side. Figure 6b, Linear regression of SeilacherÂ’s or iginal data; pink=North Carolina specimens, blue=Namibian specimens (Graz hdankin and Seilacher, 2002).
12 Pteridinium Habitat and Life Habit Unlike many Ediacaran fossils, Pteridinium appears to have been fairly cosmopolitan, is found over an extended temporal range, and has been collected from widely varied depositional environments (McCal l, 2006). It is most often associated with deltaic settings (G razhdankin and Seilacher, 2002) and marine environments where currents were rapid and unidirectional (S aylor et al, 1995). How it lived in this environment is still the s ubject of considerable debate. Various workers have hypothesized at least three di fferent life habits to Pteridinium : hyperbenthic, much like a sea pen or crinoids, (Jenkins, 1996), epib enthic (Gibson, 1984; Seilacher, 2007), and infaunal (Grazhdankin and Seilacher 2002). Jenkins (1996) interpreted Pteridinium as an alga based on its triple-bladed morphology, although there are no Modern algae that mimic this construction. He reconstructed Pteridinium as an erect organism anchored to either the cyanobacterial mat or sediment. Others have reconstructed all Vendobionts, including Pteridinium as lichens, the attachment scars of early plants, or large benthic foraminifera (Retallack, 1992, 2007; Seilacher et al, 2003). Grazhdankin and Seilacher (2002) were the first to propose a fully infaunal mode of life for Pteridinium based on the unusual three-di mensional (3-D) mode of preservation found in many Namibian specimens ; this hypothesis continues to be debated despite the limited evidence in support of their contention (Jenkins, 1995, 2007, Grotzinger, 2005, Narbonne and Gehling, 2007). Grazhdankin and Seilacher based this hypothesis on what they interpreted to be in situ specimens near the top of an event bed. They were interpreted as such due to the spec imenÂ’s lacking signs of directed stress, their convex-down orientation, and lack of evid ence for lateral colla pse (Grazhdankin and Seilacher, 2002). One counter-hypo thesis, still under debate, is that, because these were immobile organisms unaffected by winnow ing and found below the paleo-surface (implied by the top of event bed), they must have lived entirely submerged in the substrate. Although an infaunal life habit may increase Pteridinium Â’s preservation potential, it is still uncertain as to how a sessile organi sm could survive in such an environment, especially since the presence of a mat is hypothesized to make it difficult for oxygen to penetrate into the substrate (Gehling, 1999; Se ilacher, 1999). Given th at these specimens came from event beds representing a highenergy depositional mode, it seems unlikely that these fossils are in situ Jenkins and others (2007) have stated that the specimens used in Grazhdankin and Seilach erÂ’s (2002) study were curren t aligned, an unlikely event if they were not pres ent on the surface. Seilacher (2007) and others (Gibs on and Fedonkin, 1984; Grazhdankin, 2004; McCall, 2006; Fedonkin et al, 2007) have also suggested an epibenthic life position for some specimens of Pteridinium most notably from the White Sea region of Russia and North Carolina (with a few specimens from Au stralia), due to thei r preservation on the top of mat surfaces, mimicking Dickinsonia Â’s preservational mode (see Fig. 3b). In this reconstruction, the organism lay flat on the mat (two sheets touching) with one sheet oriented upright in the water column perp endicular to the mat (Seilacher and Pflug, 2007). This suggests that Pteridinium may have been more of a Â‘mat sucker,Â’ consuming the mat by external digestion, or using its er ect wall for suspension feeding possibly with of photosynthesizing symbionts. Figure 6c shows the two different life habits.
13 These last two life-habit interpretati ons stem from the simple nature of PteridiniumÂ’s body plan, which, by altering growth ra tes, can form slightly different shapes. If the vanes in each wall are short at the ends, the resulting shape will be three equal sized walls and a body plan that Seilacher (2003) would infer is adapted for life on top of the mat. If two of the walls grew l onger at the ends of the organism, the canoeshaped body is formed (Grazhdankin and Seilacher, 2002). Figure 6c, Two leading hypothesized Pteridinium life positions.
14 Geologic Setting and Pteridinium Preservation Namibia The units that contain the Namibian sp ecimens are derived from a suite of rocks collectively known as the Nama Group, which spans from ~550 Ma into the Cambrian (Fig. 7). This study, however, only examined specimens from the Ediacaran units, which are composed of two subgroups of the Nama Group: the Kuibis (550-548 mya) and the Schwarzrand (548-542 Mya). The Nama Gr oup was deposited during the initial formation of Gondwanaland, more precisely by the convergence of the Kalahari and Congo cratons with South America which woul d eventually close the Adamastor Sea and create the Transgondwanan Supermo untains (Squire et al, 2006). Figure 7, Map of Namibia including: Windhoek (cap itol), Aus and field site at Farm Aar. Shaded region is the extent of the Nama Group sediment.
15 The Nama is characterized by a series of transgressive/regressive sequences (Grotzinger et al, 1995; Saylor et al 1998; Fedonkin, 2007), although they may be eustatically driven some workers have hypothesized that local uplift during the continental collisions might have played a role in regulating relative sea-level changes (Saylor et al, 1995, 1998; Fedonkin, 2007). Th e group is characterized by shaly sandstones, coarse sandstones, and carbonates with eleven radiometrically dated ash layers (Saylor et al, 1995). Evidence from se dimentological studies indicates that these rocks formed in shallow water, most likel y a shelf-margin delta (Fedonkin, 2007). There has been extensive metamorphism in th e region (Hoffman, 1995; Grazhdankin and Seilacher, 2002), although it is surprisingly low grade in the study area given the higher grades seen in nearby regions. Due to this metamorphism, the sandstones in this sequence are now predominately quartzite. There is a th in diamictite early in the Schwarzrand, but it is unclear if it is glac ial or fluvial in origin (Kaufman, 2005; Fedonkin, 2007). Specimens used for this study come from the Kliphoek Member of the Dabis Formation, Kuibis Subgroup as exposed at Aar Farm, just southeast of the town of Aus in southern Namibia. The specimens from this locality are often found as three-dimensional casts, with an associated (non-internal) counter part mold (Gehling, 1999) representing a rare preservational mode that is seldom obs erved outside of Namibia (Grazhdankin and Seilacher, 2002). Pteridinium specimens are most often found on the bottom of coarse sandstones associated with deltaic environments repres enting unidirectional flow (Grazhdankin and Seilacher, 2002; Fedonkin et al, 2007). The K liphoek Member, a fossiliferous sandstone containing Pteridinium records strong current influence, with some units possibly being fluvial in origin (Germs 1983; Saylor et al., 1995; Grazhdankin and Seilacher, 2002). Unlike other regions, such as in No rth Carolina or Australia, Namibian Pteridinium specimens are regularly found in large groups instead of as lone specimens. This seemingly gregariousness nature appears to be more an effect of transport concentration than population dynamics (Fedonkin et al, 2007). As can be seen in Figure 8, the cyanobacterial mat on a slab termed the Â‘Seilacher BlockÂ’ appear s to have slid, though based on the amount of folding probably not much more than a meter (Grazhdankin and Seilacher, 2002; Droser, 2005). This might be indicative of a minor slumping of the mat, and not necessarily of any other high-energy proce ss as the mat remained coherent during movement. Figure 8, Seilacher Slab, with fold and crum ple structure area outlined by white square and highly visible folds poi nted out with arrows
16 North Carolina The North Carolina specimens where collect ed from Stanly County, a part of the Carolina Terrane, which extends from southwes tern Virginia to Georgia (Figs. 9a,b). It represents one of an extensive collection of peri-Gondwanan exotic terranes that were accreted onto the eastern margin of Laur entia in the Neoproterozoic and early Phanerozoic (Hibbard et al., 2002, 2009; W eaver et al, 2006). Rocks comprising the Carolina Terrane are usually lower greenschist facies due to regional metamorphism that occurred ~450 mya (Gibson and Huntsman, 1988, Weaver et al, 2006, Pollack et al, 2008). There are three commonly recognized stratigraphic units in the Carolina Terrane, which include, from oldest to youngest, the Uwharrie Formation, the Flat Swamp Member, and the Albemarle Group (Fig. 9b). The latter consists of the Floyd Church, the Yadkin, the Tillery, and the Cid formati ons (Conley, 1962; Conley and Bain, 1965; Stromquist and Sundelius, 1969; Seiders, 1978; Milton, 1984, Weaver et al, 2006). Radiometric dates taken from ash deposits i ndicate that the Yadkin Formation is unlikely to be older than Ediacaran, and Pteridinium carolinaensis specimens used in this study came from the underlying Floyd Church Form ation. There have also been possible examples of Swartpuntia sp. from the Cid Formation, which are common constituents of the Nama Assemblage (Fig. 9a).
17 Like the Nama Group, the depositiona l environment of the Floyd Church Formation has been interpreted as a shelf-ma rgin delta, below storm wave base (Weaver et al, 2006). Unlike the Nama Group, howev er, the rocks of the Albemarle Group are finer grained and are more clay rich, whic h would support an interpretation on a more distal position on the delta. The North Carolina units were deposited within a constricting basin forced by the collision between th e Gondwanaland volcanic island arc and Laurentia (Hibbard et al, 2009) This environmental interpre tation correlates well with those from other Pteridinium localities (see above). Weav er et al (2006, 2008) found evidence for depositional events with indi cators of a dominant flow direction. Unlike their Namibian counterparts, which are found in outcrop, the North Carolina material consist of float, with the specimens transported by fluvial, masswasting, or construction processes (Gib son and Fedonkin, 1984; Weaver et al, 2006). There is no primary Â‘sourceÂ’ site or quarry for these fossils, which reflects the regionÂ’s lack of well-exposed outcrops. The field locations of these fossils, in relation to the rock units they are found in, ar e displayed in Figure 9a. The initial North Carolina Pteridinium specimens were discovered in 1978 and orig inally described as trilobites by St. Jean (1979). Subsequently, they were determined to be Ediacaran organisms after additional fossils where found in the mid-1980Â’s (Weaver et al, 2006, 2008). However, in contrast to the research effort devoted to other Ediacar an localities, these have been little studied in the past ten years. Currently, based on the presence of Pteridinium and a possible Swartpuntia it is thought that the region preserve s a Nama Assemblage. There are other trace fossils found in the Slate Belt, but none are indicative of the Nama Assemblage. Figure 9b, Stratigraphic column and map of Stanly County, from Hibbard et al, 2009
18 Geometric Morphometrics Traditional morphometrics is a very useful tool that has a long history of use (e.g., Lohmann, 1983; Rogers, 1982; MacLeod, 1999; Ze lditch et al, 2004). It is the most common type of morphometrics used in paleobiology and relies on relatively straightforward measurements as the basis for comparisons (Fig. 10). However, it generally provides information focused on size, and only deals with shape in a relatively rudimentary fashion. Furthermore, this mor phometric approach records no information about the spatial re lationships between the discrete meas urements nor is there any way to determine if measurements between groups are homologous (MacLeod, 1999). An example of this is comparing measuremen ts between two differe nt arthropods with different maximum widths due to the pr esence or absence of non-homologous appendages (Fig. 11). The maximum widths for each arthropod are not homologous because the features measured on each sp ecimen do not represent identical body parts. Therefore, these two measurements cannot be used in geometric morphometrics. Fundamentally, the ability to measure the st atistical difference between two specimens while understanding how their respective shapes differ is why geometric morphometrics were used in this study. To date, the only analyses of Pteridinium Â’s morphology have largely analyzed their overall lengths and widths, as well as vane width (Ivantsov and Grazhdankin, 1997; Seilacher and Grazhdankin, 2002) Although these are quantitative, they do not properly capture many aspects of Pteridinium Â’s morphology nor the variab ility associated with its ontogenetic development. In comparison, geometric morphometrics simplifies the process by which a large number of spatial da ta points are analyzed for their relative variance while being able to compensate for confounding shape factors, such as those that might be caused by deformation during pres ervation or alteration after lithification (Hughes, 1995; McLellan and Endler, 1998). Figure 10, Traditional Morphometric measurements on a Trilobite.
19 Geometric morphometrics investigates th e degree of shape relatedness and the type of differences (disparity) each group exhibits when compar ed to one another. It uses recognizable and repeatable features of an organism as landmarks on a digitized image. These are then superimposed on a standard grid whereby each pi xel of the image, including landmarks, receives a specific coordinate. Each coordinate includes all the information about its position relative to all of the coordinates. This negates the use of numerous separate and complex measurements between features on an organism (such as total length, total width, total height, etc.; Fig. 10) that are typically required by more traditional morphometric approaches. Furthermore, this technique allows for specimens to be sampled multiple times to test whethe r deformation or biogeographic differences is the cause for the observed differences. This re -sampling technique is limited to organisms with a large amount of segmenta tion or repeatable units; this could not be applied to a lion skull, for instance (Macleod, 1999). Sp ecimens that are resampled should form statistically distinct cluste rs and act as a reference group with which to compare morphological variation. Figure 11, Non homologous width m easurements in two Arthropods. Landmarks The definition of landmarks is the basis for establishing the geometric morphometric field as they are used to de fine both shape and size. The data for each landmark consists of Cartesian coordinates, and all landmarks are taken with respect to the same set of x and y axes for any particular point on an organism. A perfect landmark, as described by Bookstein (1991), is one th at retains a homologous anatomical location, that does not alter its topol ogical position relative to other landmarks, and that can be found repeatedly as well as reli ably within the same plane as all of those other landmarks while adequately describing as much of th e entire organismÂ’s morphology as possible. These conditions are rarely completely met, as, in using 2-D photos, the third dimension can no longer be assessed. Thus, a tripartite classification has been developed consisting of type 1, 2, and 3 landmarks. Type 1 landmarks are the preferred type and are usually anatomically based, such as an intersection betw een two or more structures or tis sues (Fig. 12). Type 2 landmarks are usually mathematical landmarks (ones that are based off of the position of one or more Type 1 landmarks), which consist of su ch elements as the tips of structures or
20 maxima/minima of curves along the organism (Fig. 11). Type 3 landmarks Â– the least desirable of the three Â– are usually construc ted points or ones th at are located at a disproportionate distance from any of the ot her landmarks. Extreme distances from the main mass of landmarks tend to decrease the landmarkÂ’s ability to give accurate information about its position relative to other landmarks as there also tend to be more parts of an organism (complexity factor s) that are excluded in many instances. Landmarks should be chosen based on the question posed. If the question is simply whether or not specimens differ in shape, landmarks should be chosen to incorporate the entire shape. If th e question revolves around biomechanics or constructional stability of cer tain bones, then landmarks on parts of an organismÂ’s body that are irrelevant to the que stion will only obscure the in formation gathered. In this study, the landmarks where positioned to best capture the central s eam and the structure of three vanes, which encompasses the majo r morphological features of the organism. Differences in these main features can show major morphological disparities in Pteridinium that might result from environment, preservation, or development. There is also the need to keep the number of landmarks lower than the number of samples/specimens used, as if this is not followed it can obscure the results and/or invalidate the statistics behind the measurements. Figure 12, Landmark Types
21 Methods This projectÂ’s aim was to measure th e range of morphologic variability in Pteridinium from different sites in Namibia and North Carolina. Field specimens were augmented with additional specimens from th e Peabody Museum at Yale University and the National Earth Science Museum housed with in the Geological Survey of Namibia, Ministry of Mines and Energy. Table 1, Specimen Information Sample Group Sample # Specimen Folded Preservation type Cast Material North Carolina 1 EdiNC001 Y Flat Yes North Carolina 2 EdiNC001 Y Flat Yes North Carolina 3 EdiNC002 N Flat Yes North Carolina 4 EdiNC002 N Flat Yes North Carolina 5 EdiNC003 N 3-D No North Carolina 6 EdiNC004 N Flat No North Carolina 7 EdiNC005 N Flat No North Carolina 8 EdiNC005 N Flat No North Carolina 9 EdiNC006 Y Flat No North Carolina 10 EdiNC006 Y Flat No North Carolina 11 EdiNC006 Y Flat No North Carolina 12 EdiNC007 N Flat Yes North Carolina 13 EdiNC007 N Flat Yes North Carolina 14 EdiNC007 N Flat Yes North Carolina 15 EdiNC008 N Flat Yes North Carolina 16 EdiNC008 N Flat Yes North Carolina 17 EdiNC009 N Flat Yes North Carolina 18 EdiNC009 N Flat Yes Namibia 19 N1 Y 3-D No Namibia 20 N2 N 3-D No Namibia 21 N3 N Flat No Namibia 22 N4 N 3-D No Namibia 23 N5 Y 3-D No Namibia 24 N5 Y 3-D No Namibia 25 N5 Y 3-D No Namibia 26 N8 N 3-D No Namibia 27 N9 N Flat No Note; Table 1 continued on following page
22 Table 1(continued) Namibia 28 N10 N Flat No Namibia 29 N11 Y 3-D No Namibia 30 N11 Y 3-D No Namibia 31 N11 Y 3-D No Namibia 32 N11 Y 3-D No Namibia 33 N15 N Flat No Namibia 34 N16 N 3-D No Namibia 35 N17 N 3-D No Namibia 37 N20 N Flat Yes Namibia 38 N20 N Flat Yes Namibia 36 N18 Y Flat Yes Namibia 39 N18 Y Flat Yes Namibia 40 N18 Y Flat Yes Namibia 41 N23 Y Flat Yes Namibia 42 N24 Y Flat Yes Namibia 43 N25 N Flat Yes The North Carolina Pteridinium specimens come from collections housed in the North Carolina Museum of Natural Sciences, and the private collections of Gail Gibson and Steven Teeter. Eighteen of the 25 specime ns analyzed were original fossils, and these were augmented with seven casts from the Yale Peabody Museum (n=4) and North Carolina Museum of Natural Sciences (n=3). Th e casts were used to verify measurements recorded in the field of the original materi al. These redundant measurements were needed to test if field conditions might have had less optimal lighting conditions, changing camera angles that could skew how the fossils ap pear, or only if the size or location of the rock blocks the samples where inhibitive to the standardized photo collection method. This could potentially lead to inaccura te data acquired on those specimens. The redundancy measurements took place along the seam and along the width of the vanes. Webster and Hughes suggested an error of only 0.01 mm when using landmarks 5 mm apart. The measurements from the cast materi al varied by an aver age of 0.001 mm from their rock counterparts, which was a fraction of the average width of a vane (average width = 3 mm) and allowing cast material to be used. Table 1 disp lays the preservation type, if the sample was sampled multiple times, whether it was folded, and if it was a cast or an actual fossil. Image and Data Acquisition The specimens were photographed using a digital Olympus Stylus 790SW set at a resolution of 7.1 megapixels. Each specimen was placed 25 cm below the camera with two lights positioned to either side 1 m away from the specimen at an angle of 5 above horizontal to accentuate the relief (Fig. 13). F our different photographs of each specimen were taken; 1) in full light centered on its mu seum label, where pres ent; 2) in full light centered on the specimen; 3) with only the left light centered on the specimen; and 4) with only the right light centered on the spec imen. Pictures taken with light from only one direction (i.e., right or le ft) where used to create a co mposite containing the enhanced contrast of each orientation and these com posite photos where used for the study. Due to the parallel ridges formed as the result of Pteridinium Â’s vanes, samples where either placed with their central seam oriented perpendi cular to the lights or at a forty-five degree
23 angle from that line (figures 14a,b). This resulted in seven pictures of each specimen being taken. After combining all of these phot os into composite images it was found that the images with the seam perpendicular to the light sources e nhanced the contrast between lit and shadowed portions of the foss il more than the angled ones did, and so the former where used. Pteridinium can be found preserved in convoluted 3-D forms in addition to more compressed forms. In this study 3-D fossils we re considered to be those fossils that exhibited a wall extending perpendicular to th e main bedding plane or those fossils that where folded and where able to split apart al ong their central seam revealing a third wall; 2-D fossils where those that showed relief, but were confined to a singular preservation plane. Due to the nature of 2-D imaging and its use in morphometrics (Sheets, 2000, 2002; MacLeod, 2000), specimens were restricted to those that were lim ited in their relief (compressed) or had a relatively flat profile (in the case of 3-D fossils) and hence could be aligned parallel to the plane of th e cameraÂ’s lens along the central seam. Figure 13, Photo setup The primary geometric morphometrics soft ware used in this project is the Integrated Morphometrics Programs (IMP) suit e of applications developed and freely distributed by H. David Sheets of SUNY Buffal o. To convert the base images to the file types required for these analyses, and to position landmarks, TpsDig and TpsUtil developed by James Rohlf where used. The programs used to compute the statistics include: CoordGen6, PCAGen6, TwoGrop6, and DisparityBox. CoordGen6 formats the landmark data from TpsUtil and TpsDig so that it can be analyzed by statistical software. In this study, a generalized l east-squares Procrustes superim position was used to align the different samples using CoordGen6. The data is loaded into PCAGen6 in the form of an IMP file format (that was created in C oordGen6), and the program computes the Procrustes centroid based on all the landmar k data from the specimens analyzed. The centroid is the square root of the sum of the variances of the landmarks about that centroid in x and y -directions. The centroid size is us ed because it is uncorrelated, or not fixed, to any shape variable when land marks are distributed around a mean position (Sheets, 2002). This means that the centroid is the point that is the closest to every landmark in every direction. The principal components (i.e., the eigenvectors of the covariance matrix) are then calculated based on the covariance matrix derived from the partial warp scores. Finally, the data are plotted along any pair (though its starts with those having the heaviest loading) of princi pal component axes, in addition to plotting the deformation implied by the principal com ponent vectors should there be any. TwoGroup6 tests for significant differ ences in shape between two groups by determining an F-score using a Procrustes Superimposition. An Ftest is repeated (bootstrapped) numerous times to determine if the probability of the observed F-value
24 could have been produced by chance. To conduct the bootstrapping procedure, the two groups are combined into a common pool, and th en two groups with the same sample size as the original data sets are drawn wi th replacement from the common pool. The distribution of bootstrapped F-va lues over a large number of re sampled data sets is used to determine the probability that the obse rved F-value could have arisen by chance (Sheets, 2002); the samples were bootstrapped 2500 times because the results became consistent over 900 bootstraps. The si gnificance for this test is p=0.05. Figure 14a, Photo setup, position A Figure 14b, Photo setup, position B DisparityBox6 calculates disp arity (morphological diversit y of a group or clade) by finding the Procrustes distance between two or more groups (the distance between a specimenÂ’s Procrustes centroid and a groupÂ’s Procrustes centroid as well as between a groupÂ’s centroid and all the other groupsÂ’ cen troids). This is based on the entered landmark data using the approach developed by Foote (1993): D = (di 2 )/(N-1) Where di and i represent the distance of the centroid of each group (i.e., North Carolina vs Namibia) and from the centroid of all N gr oups, respectively; the distance metric used is the Procrustes distance. This was also bootstrapped 2500 times in the same manner as described above. Foote describes results with a negative mean distance ratio as essentially random with no clustering occurr ing (i.e. part of the same group), whereas those >1 show a considerable amount of diffe rence between clusters or groups. Zelditch suggests (2006) that as the disparity measur ement in two groups approaches 0, the size,
25 number, and tightness of cluste rs diminishes. In this study, a Foote signific ance value of 1 was used for interpreting the disparity results Statistical Tests The samples were run through two multivariate statistic tests: Principle Components Analysis (PCA), and a disparity test. PCA establishes hypothetical variables (principle components, PCs) that account for the majority of variance among individuals and plots them along two axis th at represent the most amount of variance of the first two PCÂ’s. In PCA, the first principal axis is calculated as the axis that accommodates the maximal variance which, due to the use of only x and y coordinates (Webster and Hughes, 1999), while the second principal axis is oriented orthogonally to the first axis, and shows the second-most variance; this patte rn continues with the third PC oriented orthogonally to the second showing the third most variance, and so on. The eigenvalues associated with each PC represent linear valu es of the overall amount of variance represented by each PC. When using PCA ther e are as many eigenvalues as there are variables, but a majority of them have little power to differentiate samples (see Table 2a). Zelditch (2006) recommends two methods fo r finding which eigenvalues are significant: keep any component with an eigenvalue >1 or look for an obvious break (elbow) in the list of eigenvalues, and keep any with values higher than this break; this is known as the Jolliffe cutoff. The PCA results can be refined by usi ng FooteÂ’s (1991) method to measure the amount of disparity present in the samples. Th e value of these measurements ranges from -2 to 2. Foote describes results with a nega tive mean distance ratio as essentially random with no clustering, whereas thos e >1 show a considerable amount of difference between clusters or groups. Zelditch (2006) suggests th at as the Foote disparity distance between two groupÂ’s approaches 0, the size, number, a nd tightness of clusters diminishes. While this means an increasing possibility that the two groups are parts of a larger single group, it does not diminish the statistical signif icance of the results as long as the 95% confidence interval range excludes negative numbers. The most common method to analyze th e landmark positions is to find the Â‘centroidÂ’ or center of mass of the landmarks, which assigns equal weight to each landmark. Procrustes analysis considers th e configuration of all landmarks on all specimens, calculating a best-fit superim position for each (Webster and Hughes, 1999). The optimal criteria that could be used, w ould be to calculate the fit of each sample, which calculates the least-s quares residual values across all landmarks (Sneath, 1967; Gower, 1975; Schonemann, 1970) or calculating th e best fit between landmarks that are the least different in location between speci mens, thus allowing for the possibility of shape differences being constrained to onl y a subset of the landmarks (Siegel and Benson, 1982; Rohlf and Slice, 1990). Procru stes superimposition involves specimen translation, rotation, and rescaling, but does not alter the configur ation of landmarks relative to each other (Rohlf, 1990; Ch apman, 1990; Bookstei n, 1991; Webster and Hughes, 1999; Sheets, 2007). Because specimens are independently rescaled to minimize the differences in their land mark configurations, allometr ies are represented by trends among superimposed points. As an example, rather than the plots of landmark coordinates tracing out a grow th path from small to large specimens, the landmark coordinates fall within a cloud of data points as the specimens are normalized to a similar size.
26 With a landmark system, deformation caused by lithification or from metamorphic alteration can be corrected for by the addition of a mean Procrustes centroid shape generated from sub-groups from the study population, such as the samples from only Namibia or North Carolina (Sheets, 2003) These vectors show the mean centroidÂ’s net magnitude and direction of displacement fr om all of the other landmarks within the sub-group; they are not compared to any other sub-groups. When used to compare localities, the magnitude and direction of these vectors represents the amount of deformation (compaction, stretching, scaling, et c) the samples have incurred (Figs. 15a, b) at each locality, if any. A majority of landmarks shifting in one direction is a strong indicator for deformation. Since each sample from a sub-group comes from the same respective unit, it can be assumed that the deformational factors affecting each sub-group would be similar, and should group the la ndmarks accordingly (Webster and Hughes, 1999; Sheets, 2007) Figure 15a, Procrustes deformation vectors for Namibian samples. -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 -0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25
27 Analysis In Pteridinium as is true of most Ediacarans fossils, landmark choice is difficult due to their simple body construction. Th ere are few places to position reliable landmarks, so there is an emphasis on how to best capture the shape of the vanes in a repeatable and standardized way (Brasier 2007). The only Type 1 landmarks found on most quilted fossils are along the central s eam (Fig. 16a). In this study, eight Type 1 landmarks positioned along the vaneÂ’s seam wh ere used as anchors for the placement of twelve Type 2 landmarks (see Fig. 16b), for a total of twenty landmarks for each sample. Standard square grids could not be used to constrain the positions of secondary landmarks uniformly as they could not eff ectively mimic the curvature of each vane. Therefore, a radial grid system was employed (Fig. 16c). Figure 15b, Procrustes deformation v ectors for North Carolina samples. -0.4 -0.2 0 0.2 0.4 0.6 -0.2 -0.1 0 0.1 0.2 0.3 Figure 16, Landmark placement on Pteridinium
28 The radial grid was aligned to the placement of landmarks 1 and 3. The spokes in the radial grid where spaced at 12.85 degrees (if the main x a nd y axis lines are included) to allow eight lines of intersection along the fossil (two additional lines were used by the author to aide in lining up the grid with the landmarks). Landmarks were placed where the grid lines intersected the junction between two vanes (landmarks 9-20 in Fig 16c,d). This number of landmarks was chosen primarily to allow a number of fragmentary specimens to be analyzed even though they only encompassed a limited portion of the central seam and number of vanes present. By limiting the sampling area within only four or five vanes more specimens could be sampled, some multiple times. Eighteen analyses were undertaken from nine North Carolina sp ecimens and 25 analyses where collected from 16 Namibian specimens for a total of 43 datasets collected. Specimens were resampled dependent upon the number of vanes present on each specimen; if sufficient vanes were present, a second sample was take n as far from the first sample as possible where an equivalent number of measurements could be made (see Table 1 for data on resampling). All but three specimens (due to a l ack of sufficient vanes) were sampled at least twice, with two specimens sample thr ee times. The primary aim of the re-sampling was to test whether deformation can affect the properties of a given specimen or where there is a large amount of variation pres ent in one specimen. Specimens where resampled in accordance with the number of va nes present; if a second sample could be taken, it was taken as far from the first samp le as possible where all landmarks could still be positioned on the specimen.
29 Results of Statistical Tests A review of the data set presented by Grazhdankin and Seilacher (2002) did not result in two statistically distinguishable populations, but only single population (6b), which under a linear regression (with a signi ficance level of p=0.05) was found to be insignificant. The Procrustes centroids for each landmark, every point from every sample, are plotted by locality in Fi gure 17a, with the combined means in green to gauge how tight the landmarks are cluste ring around their mean points. In figure 17b each localities means alone are plotted. In figures 17c and d each locality is plotted by landmark, using a different color for each landmark so, again, clustering can be observed. The results of each PCA plots for each variable studied are displayed in Figures 18a-c by region (Namibia or North Carolina), taphonomy (folded or not folded), and preservation (3-D or flat). The numbers next to each point correspond to the appropriate specimen (see Table 1). The bootstrapped Procrustes centroid distan ce F-test was found to have a p value of 0.1118. The Foote disparity distance was f ound to equal 0.00627. The eigenvalues are displayed in Table 2a for Pteridinium, the same data for the following piranha examples are found in Table 2b. Table 2a, Pteridinium Eigenvalues Table 2b, Piranha Eigenvalues Eigenvalue % Varience Eigenvalue % Varience PC1 0.02877 44.05 PC10.00261 58.37 PC2 0.01760 26.94 PC20.00067 14.88 PC3 0.00640 9.8 PC30.00026 5.74 PC4 0.00254 3.88 PC40.00017 3.83 PC5 0.00243 3.71 PC50.00014 3.22 PC6 0.00156 2.38 PC60.00011 2.43 PC7 0.00127 1.93
30 Figure 17a, General Procrustes centroid of all samples. Figure 17b, Mean Procrustes centroids of the two groups.
31 Figure 17c, North Carolina landmarks. -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 -0.5-0.3-0.10.10.30.50.7 Figure 17d, Namibia landmarks. -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 -0.5-0.3-0.10.10.30.50.7
32 Figure 18a, Locality PCA results by specimen Figure 18b, Preservational PCA results
33 Figure 18c, Taphonomic PCA results
34 Discussion Differing deformation at each site has been cited as a reason for the differences seen in the fossil, but there is no evidence for deformation since there is no prominent unidirectional vector acting on the landmarks (Fig. 15). There is no tight clustering of landmarks around their respective mean point s, with areas of over lap between each landmark mean. When there are only a few fact ors affecting a landmarkÂ’s position, this is reflected in fewer, higher eigenvalued PCs; this stems from a more dominate role that those fewer components have in controlli ng landmark placement, often forming tight clusters. When there are more factors affec ting a landmarkÂ’s position, th is is reflected in more, lower eigenvalued PCÂ’s with the clustering is not as tight as there are more components controlling each pointÂ’s position wi th less influence (see Tables 2a,b and Fig. 19a-c for examples from three piranha species, Sheets, 2002). When there are lower eigenvalued PCs, an ar gument could be made that the landmarks higher than 8 were Type 2 landmarks, and are prone to greater variability as they are more liable to slight differences in their pos itioning; however, the central seam landmarks should be more robust even with lower eigenva lues due to their Type 1 nature (Figs, 17c,d). Table 2b shows the same data for the piranhas as displayed in Table 2a for Pteridinium The piranhas have fewer higher eigenvalu ed PCs, showing the prevalence of only a few strong components acting on the poi nts while the p value (p=0.0408) of the bootstrapped Procrustes distances of these two species is significant. The p value of the Procrustes centroid distance and the disparity tests for the two Pteridinium species are not significant, which suggests that whil e the means of the two groups of Pteridinium appear as distinct as the known two piranha species (Figs. 17b and 19b), they more similar. Even when grouped by the differing variables, each PCA plot (Figs. 18a-c) shows the same placement of points because they are using th e same landmark data. Because the means change, these plots are useful in interpreti ng these differing variab les clustering trends even if the landmarks do not shift. The North Carolina samples appear more prevalent in the upper left corner whereas the Namibian specimens are more prev alent in the lower right. Interpretation of this is difficult as there is substantial overlap and distribution of the points, even ones that represent re-sampling from the same sp ecimen. Folded samples might display less variability, as there are groups of two or three, but these clus ters display a broad distribution (Fig. 18b). In PCA, the further from the origin (0, 0) a point lies the more distinct the point is from the others, often this is e xpressed by differing speciesÂ’ PCA points clustering away from zero (Fig. 19c piranha; Sheets, 2002). In the Pteridinium plots, all groups have their means close to the origin, with little deviation. Overall the clustering is very weak in a ll variables with all samples falling within each otherÂ’s 95% confidence intervals (Figs. 20a-c and Fig. 20c which is for comparison from the piranha study, Sheets, 2002). The lack of any two groups or a statistically significant line from
35 the traditional morphometric data set pres ented by Grazhdankin and Seilacher (2004), which suggests that all Pteridinium samples are part of the same group. Figure 19a, Landmarks from two different species of piranha clustering tightly around their means (green), from Sheets (2002). Figure 19b, Mean Procrustes centroids of th e two species of Piranha, Sheets (2002).
36 Figure 19c, PCA plot for two different species of Piranha, Sheets (2002). -0.12 -0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 -0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 Figure 20a, PCA plot with 95% ellipse for different localities.
37 Figure 20b, PCA plot with 95% ellipse for different Taphonomic factors. Figure 20c, PCA plot with 95% ellipse for different Preservational factors.
38 Figure 20d, PCA plot with 95% ellipse for tw o different Pirahna species, Sheets, 2002.
39 Conclusion Pteridinium is an enigmatic organism th at has long confounded researchers due to its tri-radial, quilted construction. Due to th e simple nature of that construction it has been difficult to assess species fidelity within the genus given more traditional approaches to morphology. Variations in observe d size, outline shape, and vane curvature have been attributed to differences in pr eservation, taphonomy, or re gionalism. None of these hypotheses is supported by robust statistical tests us ing geometric morphometrics. The analyses undertaken here suggests that all specimens studied from what have previously been defined as distinct species, Pteridinium simplex and P. carolinaensis, are morphometrically similar and he nce represent the same taxon. These findings have interesting implications for Ediacaran diversity. It is usually assumed that disparity increases rapidly and ea rly in the development of clades (i.e. the Cambrian Explosion, see Gould, 1989) follo wed by a tapering cessation of innovation due to the channelization of genotypic or phe notypic developmental pathways. This has not been seen in the Ediacaran fauna (McC all, 2006), in fact the opposite has been hypothesized (Gehling, 1991; Grotzinger et al, 1995; Fedonkin et al, 2007). Pteridinium being a genus containing a singul ar species is evidence that though there was still much niche space to be filled, the Vendobi onts do not seem to be filling them. Pteridinium was one of the few serially dividing quilted orga nisms thought to have numerous species, and therefore thought to be relati vity diverse, when compared to most genera of the Ediacaran. This lack of morphological divers ity, throughout the Ediacaran not just within the genus Pteridinium is enigmatic. It is known that there is less predation pressures during the Ediacaran, and global climate a nd oxygen was fairly stable throughout that time as well. It is possible that without any predation driving innovation the morphology of Vendobionts were slow to alter. Alternativel y, it is also possible that their biology or body construction prevented rapid burst of dive rsity. Future work should use the robust statistical tests of this study to delve into the morphological relations hips and adaptability potential of these enigmatic organisms.
40 References Antcliffe, J.B., Brasier, M.D., 2008. Charnia at 50: developmental models of Ediacaran fronds. Palaeontology 51, 11Â–26. Antcliffe, J.B., Brasier, J.B., 2007. Towards a morphospace for the Ediacara biota. In: Vickers-Rich, P., Komarower, P. (Eds.), Th e rise and fall of the Ediacaran Biota. Journal of the Geological Society of London. Special Publication, pp. 377Â–386. Antcliffe, J.B., 2004. On the form and relations of the Ediacara biota. M. Earth Sci. Thesis, Oxford University. Oxford, UK. Bookstein, F. L., 1991a, Thin-Plate Splines and the Atlas Problem for Biomedical Images. In Colchester, A. C. F. and Hawkes, D. J., editors, Information Processing in Medical Imaging number 511 in Lecture Notes in Computer Science, pages 326-342, Wye (UK). Springer-Verlag. Bookstein, F. L. 1991b., Morphometric tools fo r landmark data: geometry and biology. Cambridge University Press, Cambridge. Brasier, M.D., Antcliffe, J.B., 2004. Decodi ng the Ediacaran enigma. Science 305, 1115Â– 1117. Dzik, J., 2002. Early diversification of organisms in the fossil record. In Fundamentals of Life. Edited by G. Pa lyi, C. Zucchi and L. Caglioti. Paris: Elsevier Science S.A.: 219-248. Dzik, J., 2004. Anatomy and relationshi psof the Early Cambrian worm MyoswIex. Z o o I o g ~Sc ripfa 33: 57-69. Dzik, J., 2004. The Verdun Syndrome (A bstract). Sympos ium IGCP493. The Rise and Fall of the Vendian Biot a. Edited by P. Komarower and P; Vickers-Rich. Prato: Mona sh University: 15-1 9. Dzik, J., 2003. Anatomical information conten t. in the Ediacaran fossils and their possible zoological affinities. integr ative Comparative Biology 43: 114-126. Dzik, J., 2002. Possible Ctenophoran affi nities of the Precambrian "seapen" Rangea. Journal of Morphology 252: 31 5-334.
41 Dzik, J. and Ivantsov, 8. Yu., 2002. Internal anatomy of a new Precambrian dickinsonild dipleurozoan from northern Russia. Neues Jahrhuch Geolgie und Palaontologie Mh. 7: 385-396. Dzik, J., 2000. The origin of the mineral skeleton in chordates. Evolutionary Biology 31 : 1051 54. Dzik, J. and Ivantsov, A. Yu., 1999. An asymmetric segmented organism from the Vendian of Russia and the status of the Dipleurozoa. Hisforicaf Biology 13: 255-268. Droser, M.L., Gehling, J.G., and Jensen, S., 2005, Ediacaran trace fossils: True and false, in Briggs, D.E.G., ed., Evolving form and function: Fossils and development: New Haven, Connecticut, Yale Universi ty Peabody Museum of National History, p. 125Â–138. Fedonkin, M.A., 1977. Precambrian-Cambrian ichnocoenoses of East-European platform. In: Trace Fossils 2. Edited by T.P. Crimes and J.C. Harper. Geological Journal Special Issue 9. Li verpool: See House Press: 183-1 94. Fedonkin M.A. 1978. New locality of non-skel etal Metazoa in the Vendian of Zimnii Shore. Doklady Akademii Nauk SSSR 239 (6): 1423-1426. (in Russian) Fedonkin, M.A., 1978. Oldest trace fossils and the ways of behavioral evolution of mud eaters. Paleont ologicheskii Zhumal2: 106-112. (in Russian) Fedonkin, M. A., 1980. Fossil traces of Precambrian Metazoa. lzvestia Akademeyi Nauk, SSSR Series Geology 1: 39-46. (in Russian) Fedonkin, M. A., 1981. White Sea Biota of Vendian. Precambrian Non-Skeletal Fauna of the Russian Platform Nort h. Transactions of the Geological Institute, Academy of Scien ces of the USSR 342: 1-1 00. Fedonkin M. A., 1982. Precambrian non-skelet al fauna and the earliest stages of metazoan evolution. Third North American Paleontological Convention, Abstracts of papers. Journal of Paleontology 56 (2), Supplement 9. Fedonkin M. A. 1982. Precambrian soft-bodied fauna and the earliest radiation of invertebrates. Third North Am erican Paleontological Convention Proceedings 1 : 165-1 67. Fedonkin, M. A,, 1983. Organic World of the Vendian. Transactions of the Institute of Scientific and Technica l Information. ltogi nauki I techniki:
42 Stratigraphy, Paleontology. Vl NlTl12: 1-1 28. (in Russian) Fedonkin, M. A., 1983. Promorphology of the Vendian Radialia as a key to understanding the early e volution of Coelenterata Fourth lnternational Symposium on Fossil Cnidaria. Abstracts. 7-12 August 1983, Washington, D.C.: 6. Fedonkin, M. A., 1984. Promorphology of Ve ndian Radialia. In: Stratigraphy and Paleontology of the Most Ancient Phanerozoic. Edited by A. B. lvanovsky and A. B. Ivanov. Moscow: Nauka: 30-58. Fedonkin, M. A,, 1985. Paleoichnology of Vendian Metazoa. In: The Vendian System: 9toric-Geological and Paleon tological Basis Vol. 1. Edited by B. S. Sokolov and M. A. Ivanovskiy. Moscow: Nauka: 112-116 (in Russian) Fedonkin, M. A,, 1985. Precambrian metazoan s: The problems of preservation, systematics and evolution. Philosophi cal Transactions of the Royal Society of London B311: 27-45. Fedonkin, M. A., 1986. Precambrian problematic animals: Their body plan and phylogeny. In: Problematic Fossil Taxa. Edited by A. Hoffman and M.H. Nitecki. Oxford: Oxford University Press: 59-67. Fedonkin, M. A., 1987. Fossil traces. Priroda, 864 (8): 82-94. (in Russian) Fedonkin, M. A,, 1990. Non-skeletal fauna of the Vendian: Promorphological analysis. In: The Vendian System 1. Paleontology. Edited by B.S. Sokolov and A.B. Iwanowski. New York: Springer-Verlag: 7-120, 132-1 37. Fedonkin, M. A,, 1990. Precambrian metazoans. In: Palaeobiology. A Synthesis. Edited by D.E.G. Briggs and P.R. Crowther. New York: Blackwell Scientific: 17-24. Fedonkin, M. A,, 1992. Neoproterozoic ecosyst em restructuring: From net to pyramid. In: Fifth lnternational Conf erence on Global Bioevents. Abstract volume. Goettingen: 33-34. Fedonkin, M. A., 1992. Vendian faunas and th e early evolution of Metazoa. In: Origin and Early Evolution of the Me tazoa. Edited by J. H. Lipps and P. W. Signor. New York: Plenum Press: 87-129. Fedonkin, M. A., Gehling, J. G., Grey, K., Narbonne, G. M., Vickers-Rich, P., 2007, The Rise of Animals; evolution and divers ification of the kingdom Animalia, Johns Hopkins University Press, pp 326
43 Gurich, G., 1920. Die altesten Fossilien Suda frikas. ZeitschM PraMik Geologie 37: 85. Gurich, G., 1930. Die bislang altesten Spuren von Organismen in Sudafrick. lnternational Geological Congre ss of South Africa 1929, 15: 670-680. Gurich, G., 1930. Uber den Kuibisquarzit in Sudwest-afrika. Zeits chrifl Dec/tscli Geologie Gesellschafl82: 637. Gurich, G., 1933. Die Kuibis-Fossilien der Nama Formation von Sudwestafrika. I Palaeontologische Ze itschrifl 15: 137-1 54. Grotzinger, J. P., Bowring, S. A,, Sa ylor, B. Z. and Kaufman, A. J., 1995. Biostratigraphic and geochronologic constraints on early animal evolution. Science 270: 598-604. Grotzinger, J. P. and Rothman, D. H. 1996. An abiotic model for stromatolite morphogenesis. Nature 383: 423-425. Grotzinger, J. P., Watters, W. A. and K noll, A. H., 2000. Calcified metazoans in thrombolite-stromatolite reefs of th e terminal Proterozoic Nama Group, Namibia. Paleobiology 26 (3): 334-359. Germs, G. J. B., 1983. Implications of a sedimentary facies and deposition environmental analysis of the Nama Group in South West Africa1 Namibia. Special Publication of the Geological Society of South Africa 11 (1983): 89-114. Germs. G. J. B., 1995. The Neoproteroz oic of southweste rn Africa, with emphasis on platform stratigraphy and paleontology. In: Neoproterozoic Stratigraphy and Earth History Edite d by A. H. Knoll and M. Walter. Precambrian Research 73: 137-1 51. Germs, G. J. B., Knoll, A. H. a nd Vidal, G., 1986. Latest Proterozoic microfossils from the Nama Group, Namibia (South West Africa). Precambrian Research 32 (1 986): 45-62. Gibbons, W., 1989. Trace fossils from late Precambrian Carolina Slate Belt, south-central Carolina. Joumal of Paleontology 63: 1-10. Germs, G. J. B., 1972. The stratigraphy and paleonto logy of the lower Nama Group, South West Africa. Bulletin of the Precambtian Research Unit, University of Cape Town 12. Germs, G. J. B., 1972. Trace fossils fro m the Nama Group, South-West Africa. Journal of Paleontology, 46 (6): 864-870. Germs, G. J. B., 1968. Discovery of a new fossil in the Nama System, South
44 West Africa. Nature 219: 53-54. Gehling, J. G. 1982. The sedimentol ogy and stratigraphy of the late Precambrian Pound Subgroup, central Fli nders Ranges. South Australia, University of Adelaide, M.Sc. Thesis (unpublished). Gehling, J. G., 1987. Earliest known echinode rm a new Ediacaran fossil from the Pound Subgroup of South Australia. Alcheringa 11: 337-345. Gehling, J. G., 1988. A cnidarian of ac tinian-grade from the Ediacaran Pound Subgroup, South Australia. Alcheringa 12: 299-314. Gehling, J.G., 1991.The case for Ediacaran fossil roots to the Metazoan tree. In: The World of Martin Glaessn er. Edited by B. P. Radhakrishna. Geological Society of I ndia Memoir 20: 181-224. Gehling, J. G., 1996. Taphonomy of the Term inal Proterozoic Ediacara biota, South Australia. Unpublished Ph.D. Thesis, U.C.L.A. Gehling, J. G. and Rigby, J. K., 1996. Long-expected sponges from the Neoproterozoic Ediacara Fauna Pound Subgroup, South Australia. Journal of Paleontology 70: 185-1 95. Gehling, J. G., Runnegar, B. and Seilacher, A. 1996. Rasping markings of large metazoan grazers, termin al Neoproterozoic of Australia and Cambrian of Saudi Arabia. SEPM Meeting, Abstracts Gehling, J. G., 1999. Microbial mats in terminal Proterozoic siliciclastics: Ediacaran death masks. Palaios 14: 40-57. Gehling, J. G.. 2000. Environmental interpretation and a sequence stratigraphic framework for the termin al Proterozoic Ediacara Member within the Rawnsley Quartzite, South Australia. Precambrian Research 100: 65-95. Gehling, J.G., 2000. Sequence stratigraphi c context of the Ediacara Member, Rawnsley Quartzite, South Australia: A taphonomic window into the Neoproterozoic biosphere. Precambrian Research, 100: 65-95. Gehling, J. G., Narbonne. G. M. and Anderson, M. M., 2000. The first named Ediacaran body fossil: Aspidella te rranovica Billings 1872. Palaeontology 43: 427-456. Gehling, J. G., Jensen, S., Droser, M. L., Myrow, P. M. and Narbonne, G. M.. 2001. Burrowing below the basal Cambrian GSSP, Fortune Head, Newfoundland. Geological Magazine 138: 213-218.
45 Gehling, J. G., Droser, M. L., Jensen, S, and Runnegar, B. N., 2005. Ediacara Organisms: Relating Form to Funcfion. In: Evolving Form and Function; Fossils and Development. Edited by D. E. G. Briggs. Proceedings of a Symposium honouring Adolph Seila cher for his contributions to paleontology in celebration of his 80th birthday. New Haven: Peabody Museum of Natural Histor y, Yale University: 1-25. Gibson, G.G., Teeter, S.A., Fedonkin, M.A., 1984. Ediacaran fossils from the Carolina slate belt, Stanly County, North Carolina. Geology 12, 387Â–390. Gibson, G.G., Teeter, S.A., 1984.AstratigrapherÂ’s viewof the Carolina Slate belt, south central North Carolina. In: Carolina Geological Society 1984 Field Trip Guidebook, 43 pp. Gibson, G.G., Huntsman, J.R., 1988. Re-examination of the Gold Hill shear zone Cabarrus and Stanly County area, Southcen tral North Carolina. Southeast. Geol. 20, 51Â–64. Gibson, G.G., 1989. Trace fossils from late Precambrian Carolina Slate Belt, southcentral North Carolina. J. Paleontol. 63 (1), 1Â–10. Gibson, G.G., Teeter, S.A., 2001. Additional Ediacaran fossils from the late Precambrian Carolina Terrane, South-central, North Carolina. Southeast. Geol. 40 (4), 231Â– 240. Gould, S. J., 1989, Wonderful Life: The Burg ess Shale and the Nature of History: Hutchinson Radius, London, 347 p. Grazhdankin, D., 2004, Patterns of distribution in the Ediacaran biotas: facies versus biogeography and evolution: Paleobiology, v. 30, 179Â– 221. Hughes, N. C. 1999. Statistical and imaging methods applied to deformed fossils. Pp. 127Â–155 in D. A. T. Harper, ed. Numeri cal palaeobiology. Wiley, London. Hughes, N. C., and R. E. Chapman. 1995. Growth and variation in the Silurian proetide trilobite Aulacopleura konincki and its implications for trilobite palaeobiology. Lethaia 28:333Â–353. Hughes, N.C., 1995, Trilobite Taphonomy and Taxonomy: Palaios, v. 10, p. 283-285. Hoffman, P. F., Kaufman, A. J., Halver son, G. P. and Schrag, D. P., 1998. A Neoproterozoic Snowball Earth. Science 281 : 1342-1 346. Hoffman, P. F. and Schrag, D. P., 2000. Snowball Earth. Scientific American 282 (I): 50-57.
46 Hoffman P. F. and Schrag D. P., 2002. The Snowball Earth hypothesis: Testing the limits of global change Terra Nova 14 (3): 129-155. Hibbard, J.P., Stoddard, E.F., Secor, D.T ., Dennis, A.J., 2002. The Carolina zone: overview of Neoproterozoic to early Pa leozoic peri-Gondwanan terranes along the eastern flank of the southern Appalachians. Earth Sci. Rev. 57, 299Â–339. Hibbard, J.P., Pollock, J. C., Brennan, M., Samson, S. D., Secor, D., 2009, Significance of New Ediacaran fossils and U-Pb zircon ages from the Albemarle Group, Carolina Terrane of North Carolina, The Journal of Geology, volume 117, p. 487Â– 498 Halverson, G. P., Hoffman, P. F., Schrag, D. P., Maloof, A. C. and Rice, A. H. N., 2005. Toward a Neoprot erozoic composite carbon-isotope record. Geological Society of Ameri ca Bulletin 11 7 (911 0): 11 81 -1 207. Ivantsov, A., Grazhdankin, D.V., 1997, A new re presentative of the Petalonamae from the upper Vendian of the Arkhangelsk regi on: Paleontological Journal, v. 31, p. 1Â–16. Jenkins, J.F., 1992, Functional and ecological as pects of Ediacaran a ssemblages: Lipps, J.H., and Signor, P.W., eds., Origin and Ea rly Evolution of the Metazoa: Plenum Press, New York, p. 131-176. Jensen, R.J., 2003, The Conundrum of Mo rphometrics: Taxon, v. 52, no. 4, p. 663-671. Jensen, S., Gehling, J., Droser, M.L., 1998; Edia cara-type fossils in Cambrian sediments: Nature, v. 393, p. 567-569. Jensen, S., Saybr, B. Z., Gehling, J. G. and Germs, G. J. 8.. 2000. Complex trace fossils from Re terminal Proter ozoic of Namibia. Geology 28: 143-146. Jensen, S., Gehling, J. G., and Droser, M. L., 1998. Ediacaran-type fossils in Cambrian sediments. Nature 393: 567-569. Jensen, S., Droser, M. L. and Gehl ing, J. G., 2005. Trace fossil preservation and the early evolution of animals. Palaeogeography, Palaeociimatology, Palaeoecology 220: 19-29. Jensen, S., Droser, M. L. and Gehl ing, J. G., 2006. Acritical look at the Ediacaran trace fossil record. In : Neopmtemzoic Geobiology and Paleobiology. Edited by S. Xiao and A. J. Kaufman. Dordrecht: Springer: Chap. 5.
47 Knoll A. H., 1990: hathing room for early animals. Nature 382: 111-112. W A H., 2000. Leamlng to tell Neopr oterozoic time. Precambrian Reseam 100: 3-20. Knoll, A. H. and Cmll, S. B., 1999. Ea rly animal evodution: Emerging views From biology and geology. Science 284: 2129-21 37. Kaufman, A. J. and Xiao, S., 2003. Hi gh C02 levels in the Proterozoic atmosphere estimated from analyses of individual microfossils. Nature 425: 279-282. Kaufman, A. J., Knoll, A. H. and Narbonne, G. M., 1997: Isotopes, ice ages, and terminal Proterozoic ea rth history. Proceedings of the National Academy of Sciences of the United States of America 94: 6600-6605. LaFlamme, M., Narbonne, G., Anderson, M.M ., 2004, Morphometric analysis of the Ediacaran frond Charniodiscus from the Mistaken Point Formation, Newfoundland: Journal of Paleontology, v. 78, p. 827Â–837. LaFlamme, M., Narbonne, G. M., Anderson, M. M., 2004, Moprhometric analysis of the Ediacaran frond Charniodiscus from the Mistaken Point Formation, Newfoundland, Journal of Paleontology, v. 78, no. 5 Narbonne, G.M., Myrm, P.M., Landing, E. a nd Anderm, M.M., 1987. A New fossil type from the Precambrian-Cambrian boundary, Fortune Head, Avalon Peninsula, Newfoundland Canadian Journal of Paleontology, 24: 1277-1 293. Narbonne, G. M., Myrew, P.M.. Landing, E. a d Anderson, M.M., 1991. Medusoid hydrozoan from the basal Ca mbrian of Newfoundland. Canadian Journal of Paleontology 6.5 (2): 186-191. Milton, L. W., 1985. Chimney fossils Earth Science 38 (3): 24-27. McMenamin, M. A. S., 1986. The Garden of Ediacara. Palaios 1: 178-182. McMenamin M. A. S., 1998. The Garden of Ediacara: Discovering the First Complex Life. New York: Columbia University Press. McMenamin, M. A. and McMenamin, D. L. S., 1989. The Emergence ofAnimals. The Cambrian Breakthrough. New York: Columbia University Press. McCall, G. J. H., 2006. The Vendian (Edi acaran) in the geological record: Enigmas in geology's prelude to the Cambrian explosion. Earth-Science Reviews 77 (1 -3): 1-229.
48 McLellan, T., Endler, J.A., 1998, The Relative Success of Some Methods for Measuring and Describing the Shape of Complex Ob jects: Syst. Bio., v. 47, Part 2, p. 264281. McLellan, T., Endler, J. A., 1998, The relativ e success of some methods for measuring and describing the shape of complex obj ects, Systematic Biology: 47(2): 264-281 Miriam L. Z., Swiderski, D. L., Sheets, D. H., William, L. F., 2004, Geometric Morphometrics for Biologists, Cambridge Press MacLeod, N., Generalizing and extending the eigenshape method of shape visualization and analysis, Paleobiology 25 (1), 107Â–138, 1999. MacLeod, N., PaleoBase: Macrofossils (Par t 1) Blackwell Science and The Natural History Museum, Oxford, 2000. Pflug, H. D., 1970. Zur Fauna der Nama-Schichten in Sudwest-Afrika. II. Rangeidae, Bau und Systematisch e Zugehorigkeit. Sonder-Abdruk, Palaeontographica. Beitrage zur Natu rgescliichte der Vorzeit 135, Abt. A: 198-231. Pflug, H. D., 1972. Systematik der jungprakambrischen Petalonamae Pflug 1970. Palaeontologische Zeitschriff 46: 56-67. Pflug, H. D., 1972. Zur Fauna der Nama-Schichten in SudwestAfrika. Ill. Erniettomorpha, Bau und Systematik. Sonder-Abdruk, Palaeontographica. Beitrage zur Na turgeschichte der Vorzeit Bd. 139, Abt. A: 134-170. Pflug, H. D., 1973. Zur Fauna der Nama-Schichten in Sudwest-Afrika. IV. Mikroskopische Anatomie der Peta l-Organismen. Palaeontographica Palaeontographica. Beitrage zur Na turgeschichte der Vorzeit Bd. 144, Abt. A: 166-202. Pflug, H. D., 1974. Vorund Fruhgeschichte der Metazoen. Precambrian history of the Metazoa. Neues Ja hrbuch Geologie und Palaeontologie 145: 328-374. Pfluger, F., 1999. Matground structures and redox facies. Palaios 14: 25-39. Rohlf, F. J., 1990, Rotational fit (Procrustes) Methods. In Rohl f, F. J. and Bookstein, F. L., editors, Proceedings of the Michigan Morphometrics Workshop pages 227236, Ann Arbor, Michigan (USA). The Un iversity of Michigan, Museum of Zoology.
49 Rohlf, F. J. and Slice, D., 1990, Extensions of the Procrustes Method for the Optimal Superimposition of Landmarks. Systematic Zoology 39(1):40-59. Sheets, H. D., 2001. Regress6. Department of Physics, Canisius College, Buffalo, N.Y. Sheets, H. D., 2001. VecCompare12. Department of Physics, Canisius College, Buffalo, N.Y. Seilacher, A., 1989. Vendozoa: organismic construction in the Proterozoic biosphere. Lethaia 22, 229Â–239. Seilacher, A., Meschede, M., 1998. Precambrian fossil turns into tectonic diary. Geol. Soc. Am. Abstr. Programs 30 (7), 117. Seilacher, A., Meschede, M., Bolton, E., Luginsland, H., 2000. Precambrian Â“fossilÂ” Vermiforma is a tectograph. Geology 28, 235Â–238. St. Jean, J., 1973. A new Cambrian tr ilobite from the Piedmont of North Carolina. Am. J. Sci., Cooper Volume 273-A, 196Â–216. Seilacher, A., Grazhdankin, D., Leguta, A., 2003, Ediacaran biota: the dawn of animal life in the shadow of giant prot ists: Paleontological Research 7, 43Â–54. Seilacher, A., 1992, Vendobionta and Psammocoral lia: lost constructions of Precambrian evolution: Geo. Soc. of London, v. 149, 607Â–613. Seilacher, A., Grazhdankin, 2002, Unde rground Vendobionta from Namibia: Paleontology, v. 45, p. 57-78, 1pl. Waggoner, B. M., 1995. Ediacaran lichen: A critique Paleobiology 21 (3): 393-397. Waggoher, B. M,, 1999. Interpreting the ea rliest Metazoan fossils: What can we learn? Americen Zoologist 38: 9756-982. Waggoner, B. M., 2003, The Ediacaran biot as in space and time. Integrative and Comparative Biology 43: 1 041 1 3. Waggmer, B. M, Hagadorn, J. W., 2002. New fossils from terminal Strata of southern Nye Count y, Nevada. In: ProterozoicCambrian of the Great Bash end Beyond Edited by F. A. Corsetti. Field Ttii, Guidebook and Valutne Pvepaxed for the Annual Pac Section SEPM. Weaver, P., McMenamin, M., Tacker, C., 2006, Paleoenvironmental and paleobiogeographic impli cations of a new Ediacaran body fossil from the Neoproterozoic Carolina Terrane, Stanly County, North Carolina Precambrian Research 150, pg 123Â–135
50 Weaver, P.A., McMenamin, M., Tacker, R. C., 2006, Paleoenvironmental and paleobiogeographic impli cations of a new Ediacaran body fossil from the Neoproterozoic Carolina Terrane, Stanly County, North Carolina: Precambrian Research, v. 150, p. 123Â–135 Webster, M., H. D. Sheets, and N. C. Hughe s. 2001. Allometric patterning in trilobite ontogeny: Testing for heterochrony in Nephrolenellus Pp. 105-144 in M.L. Zelditch (ed.), Beyond Heterochrony: The Evolution of Development John Wiley and Sons, New York. Webster, M., and N. C. Hughes. 1999. Comp action-related deformation in Cambrian olenelloid trilobites and its imp lications for fossil morphometry Journal of Paleontology 73: 355-371. Xiao, S-H, AH Knoll, X-L Yuan, & CM Pu eschel (2004), Phosphatized multicellular algae in the Neoproterozoic Doushant uo Formation, China, and the early evolution of florideophyte red algae. Am. J. Bot. 91: 214Â–227. Zelditch, M. L., Swiderski, D. L., Sh eets, H. D., Fink, W. L., 2004, Geometric Morphometrics for Biologists: Maple-Vail, New York, 443 p. Zhang, Y., Yin, L., Xiao, S. and Knoll A. H., 1998. Perrninerelized fossils from the Terminal Proterozoic Doushantuo Formation, South China. Tm Paleontological Society Memoir 50: 1-52.