The Proterozoic Record of Eukaryotes Phoebe A. Cohen and … · 2015-09-15 · biases inherent in the Proterozoic fossil record, including issues of rock record availability, poor

Paleobiologyhttp://journals.cambridge.org/PAB

Additional services for Paleobiology:

Email alerts: Click hereSubscriptions: Click hereCommercial reprints: Click hereTerms of use : Click here

The Proterozoic Record of Eukaryotes

Phoebe A. Cohen and Francis A. Macdonald

Paleobiology / FirstView Article / September 2015, pp 1 - 23DOI: 10.1017/pab.2015.25, Published online: 10 September 2015

Link to this article: http://journals.cambridge.org/abstract_S0094837315000251

How to cite this article:Phoebe A. Cohen and Francis A. Macdonald The Proterozoic Record of Eukaryotes. Paleobiology,Available on CJO 2015 doi:10.1017/pab.2015.25

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/PAB, IP address: 65.112.10.78 on 11 Sep 2015

The Proterozoic Record of Eukaryotes

Phoebe A. Cohen and Francis A. Macdonald

Abstract.—Proterozoic strata host evidence of global “Snowball Earth” glaciations, large perturbations tothe carbon cycle, proposed changes in the redox state of oceans, the diversification of microscopiceukaryotes, and the rise of metazoans. Over the past half century, the number of fossils described fromProterozoic rocks has increased exponentially. These discoveries have occurred alongside an increasedunderstanding of the Proterozoic Earth system and the geological context of fossil occurrences, includingimproved age constraints. However, the evaluation of relationships between Proterozoic environmentalchange and fossil diversity has been hampered by several factors, particularly lithological and tapho-nomic biases. Here we compile and analyze the current record of eukaryotic fossils in Proterozoic stratato assess the effect of biases and better constrain diversity through time. Our results show that meanwithin assemblage diversity increases through the Proterozoic Eon due to an increase in high diversityassemblages, and that this trend is robust to various external factors including lithology and paleogeo-graphic location. In addition, assemblage composition changes dramatically through time.Most notably,robust recalcitrant taxa appear in the early Neoproterozoic Era, only to disappear by the beginning of theEdiacaran Period.Within assemblage diversity is significantly lower in the Cryogenian Period than in thepreceding and following intervals, but the short duration of the nonglacial interlude and unusualdepositional conditions may present additional biases. In general, large scale patterns of diversity arerobust while smaller scale patterns are difficult to discern through the lens of lithological, taphonomic,and geographic variability.

Phoebe A. Cohen. Department of Geosciences, Williams College, Williamstown, Massachusetts 01267.E-mail: [email protected]

Francis A. Macdonald. Department of Earth and Planetary Science, Harvard University, Cambridge,Massachusetts 02138. E-mail: [email protected]

Accepted: 2 June 2015Supplemental material deposited at Dryad: doi:10.5061/dryad.5pc3g

Introduction

The Proterozoic Eon encompasses some ofthe most dramatic changes in the history ofEarth and life, including the evolution andradiation of eukaryotes, the evolution of stemgroup metazoans, the origins of both eukar-yotic and metazoan biomineralization (Knoll2003), possible changes in the oxidative state ofthe planet’s atmosphere and oceans (e.g., Scottet al. 2008; Planavsky et al. 2014), and twoglobally distributed, low latitude glacial events(Hoffman 1998; Rooney et al. 2015). Protero-zoic environmental change has been recon-structed through a wide variety of geochemicaland sedimentological proxies (e.g., Lyons et al.2014; Li et al. 2013), whereas the record of life isreconstructed solely from the fossil record andmolecular clocks (Peterson et al. 2004; Knollet al. 2006; Erwin et al. 2011; Parfrey et al. 2011).

Since the proliferation of molecular clock data,a first order goal of Proterozoic geobiology hasbeen to reconcile these data with the fossilrecord (e.g., Sperling et al. 2009). While therehas been much concern about limitations ofmolecular clock data (Roger and Hug 2006),little has been done to address the myriadbiases inherent in the Proterozoic fossil record,including issues of rock record availability,poor age constraints, taphonomy, and theenigmatic nature of many Proterozoic fossils.Although Proterozoic paleobiology cannot yetattempt the massive sample standardized longterm analyses possible in the Phanerozoic(Alroy et al. 2008) the significant increase inpublications describing new eukaryotic taxafrom the Proterozoic (Fig. 1) combined with astatistical approach and a close analysis ofpotential biases enables us to attain a morenuanced understanding of what the fossil

Paleobiology, page 1 of 23DOI: 10.1017/pab.2015.25

© 2015 The Paleontological Society. All rights reserved. This is an Open Access article, distributed under the terms of theCreative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use,distribution, and reproduction in any medium, provided the original work is properly cited. 0094-8373/15

doi:10.5061/dryad.5pc3g

record can and cannot tell us about changingenvironments and biota in the Proterozoic Eon.

Here, we assess the existing record ofProterozoic fossils and test the robustness ofthis record by investigating potential biasespresented by taphonomy, fossil categoriza-tions, regional sampling, and uncertainties inage models. We then layer on existing paleo-geographic, geochemical, and climatologicaldatasets and assess potential relationshipsbetween eukaryotic diversification and environ-mental change. Questions we seek to illuminatewith improved datasets include: What was therelationship between eukaryotic diversificationand a putative rise in oxygen (Lenton et al. 2014;Planavsky et al. 2014)? Did the breakup of thesupercontinent Rodinia lead to changes in thediversity and distribution of microfossil assem-blages (Valentine andMoores 1970; Dalziel 1997;Hoffman 1998)? Was the diversification ofcrown group eukaryotes and origin of biominer-alization (Parfrey et al. 2011, Cohen et al. 2011)driven by tectonically modulated changes inocean chemistry (e.g., Halverson et al. 2010;Squire et al. 2006)? Did increased sinking ofnewly evolved mineralized tests drive changesin the biogeochemical cycles and climate(Tziperman et al. 2011)? What were the effectsof global glaciation (a.k.a. Snowball Earth;Hoffman et al. 1998) on microeukaryotes? Weremicroeukaryotic diversification and the appear-ance of metazoans driven by predation

(Porter 2011), changing ocean chemistry, orother factors? These questions can only be fullyanswered by a detailed and critical view of bothchemical proxy data and the fossil record itself.Below, we discuss hypotheses related to thenature of Proterozoic evolution and argue thattests must necessarily be limited to the fidelity ofthe fossil record.

Previous Assessments of ProterozoicDiversity

Proterozoic diversity has been assessed ineach of the last three decades (Vidal and Knoll1983; Vidal and Moczydłowska-Vidal 1997;Knoll et al. 2006). The earliest reviews (Vidaland Knoll 1983; Vidal and Moczydłowska-Vidal 1997) focused solely on acritarch taxa,assumed by the authors to be the remains ofphotosynthetic taxa. Other have looked atdifferent metrics of diversity, such as morpho-logical disparity, again focusing solely onacritarch taxa (Huntley et al. 2006) or macro-algal taxa (Xiao and Dong 2006). A laterassessment by Knoll et al. (2006) expandeddiversity analyses to include taxa not pre-sumed to be photosynthetic, such as vaseshaped microfossils, and looked at withinassemblage (alpha) diversity, as opposed tototal (gamma) diversity, in the Proterozoic.

The most critical differences between thisassessment of Proterozoic eukaryotic diversityand previous ones are the improved temporalresolution and the increased number ofpublications describing fossils, especially frombetween the Sturtian and Marinoan agedglaciations. For example, the last comprehen-sive assessment of eukaryotic diversity wasdone in 2006 by Knoll and co-authors; ourdatabase includes 32 manuscripts that havebeen published since 2006, which representsalmost a doubling of records. In addition, thisanalysis takes a more statistical approach andconsiders potential biases and influencingfactors in a way that has not been previouslyaddressed, including geography, lithology,and preservation pathways.

Here we followKnoll et al. (2006) by assessingdiversity within individual assemblages. Thereare several advantages of the within assemblage

FIGURE 1. Cumulative graph of publicationsdocumenting Proterozoic eukaryotic fossils; note largeincrease in publications since 2010. See SupplementaryAppendix for details.

2 PHOEBE A. COHEN AND FRANCIS A. MACDONALD

approach (Bambach 1977), especially for theProterozoic: 1) With poor age constraints,determining the relative or absolute temporalrelationship between stratigraphic units can bedifficult or impossible. For example, if twodiverse assemblages of different ages arelumped together because age constrains arepoor, summed diversity in a particular bin maybe exaggerated. Conversely, extracting a fossilassemblage from a bin due to poor ageconstraints could also lead to anomalouslylow diversity. 2) Most eukaryotic fossils fromProterozoic strata have poor taxonomic control.For example, smooth walled leiospherid acri-tarchs have been hypothesized to be algal,metazoan, or fungal, but these fossils likelyrepresent multiple eukaryotic and potentiallyeven non eukaryotic organisms. Although leio-spheres may be the most challenging fossilgroup in terms of assigning affinity, the problemof uncertain taxonomic affinity remains acrossmany other fossil types. In addition, differenttaphonomic windows make it challenging todetermine if microfossils from different litholo-gies are indeed the same species or even genus(e.g., Moczydłowska 2005). Many microfossilgroups have a small number of distinctivemorphological characters, and morphologicaldifferences or similarities may be obscured byvariable preservation. Thus, determining genusor species level diversity will be influenced bytaphonomy and the taxonomic decisions ofauthors, and determining overlapping speciesbetween varying lithologies may be intractable.In the future, we hope that taxonomic work

will move forward to the point where over-lapping taxa can bemore accurately determined,allowing for the distinction of origination, turn-over, and extinction rates in the Proterozoicfossil record. However, the scope of this paper islimited to within assemblage diversity patternsthrough time, which we believe at present is themost reliable indicator of biotic changes in theProterozoic.

Materials and Methods

Scope of Literature SearchAll data was compiled from published

literature; a broad search was undertaken

using Google Scholar and GeoRef databases.Search terms included the time designations‘Proterozoic’, ‘Precambrian’, ‘Vendian’, ‘Meso-proterozoic’, ‘Neoproterozoic’, and ‘Edia-caran’, plus ‘fossil’ and ‘microfossil’. Not allnon English literature is included, but thosenon English publications documenting a highdiversity of novel eukaryotic fossils wereincluded. Fossils were only included in thedatabase if they could confidently be assignedto Eukarya, with the exception of smoothwalled leiospherid acritarchs, which wereincluded despite the fact that there is thepotential for some of them to be non eukar-yotic. The Ediacaran fauna and end Ediacaranmetazoan taxa such as Cloudina are notincluded as we chose to focus on nonmetazoancomponents of the eukaryotic record. Thedatabase includes all found literature pub-lished before 01 March 2015. Each publicationreporting eukaryotic fossils was entered sepa-rately; therefore multiple entries may exist forone stratigraphic unit. However, overlappingtaxa were only counted once. For example, ataxon described in a publication on a formationfrom 1990 was not re-counted in anotherpublication on the same formation from 2000even if it is described in the later work. Thenumber of taxa described from each publica-tion was reported at the species level whereavailable, where no species names are given,we used morphotypes described and docu-mented by the authors of each publication.

Categorical Assignments of Fossil TaxaBecause of issues with taxonomic assign-

ments discussed above, we did not to categor-ize fossils by genus or species. In addition, notall Proterozoic fossils of interest have beenassigned taxonomic names (e.g., Bosak et al.2011a,b,c). Our approach was thus to count thetotal diversity described in each publicationand then assign fossils to one of ten morpho-logically based categories, building off of thecategories described in Knoll et al. (2006).These categories were chosen to capture themorphological diversity of any given strati-graphic unit (Fig. 2). The categories and theirdescriptions are as follows:

Smooth walled.—These fossils are defined asclosed, organicwalled structureswith no surface

PROTEROZOIC EUKARYOTES 3

ornamentation or processes. Generally, fossils inthis category fall into the leiosphaerid acritarchgrouping, a likely polyphyletic group of fossilswith few morphological characters, whichmakes their taxonomic affinity challenging toassess.While leiosphaerids are likely eukaryotic,some structures that would be considered in thesmooth walled category may not be eukaryotic;this possibility is dealt with in subsequentanalyses of the data.

Ornamented.—These fossils are defined asclosed, organic walled structures with surface

ornamentation, but no external processes.Examples include a variety of acritarch taxaincluding Valeria lophostriata. These structureshave a higher likelihood of eukaryotic affinitythan the smooth walled category due to theirmore complex exterior morphologies (Javauxet al. 2004) and general larger size in theProterozoic, however, they still may representa polyphyletic group.

Asymmetrical Processes.—These fossilsare defined as closed, organic walledstructures with external processes arranged

FIGURE 2. Representative images of fossil categories. A, Smooth walled organic microfossil. B, ornamented organicwalled microfossil Satka favosa, from Javaux et al. (2004). C, Organic walled microfossil with asymmetrical processes,Ceratosphaeridium sp. D, Organic walled microfossil with symmetrical processes. E, Vase shaped microfossil (VSM).F, Test of putative ciliate from Mongolia. G, Microscopic multicellular, Proterocladus from Butterfield (2009). H, Scalemicrofossil, Characodictyon skolopium. I, Macroscopic MOWS (macroscopic organic warty sheet) from Mongolia.


asymmetrically across the surface of thevesicle. Examples include the acritarch genusTrachyhystrichosphaera. Because of their complexexternal morphological structure, these fossilscan confidently be assigned to groupswithin theEukarya (Javaux et al. 2003). Further taxonomicaffinities are challenging to determine, thoughboth algal and fungal affinities have beenproposed (e.g., Butterfield 2005).Symmetrical Processes.—These fossils are

defined as closed, organic walled structureswith external processes arranged symmetricallyacross the surface of the vesicle. Examplesinclude the acritarch genera Gyalosphaeridiumand Alicesphaeridium. Due to their complexexternal morphological structure, these fossilscan confidently be assigned to groupswithin theEukarya including metazoans, microalgae, andpossibly fungi as well (Knoll et al. 2006; Yin et al.2007; Moczydłowska et al. 2009; Cohen et al.2009).Vase Shaped Microfossils.—Vase Shaped

Microfossils (VSMs) refer to microfossils withvariously shaped tests that in profile canresemble that of a vase, though morphologiesare variable. These fossils are believed to be thepreserved tests of Amoebazoan and Rhizarianorganisms (Porter and Knoll 2000).Macroeukaryotes.—This category captures

fossils that are larger than mm scale andconfidently assigned to the Eukarya. Examplesinclude various carbonaceous compressionfossils such as Chuaria (Vidal et al. 1993) andmacroscopic organic warty sheets found in theTaishir Formation of Mongolia (Cohen et al.2015). We have excluded somemacroeukaryoticfossils, such as Grypania, and other unnamedcarbonaceous compression fossils, whichcannot be confidently assigned to the Eukarya(Butterfield 2009; Sharma and Shukla2009; Srivastava 2012). While some of thesemacroscopic fossils may eventually bedetermined to be eukaryotic, we have erred onthe side of caution by excluding them from theseanalyses. For many carbonaceous compressionfossils, we follow the guidelines outlined in XiaoandDong (2006) to determine general consensuson the eukaryotic nature of contentious taxa.Microscopic Multicellular.—This category

captures multicellular forms that are sub mmscale in size. Examples include various algal

taxa such as Palaeovaucheria and Bangiomorpha(Butterfield 2004; Butterfield 2000).

Tests.—This category captures fossilsinterpreted as eukaryotic tests that are notVSMs. The most conspicuous example is theputative ciliate tests found in Sturtian aged capcarbonates of Mongolia (Bosak et al. 2011c).Fossils that are similar to VSMs but cannotconfidently be placed within the VSM categoryare also included here.

Scales.—This category captures the uniquescale microfossils from the Fifteenmile Group,Yukon (Allison and Hilgert 1986; Cohen andKnoll 2012).

Other.—This category captures enigmaticforms such as the “string of beads” inferred bysome authors to be eukaryotic (Calver et al. 2010).

Splitting vs. Lumping.—There is a risk thatdiversity assessments can be affected byindividual taxonomic assignments of authors.Thus, diversity could be inflated or deflated fora particular publication or stratigraphic unitbecause of the taxonomic techniques andstandards used by a particular researcher.However, the scope of this project and the useof statistical techniques such as binning and theproportional composition of biotas throughtime help mitigate these biases. In addition,some discretion was applied when determiningdiversity as presented in the publications, that is,some taxa previously split were lumped.

Age ConstraintsAge constraints for each fossil bearing strati-

graphic unit were determined based on thecurrent literature. Updated age constraintswere determined for many fossil assemblagesbased on new geochronological data and onrevised sequence and chemo stratigraphiccorrelations. Preference is given to preciseU/Pb zircon dates, but other dating techniquesare considered where these are lacking. SeeAppendix for data sources and information onage constraints of specific units.

In this review, we focus on the Mesoproter-ozoic and Neoproterozoic eras because few, ifany, definitively eukaryotic fossils have beenidentified in the Paleoproterozoic. We treatthe Mesoproterozoic as a whole due to thesmall number of fossiliferous assemblages. We


subdivide the Neoproterozoic following theinterim divisions approved by the Interna-tional Commission on Stratigraphy (ICS). TheICS recommended shortening the CryogenianPeriod, defining the base below the oldestCryogenian glacial deposit. The CryogenianPeriod is characterized by at least two globalglaciations that left diamictites on every paleo-continent between 720 and 635Ma (Condonet al. 2005; Macdonald et al. 2010b; Rooneyet al. 2015). The base of the Cryogenian Periodis thus taken here at 720Ma; the Tonian is thusdefined from between 1000–720Ma. TheEdiacaran is defined from between 635Maand 541Ma; additional stage boundaries arenot included here due to the relatively lownumber of fossil assemblages per stage.

Lithological, Taphonomic, Paleogeographic,and Environmental Assignments

Lithology was determined from data pre-sented in each publication. Where more thanone fossiliferous lithology was documented,

this was taken into account. For example, if asingle taxon was found in both carbonate andshale lithologies in a single publication, thenthe total diversity counted would only be one,but “carbonate” and “shale” would also bothreceive one diversity count each. Thus, the sumof the lithology diversity totals is potentiallyhigher than the total diversity for that publica-tion, if that publication describes multiple taxafrom more than one lithology. This is notedwhen applicable. Preservational categorieswere also taken from each publication accord-ing to those documented by the authors.

Paleogeographic assignments were deter-mined using information provided in eachpublication, along with recent refinementsto the Proterozoic geological record andpaleogeographic reconstructions. To avoidaddressing all possible paleogeographicreconstructions, we plot Proterozoic fossilfinds on a modern paleogeography (Fig. 3).This is sufficient for our purposes because weare concerned primarily with regional biases inreporting. However, these localities can easily

FIGURE 3. Map of locations of fossiliferous stratigraphic units in this study. Dotted margins represent approximateProterozoic paleo cratons used here.


be transferred to future paleogeographic mapsto address issues of endemism and potentialrelationships between paleogeography anddiversification.

Analytical MethodsAll analyses were performed using our

database and the statistical computingsoftware R, v. 3.0.1. For box plots, mean andmedian within assemblage diversity wascalculated for each Period with and withoutsmooth walled acritarchs. Upper and lower“hinges” were calculated as the first and thirdquartiles (the 25th and 75th percentiles). Forscatterplots, assemblages or publications wereplotted using their mean ages and trend lineswere calculated using LOESS smoothing(locally weighted polynomial regression).When the data allows, 95% confidence inter-vals were calculated and are shown aroundtrend lines using a t-based approximation.We used data based Monte Carlo simulation

(Kowalewski and Novack Gottshall 2010) toevaluate the potential role of sampling inproducing observed diversity trends. Withinassemblage diversity values were randomlyshuffled across all Periods, and for each Perioda number of within assemblage diversityvalues equal to the number actually sampledfrom that Period were extracted withoutreplacement and the mean within assemblagediversity calculated. We repeated thisprocedure 1000 times to produce 95% con-fidence intervals on mean within assemblagediversity values expected in the absence of anytemporal trend. This was done with all dataand then again without the two highestdiversity assemblages.

Results

Overall Trends.—Within assemblagediversity is low through the MesoproterozoicEra and into the early to middle Tonian Period.Towards the end of the Tonian Period, withinassemblage diversity increases, only to declinein the Cryogenian Period (Figs. 4, 5).A resurgence in within assemblage diversityoccurs in the aftermath of the Marinoan agedglaciation, and appears to be relatively stable

FIGURE 4. Within assemblage diversity (WAD) of allfossiliferous Proterozoic stratigraphic units. The height ofeach individual bar represents the total number of uniquespecies or morphotypes described per stratigraphic unit;stratigraphic unit age uncertainties or ranges are shownas the width of each bar.

FIGURE 5. Scatterplot of within assemblage diversity (WAD;number of unique species or morphotypes described) in eachpublication or stratigraphic unit by the stratigraphic unit’smean age. Trend line is LOESS smoothing (fitted locally).Grey shaded area represents 95% confidence intervalsaround the trend line using a t-based approximation. A,Diversity per publication. B, Diversity per stratigraphic unit.


until the end of the Ediacaran Period whendiversity drops off again. These overallpatterns hold when the data is analyzed bypublication and by formation and when thedata is binned by Period (Figs. 4, 5, 6). MonteCarlo simulation of the data support both thelow diversity in the Cryogenian, and thehigh diversity in the Ediacaran, thoughMesoproterozoic and Tonian diversity is within95% confidence intervals for random sampling(Fig. 7A). These trends hold even when the twohighest diversity assemblages (the DoushantuoFormation and the Fifteenmile Group) areexcluded (Fig. 7B). Importantly, low diversity

assemblages persist throughout the entireProterozoic. Thus, the increase in time binmean within assemblage diversity is due to anincrease in the maximum within assemblagediversity, not an increase in the minimum.Because the taxonomy of smooth walledmicrofossils is contentious, we also randiversity analyses excluding all smooth walledtaxa. These metrics show similar overall trendsas the full data set, but with slightly lowereddiversity in the earlier parts of the Proterozoic,and a decrease in the apparent higher diversityin the early Mesoproterozoic (Fig. 6B).

Our overall results are consistent with pre-vious assessments that document an increasein diversity metrics through the Proterozoicand into the Ediacaran (e.g., Knoll et al. 2006,Vidal and Moczydłowska-Vidal 1997, Huntleyet al. 2006). However, our results represent alarge increase in temporal resolution, espe-cially with relation to Cryogenian diversitypatterns.

Sensitivity to Stratigraphic and LithologicalBiases

Sensitivity to Lithology.—The fossil record ofeukaryotes in the Proterozoic is dominated byshale hosted biota (Fig. 8). Thus, there is a riskthat this dominance may be skewing the viewof overall diversity, as shale is deposited ina limited set of depositional environments,predominantly deep water or quiet watersettings restricted from open ocean conditions.The record of Proterozoic eukaryotes whenshale hosted taxa are excluded is challenging tointerpret because it is so data poor (Fig. 9).However, broad scale patterns of increaseddiversification through the Proterozoic remainapparent even without shale hosted taxa. Thishighlights the importance of new taphonomicwindows, including carbonates; without them,the non shale hosted record of eukaryoticdiversity would be uninterpretable.

The predominance of shale preservation inthe Proterozoic fossil record highlights the factthat changing dominance of taphonomicpathways and depositional environments overspace and time affect our view of fossil diver-sity. For example, prior to the evolution ofpelagic silica biomineralizing organisms,

FIGURE 6. Box and whisker plot showing a summary ofWAD data (number of unique species or morphotypes)binned by Period. Horizontal lines are the median. Theupper and lower box lines correspond to the first andthird quartiles. The upper whisker extends to the highestvalue that is within 1.5 of the inter quartile range (IQR).The lower whisker extends to the lowest value within1.5 * IQR of the hinge. Data beyond the end of thewhiskers are outliers and plotted as points. Solid blackdiamond represents the mean. A, Per stratigraphic unit.B, Per stratigraphic unit with smooth walled acritarchsexcluded.


Proterozoic silicification occurred pre-dominantly on carbonate platforms in peritidalsettings (Maliva et al. 1989), creating a rela-tively limited window for fossil preservationin shallow water carbonate environments.Conversely, preservation of organic walledtaxa such as acritarchs is most common in quietwater siliciclastic settings. Although recentwork has shown that carbonate rocks can alsopreserve organic walled tests and other organicstructures (Bosak et al. 2011c; Cohen et al.2015), acritarchs are comparatively rare from

carbonate macerations, suggesting tapho-nomic selectivity.

Biases may also be the product of diageneticenvironments or sample preparation. Forexample, biomineralized structures are rare insiliciclastic settings. This may be due to theeffects of different diagenetic pathways andpreservation in differing lithologies. Alter-natively, the lack may be due to the process bywhich these samples are prepared; traditionalhydrochloric and hydrofluoric acid macera-tions techniques would destroy any existing

FIGURE 7. Monte Carlo simulation of mean within assemblage diversity per Period. Open circles are the means of 1000random means (without replacement). Filled triangles are true means per Period. Bars represent 95% and 5%confidence intervals on simulation means. A, All data. B, Doushantuo Formation and Fifteenmile Formation excludedfrom the analysis.

FIGURE 8. Counts of fossil taxa described in each majorlithology per Period. Carb= carbonate, Phosph=phosphorite.

FIGURE 9. Within assemblage diversity (number ofunique species or morphotypes described) of fossiliferousProterozoic stratigraphic units, excluding all shale hostedbiota. The height of each individual bar represents thetotal diversity per formation; formation age uncertaintiesor ranges are shown as the width of each bar.


biomineralized material present. Microfossilsin siliciclastic lithologies can also be identifiedin thin section (Butterfield et al. 1994), but theuse of thin sections of shale in Proterozoicpaleontology is not widespread, and thus maynot be prevalent enough to discount a biasagainst mineralized fossils in shale lithologies.

In general, formations with multiple lithol-ogies have higher diversity. Examples includethe Fifteenmile Group, which has carbonateand chert hosted biota, and the Doushantuo,which has fossils preserved in chert, shale, andphosphorites. This is unsurprising, but usefulto consider when assessing diversity trendsthrough time.

Sensitivity to Geographic Location ofCollection.—The majority of publicationscurrently available document Proterozoiceukaryotic assemblages from Laurentia(Figs. 3, 10). When Laurentian localities areexcluded from the analysis, the increase indiversity into the Ediacaran is clearly apparent,but the increase in diversity seen in the Tonianin global compilations disappears entirely(Fig. 11). Thus, areas of the record do seemsensitive to a “Laurentian bias”. This bias

may be due to more micropaleontologicalwork in North America than in otherregions. Alternatively, in a macrostratigraphicframework, paleontological trends may bedriven by the abundance of rock packages

FIGURE 10. Histogram of number of stratigraphic units by mean age, separated out by craton. Only cratons with morethan two fossiliferous stratigraphic units are shown. Note the peak in Laurentian diversity in the mid late Tonian, andgap in the earliest Neoproterozoic.

FIGURE 11. Scatterplot of total within assemblagediversity (number of unique species or morphotypesdescribed) from non Laurentian localities described ineach stratigraphic unit by its mean age. Trend line isLOESS smoothing (fitted locally). Grey shaded arearepresents 95% confidence intervals around the trendline using a t-based approximation. WAD=withinassemblage diversity.


during specific intervals (Peters andHeim 2010) (Fig. 10). An abundance ofMesoproterozoic and Neoproterozoic basinsin North America may in part explain thecorresponding relative abundance of fossilreports; this also creates a bias in the recordtowards particular basin forming events onLaurentia. For example, gaps in the middleMesoproterozoic and early Tonian fossil recordcoincide with depositional hiatuses andlimited basin formation on Laurentia, and aminimum in passive margin developmentglobally (Bradley 2008). Thus, strengtheningmacrostratigraphic resolution for the Proterozoicwill be key in assessing the influence of regionalor basin scale dynamics on the fossil record.Sensitivity to Age Uncertainties.—Many

stratigraphic units in our compilation haveage constraints greater than 50 million years(Fig. 4, Supplementary Appendix) and the lackof better age constraints puts limits on theinterpretive power of some aspects of ourdataset. This emphasizes the importance ofadditional work on age constraints andcorrelations. For example, until recently,fossiliferous strata from Death Valley nowinterpreted as late Tonian (Macdonald et al.2013) were believed to be Cryogenian in age(Corsetti et al. 2003), which would haveelevated the diversity of that time bin by 40%.Noticeable Data Gaps.—The newly compiled

data and age constraints suggest that there is agap in fossil data in the early Tonian (Fig. 4).This gap coincides with an apparent low indeposition on Laurentia (Fig. 10), and thusmay be an artifact of rock availability on themost dominant craton in the Tonian. If so,this compilation presents a new perspectiveon sampling strategies. The late Tonian toCryogenian rifting of the supercontinentRodinia created a late Neoproterozoic peak inthe abundance of passive margin deposits(Bradley 2008; Fig. 10). Similarly, there isan apparent increase in within assemblagediversity during the late Mesoproterozoicfrom ca. 1250 to 1000Ma, but this may justbe due to the lack of well studied earlyMesoproterozoic (1600–1250 Ma) deposits.Currently, early Mesoproterozoic assemblagesare dominated by smooth walled taxa ofuncertain taxonomic affinity (Fig. 12), and the

discovery of one moderately diverse earlyMesoproterozoic fossil assemblage wouldeliminate the apparent late Mesoproterozoicdiversification. Thus, it is difficult to argue for arobust increase in within assemblage diversityuntil the late Tonian Period.

The Carbonate Taphonomic Window.—Recentwork has shown that carbonate rocks can hostdiverse Proterozoic fossil assemblages (Figs. 8,12) (Bosak et al. 2011a,b,c; Cohen and Knoll2012; Cohen et al. 2015). The addition of thecarbonate hosted window fundamentallychanges the record of eukaryotes in theProterozoic by populating the Cryogeniannonglacial interlude, as well as creating a newsearch image for future paleontological research.In addition, the types of fossils preserved incarbonate are different from those preserved inother lithologies, especially shale (Fig. 12). Thinorganic walled microfossils such as acritarchsare not often preserved in carbonate successions,whereas more robust forms such as theFifteenmile scale microfossils and putativeciliate tests from the Taishir Formation ofMongolia are (Cohen and Knoll 2012; Cohenet al. 2011; Bosak et al. 2011c). Thus, while thecarbonate window provides an additional andimportant view of Proterozoic diversity, it alsoserves as a reminder that no single lithology canaccurately capture the true fossil diversity of anytime period in Earth history (e.g., Porter 2004).

Importantly, most of the carbonate hostedmicrofossils have been recovered from a verynarrow temporal window in the ~ 10Myr afterthe ca. 717–660Ma Sturtian glaciation. Thismay be due to a biased sampling strategyfocused on the aftermath of Snowball Earthevents. Alternatively, rapid depositional rates(Rooney et al. 2014) and the peculiar carbonateenvironments of a post Snowball world (Prusset al. 2010) may have facilitated preservation ofthese fossils. This can only be addressed withmore systematic searches for acid resistantfossils in carbonate rocks throughout theNeoproterozoic

Distribution of Fossil CategoriesTemporal Patterns of Fossil Categories.—Some

fossil categories show variability in withinassemblage diversity through the Proterozoic,


while others do not (Figs. 12, 13). For example,smooth walled acritarchs show little variabilityin their diversity over the Proterozoic,maintaining low within assemblage diversitythrough the entire Proterozoic. As notedearlier, smooth walled acritarchs are likelypolyphyletic, and have little identifiablemorphology, making trends in diversitychallenging to asses. On the other hand, withso few morphological features, convergence islikely, and there could be patterns within thedata that are essentially invisible. The samelack of variability is apparent for ornamentedacritarchs (Fig. 13) perhaps for the samereasons: groups that have lower taxonomicspecificity will inherently have less of atemporal pattern. Other groups, such asacritarchs with symmetrical processes, VSMs,and macroeukaryotes, have distinct temporaltrends (Fig. 13). Some of this pattern is due toevolutionary innovation—for example,sampling of shale in the Proterozoic is highenough that it is unlikely that we are missing

many symmetrically acanthomorphicacritarchs or eukaryotic carbonaceouscompressions in older strata. Thus, we feelconfident that the patterns of withinassemblage diversity seen in these groups inour data are real. Other groups may showtemporal patterns because they are newlydiscovered or represent a new taphonomicwindow (i.e., the Fifteenmile Group scalemicrofossils). Thus, data on a specificcategory of fossil must be interpreted in lightof factors that may influence its occurrence inthe sedimentary record as well as in theliterature.

Distribution of Resistant Taxa.—Taxacategorized as resistant, which include VSMs,other tests, and scales, have only been describedin late Tonian and Cryogenian strata with oneexception (Fig. 14). As we would expect moreresistant taxa to have a more complete fossilrecord, this suggests that there is a true lack ofthese fossils in Ediacaran and Mesoproterozoicstrata. However, lithological and sampling

FIGURE 12. Number and type of fossil categories within each stratigraphic unit. Each bar represents one stratigraphicunit, colors represent the number of described species or morphotypes from each of the fossil categories, total heightrepresents the total number of described species or morphotypes per stratigraphic unit. Bars above each stratigraphicunit code represent the fossil bearing lithologies of that unit. WAD=within assemblage diversity.


biases may yet play a role, as there is a lack ofpublished microfossil assemblages and studieson Ediacaran carbonates, and many of the olderresistant forms have been described fromcarbonate environments (e.g., Cohen and Knoll

2012; Bosak et al. 2011a). The Ediacaran hasmore lithological units sampled per millionyears than the rest of the sampled Periods(Fig. 15), so we cannot account for the lack ofresistant taxa in shale, chert, and phosphoritethrough a lack of sampling.

Discussion

Diversity Patterns in Relation to EcologicalFactors

Relationship to Eukaryotic CladeDiversification.—Molecular clock estimatescalibrated from the fossil record (e.g., Parfreyet al. 2011) indicate that major eukaryoticgroups originated in the Paleoproterozoicand Mesoproterozoic eras, and furtherdiversified during the Neoproterozoic. TheMesoproterozoic and early Neoproterozoicrecords are currently too scant to corroboratethe earliest branches in the eukaryotic tree withfossil data. These clocks are calibrated with keyfossils such as ca. 1100Ma Bangiomorpha (Turnerand Kamber 2012; Butterfield et al. 2000),

FIGURE 13. Scatterplot of the number of described species or morphotypes in each stratigraphic unit by the unit’s mean age,separated by each fossil category. Scales not shown as they only have one occurrence. WAD=within assemblage diversity.

FIGURE 14. Within assemblage diversity of taxacategorized as resistant (VSMs, tests, and scales). Theheight of each individual bar represents the total numberof described species or morphotypes per stratigraphicunit; stratigraphic unit age uncertainties or ranges areshown as the width of each bar.


interpreted as a crown group red alga, the ca.800Ma Paleovaucheria, interpreted as a crowngroup Vaucherian (chromalveolate) alga, ca.800Ma Proterocladus, interpreted as a crowngroup green alga (Butterfield 2004; 2011), andca. 742Ma VSM taxa from the Chuar andCallison Lake formations representingamoebozoans and rhizarians (Porter and Knoll2000; Strauss et al. 2014). Thus, although withinassemblage fossil diversity remains low until thelate Tonian, eukaryotic clade originations arehigh during the Tonian, with at least four Toniantaxa that can be attributed to eukaryotic crowngroups appearing in the fossil record, andmolecular clock results showing a large numberof inferred originations before the late Tonian

(Parfrey et al. 2011). During the Cryogenian,putative foraminifera (Bosak et al. 2011a) andciliates (Bosak et al. 2011c) appear at ca. 660Ma.These three data sets—within assemblage fossildiversity, crown group FADs, and molecularclock estimates—indicate a possible discrepancybetween within assemblage diversity and theorigination of major eukaryotic clades (Fig. 16).

Role of Predation in Eukaryotic Diversification.—Was the microeukaryotic diversification seen inthe Tonian and Ediacaran driven by predation?The presence of ornate mineralized scales(apatitic scale microfossils, or ASMs) in the midTonian Fifteenmile Group, Yukon, has beensuggested to be a response to protistanpredation pressure (Cohen et al. 2011; Porter2011). This hypothesis is supported by theinterpretation of some VSMs as predatoryamoeba (Porter 2011). In fact, many singlecelled eukaryotes can predate on othereukaryotes of a similar or even larger size(Fenchel 1968; Han et al. 2007; Sayre 1973).More broadly, the role of predation in the fossilrecord has often been invoked to explaindiversification events in the history of life(Stanley 1973; Vermeij 1977; Huntley andKowalewski 2007). One issue with predationdriven hypotheses is that the Fifteenmile scalemicrofossils do not co exist with VSM fossils,and the majority of VSM occurrences appearapproximately 40Myr later in the Tonian.However, other heterotrophic predatoryeukaryotes may have co existed with the ASMtaxa, perhaps not leaving behind a robust fossilrecord. In addition, we would expect protistanpredation to increase, or at least remainrelatively constant throughout the Proterozoic,thus presenting a puzzle—why do resistantforms stop appearing in the fossil record beforethe Marinoan glaciation?

The Question of the VSMs.—As noted in ouranalyses, VSMs are relatively common duringthe late Tonian, yet are apparently absent in theearly Tonian, Cryogenian, and Ediacaran(Porter and Knoll 2000; Strauss et al. 2014).Many of these VSMs have been assigned tomodern eukaryotic clades that containstrikingly similar taxa. Thus, a conundrum ispresented—were testate amoeba not formingtests during the Cryogenian, Ediacaran and themajority of the Phanerozoic? Are the Tonian

FIGURE 15. A, Correlation between mean diversity (thenumber of described species or morphotypes) per Periodand the number of stratigraphic units with described fossilassemblages per Period. B, Correlation between meandiversity (the number of described species or morphotypes)per Period and the number of stratigraphic units per Maduration of Period. The length of the Cryogenian has beenshortened to account for the amount of time now estimatedthat sediments were being deposited during the twoglacial events. Cry=Cryogenian, Ed=Ediacaran, Mes=Mesoproterozoic, Ton=Tonian.


VSMs phylogenetically related to moderntestate amoeba or was there an extinction andre evolution of test formation in amoeboidgroups? Alternatively, perhaps there istaphonomic bias present in Ediacaran and

Phanerozoic lithologies that reduces thelikelihood of preservation? One interestingpossibility is provided by the fact that in themodern, testate amoeba are most common inlacustrine environments, so perhaps these

FIGURE 16. Overview of major events and fossil diversity in the Proterozoic. Carbon isotope data compiled fromMacdonald et al. (2009, 2010), Halverson et al. (2010), and Cox et al. (unpublished). SE= Shuram carbon isotopeexcursion, Tr=Trezona carbon isotope excursion, Tai=Taishir carbon isotope excursion, ICIE= Islay carbon isotopeexcursion, BSS=Bitter Springs stage, SGE= Sturitian aged glacial event, MGE=Marinoan aged glacial event,GGE=Gaskiers aged glacial event. Trend line and confidence interval for within assemblage fossil diversity fromFigure 5. Eukaryotic clade ranges from this analysis.


organisms experienced a change inenvironmental distribution and tolerance,which affected their preservation potential.

Relationship to Metazoan Origins andDiversification.—While we have assumed thusfar that fossils in our database reflect nonmetazoan eukaryotes, some component ofthis record, especially the Ediacaranacanthomorphic acritarchs, may actually be arecord of metazoans (Yin et al. 2007; Cohen et al.2009). Thus, this record is not necessarily entirelycomplementary, but may have taxonomicoverlap with a record of metazoandiversification. Molecular clocks of metazoansindicate a deep branching in the laterNeoproterozoic, with the roots of the metazoanclade seeded within latest Tonian to Cryogenian(Erwin et al. 2011). Molecular data has thusgiven us a rich picture of what we expect to findin the fossil record.However, the search for earlymetazoans has thus far been limited tobiomarkers from 711–635Ma rocks in Oman(Love et al. 2008) and various fossils of uncertaintaxonomic affinity (e.g., Maloof et al. 2010). It ispossible that other fossils categorized here asnon metazoan eukaryotes actually representearly branching or stem metazoans but areunrecognizable as such. The earliest metazoanswould have had physiologies very similar totheir microeukaryote cousins, thus it isreasonable to assume that any forces (such aschanging redox conditions) that influenced nonmetazoan eukaryotic diversification would alsoaffect metazoan eukaryotes.

Diversity Patterns in Relation to Tectonic andGeochemical Factors

The Effects of Snowball Earth Events.—Whatwere the effects of global glaciation (a.k.a.Snowball Earth) onmicroeukaryotes? Our onlyrobust view of eukaryotic life in theCryogenian nonglacial interlude comes from alimited number of low diversity carbonatesamples from Mongolia and Namibia (Bosaket al. 2011a,b,c; Dalton et al. 2013) and shalehosted biota from Australia (Riedman et al.2014). Prior to the identification of thecarbonate window it had previously beensuggested that the Cryogenian Period wasvariously depauperate, or contained the rich

Fifteenmile scale biota (Kaufman et al. 1992)and VSMs of the Pahrump Group (Corsettiet al. 2003), however, this was based on theabsence of fossils other than leiospheres inAustralian drill cores and poor age models(Macdonald et al. 2010a,b; Macdonald andCohen 2011). Recent geochronologicalconstraints have trimmed the duration of theCryogenian nonglacial interlude to between660 and 635Ma (Zhou et al. 2004; Condon2005; Rooney et al. 2014), which indicates thisinterval may be additionally biased by itsnarrow temporal range. In addition, much ofthe Cryogenian record is composed of glacialdiamictites that are not well suited topreserving microfossils. If we consider thatapproximately two thirds of Cryogenian strataformed during global glaciations, thislithological difference could have stronglybiased fossil preservation rates.

The Cryogenian carbonate hosted fossilsthat have been discovered in recent years allcome from the Cryogenian nonglacial inter-lude. Our understanding of life during theglaciations themselves may remain elusive,due to a reduction of normal sedimentationduring ice coverage. Molecular clocks, cali-brated with older fossil occurrences, allow us aglimpse into the evolution of eukaryotes at thistime. These results indicate that major diversi-fications were occurring during the Cryogen-ian (Parfrey et al. 2011). However, if eukaryoticcommunities were provincial, surviving insmall population sizes in relatively small geo-graphic areas, the limited sampling from theCryogenian nonglacial interlude is likely notadequate to capture the full record of evolu-tionary innovation. Geographic isolation mayalso be responsible for the diversification ofeukaryotic groups during the Cryogenian.Several recent studies have documented evi-dence of geographic isolation and divergencein modern protist groups (Boenigk et al. 2006;Foissner et al. 2007; Casteleyn et al. 2010),suggesting that protists in modern ecosystemsare not cosmopolitan. As such, global glacia-tion may have potentially stimulated diversi-fication, while keeping overall abundance low.

Despite the issues outlined above, fossilsampling of the Cryogenian is much morecomplete than it was a decade ago


(e.g., Riedman et al. 2014). The question arisesthen as to whether the signal of low diversityand depauperate fossil assemblages is real, or abias of the record as described above (Fig. 5). Inorder to test the robustness of the low diversityseen in the Cryogenian nonglacial interlude,we ran a Monte Carlo simulation (Figure 7)which shows that the within assemblagediversity in the nonglacial interlude is lowerthan would be expected from a random dis-tribution of the diversity data.

To further analyze the unique nature of theCryogenian, we plotted the number of strati-graphic units per period against the meandiversity in that period (Fig. 15A). At firstglace, the lowmean diversity of assemblages inthe Cryogenian is consistent with the lownumber of sampled stratigraphic units (Fig.15A). However, the Cryogenian stands out instark contrast to the Ediacaran, which has asimilar number of fossiliferous stratigraphicunits, but a much higher mean diversity. Whenthe same data is plotted against the number ofstratigraphic units per million years, and theCryogenian is constrained only to the non-glacial interlude, the Cryogenian again standsout as an outlier with respect to mean diversity(Fig. 15B). Whether or not this is exclusively atrue biological signal, or an artifact of the non-glacial period’s short temporal time span andunique sedimentological regime remains anopen question.

Nonetheless, the current record does pre-serve a remarkable turnover in eukaryoticmicrofossils from diverse assemblages ofresistant tests and microscopic multicellulareukaryotes to low diversity assemblages ofagglutinating tests in the Cryogenian non-glacial interlude, and a radiation of acritarchswith symmetrical processes and algal taxa inthe Ediacaran (Fig. 12). Thus, it is clear thatmany clades of eukaryotes survived theSnowball earth events, but different membersof those clades appeared both between andafter the both global glaciations. Further workis needed both to confirm this apparent trend,as well as to understand taxonomic selectivityacross the glaciations.Break up of Rodinia: Weathering and

Biomineralization.—The apparent coincidenceof the break up of the supercontinent Rodinia

(Li et al. 2008) and the diversification ofeukaryotes including the origin of themetazoan clade, is striking, and has been notedby several authors (Valentine and Moores 1970;Dalziel 1997; Hoffman 1998; Fig. 16). Potentialdriving mechanisms could include an increasein geographic isolation driving allopatricspeciation, but assessing the likelihood of thishypothesis will require a better handle onboth taxonomy and paleogeography in theNeoproterozoic.

Although the timing and nature of the breakup of Rodinia is still debated (for a recentreview, see Evans 2013), it is apparent thatseveral margins began to rift between ca. 830and 780Ma with the deposition of continentaldeposits in narrow grabens (Wang et al. 2011;Li et al. 2013), which coincided with theemplacement of the ca. 830Ma compositeGuibei Willouran large igneous province (LIP)in South China and Australia and the ca.780Ma Gunbarrel and Kanding LIPs ofwestern Laurentia and South China (Fig. 16;Wingate et al. 1998; Li et al. 1999; Ernst et al.2008; Wang et al. 2008).

Rifting of Rodinia and the widespreademplacement of large igneous provinces mayhave led to a myriad of chemical changes in theocean that culminated in both evolutionaryand climate change. A long term rise in stron-tium isotope values in carbonates through theTonian are consistent with low latitude rifting,an increase in continental margin length, andincreased global silicate weathering (Halver-son et al. 2010). This rise in strontium isotopevalues through the Tonian is stalled by shortterm falls coincident with the emplacement ofunradiogenic LIPs (Halverson et al. 2010). Onaverage, basalt contains ~3 times as muchphosphorus as granite (Ronov 1982; Halversonet al. 2014; Cox et al. unpublished) and the lowlatitude weathering of extensive LIPs between820 and 720Ma may have led to an additionalincrease in phosphorous delivery to the oceans.Fe speciation studies through Tonian andCryogenian strata have found predominantlyanoxic and ferruginous subsurface conditions(Poulton and Canfield 2011) with oxic shallowwater (Sperling et al. 2013) and local euxiniathrough organic carbon loading (Johnston et al.2010). These conditions are ideal for the


“Fe-P shuttle” (Berner 1973; Poulton and Can-field 2006; Creveling et al. 2013) and enhanceddelivery of phosphorite to the ocean, whichmay have been limited in the Mesoproterozoic,not only by the lack of weatherable phosphor-ous rich basalts, but also by low surface oxy-genation and a riverine phosphorous trap(Laakso and Schrag 2014). If phosphate was thelimiting factor on primary productivity duringthe Tonian Period, then an increase in bothtotal silicate weathering and the weathering ofphosphorous rich basalts may have allowedsufficient phosphorous to skip the riverine trapand resulted in both an increase in primaryproductivity in the oceans, fractional organiccarbon burial, and free oxygen. These ideas areconsistent with a rise in phosphorous ironratios recorded in iron formation in Sturtianglacial deposits (Planavsky et al. 2010). Asdiscussed below, δ13C data is also consistentwith an increase in fractional organic carbonburial during the early Tonian Period,although other factors are likely at play indriving variability in δ13C records (e.g., Schraget al. 2013).

Studies in modern ecosystems indicate thatincreasing nutrients, including phosphate, arepositively correlated with biodiversity andspecies richness over large spatial andtemporal scales (Chase and Leibold 2002).Modern studies of phosphate are mainlyrestricted to lacustrine environments, as P isless limiting than other key nutrients in themodern ocean. However, as noted above,P may have been more limiting in lateMesoproterozoic to early Neoproterozoicmarine ecosystems, thus providing a potentialabiotic control on eukaryotic diversification.Increased phosphorous concentrations in theocean would have also increased the avail-ability of phosphate for biomineralization.Indeed, the first occurrence of phosphaticbiomineralization in eukaryotes (Cohen et al.2011) occurs directly on the heels of theemplacement of the ca. 830Ma LIPs in SouthChina and Australia (Wingate et al. 1998; Liet al. 1999; Wang et al. 2008).

The rifting of the supercontinent Rodiniamay have not only changed chemical environ-ments, but also physical environments.Thermal subsidence of many continental

margins (Bradley 2008) increased the area ofnear shore and epicontinental sedimentaryenvironments (Li et al. 2013), thus increasinghabitat area for eukaryotes living in relativelyshallow near shore environments, includingearly metazoans. In addition, organisms livingin near shore environments have a higherchance of being preserved in the sedimentaryrecord. Quantifying the potential bias for moreepicontinental preservation will require amacrostratigraphic analysis (c.f. Peters andHeim 2010) of Proterozoic sedimentaryrecords.

Relationship to Geochemical Proxies for OceanOxygenation.—Hypotheses to explain theorigin and radiation of metazoans commonlycall on a Neoproterozoic rise of oxygen (Cloud1968; Rhoads and Morse 1971; Knoll 1999),however, increasing oxygen levels do notprovide a driving mechanism for thegeneration of biodiversity (Marshall 2006;Erwin et al. 2011). In fact, early animals maynot have required elevated oxygen levels (Millset al. 2014), but rather oxygenation was likelymore important for metabolically expensiveactivities like predation (Knoll and Sperling2014). The role of oxygen in the radiation ofeukaryotes is even less clear. Almost alleukaryotes require some amount of freeoxygen, though many modern taxa live inlow to dysoxic environments for extendedperiods of time, and some can live in anoxicenvironments for weeks to months (Heinz andGeslin 2012). Thus, it is unlikely thatNeoproterozoic oxygen levels in and ofthemselves would have directly constrainedthe radiation of eukaryotic groups.

Moreover, independent geochemical evi-dence for a Neoproterozoic oxygenation eventis not well resolved in space and time. Theconcept of a Neoproterozoic oxygenation eventwas reviewed (Kah and Bartley 2011) andextended by Och and Shields (Och andShields-Zhou 2012) who used molybdenum,uranium, and vanadium concentrations toargue that a rise in oxygen had occurred by theEdiacaran, but the lack of analyses onCryogenian samples precluded an under-standing of the relationship to glaciation.An early Ediacaran rise in oxygen was alsosuggested from molybdenum, uranium,


vanadium concentration measurements andsulfur isotope analyses (Partin et al. 2013;Sahoo et al. 2013), however, it has remaineduncertain if there was an earlier rise, and ifoxygen levels remained high throughout theEdiacaran (Johnston et al. 2013). In fact, statis-tical analysis of iron geochemical data doesnot show a significant change through theEdiacaran and Cambrian periods (Sperlinget al. 2015).

Planavsky et al. (2014) used Cr isotope stu-dies to argue for an increase in oxygen at ca.810Ma to >0.1% PAL. This would seem tocorrelate nicely with the appearance of severaleukaryotic clades in the fossil record (Fig. 16);however, this estimate is model dependent andlacks global coverage from earlier strata tosubstantiate that it represents a significant rise.

If oxygenation was a trigger to diversifica-tion, it is likely that the effects of rising oxygenwere felt by eukaryotes most strongly throughoxygen’s effects on the bioavailability of otherelements and nutrients, and food webs, asopposed to through oxygen concentrationsthemselves (e.g., Anbar and Knoll 2002).Examples of these effects can be found bylooking at experiments on metal utilizationamong modern eukaryotes as well as examin-ing genomic information to determine themetal co factor needs of ancient eukaryoticgroups. For example, work by Dupont andcolleagues (2010) indicates that eukaryotes userelatively late evolving proteins for Zn, Ca, andFe utilization as compared to akaryoterelatives. Eukaryotes depend on Zn for a vari-ety of protein functions, and Zn is less bioavailable in the low oxygen conditions, and isthus presumed to have been low early in theProterozoic (Saito et al. 2003). Thus, increasinglevels of atmospheric oxygen would haveenabled eukaryotes to better access key bioelements. While the picture of redox changes inglobal oceans during the Proterozoic is not yetclear, if a rise in oxygen facilitated eukaryoticdiversification, it likely did so through alimited rise past critical ecological thresholds(Sperling et al. 2015).Evolution of Sinking, Carbon Burial, and the

Carbon Cycle.—One interesting implication forthe temporal distribution of resistant taxa(Fig. 14) is the role that such resistant

structures may play in increasing the rate ofsinking of organic matter to the seafloor.Tziperman et al. (2011) proposed that theradiation of eukaryotes in the Neoproterozoiccould be partially responsible for fluctuationsthe carbon cycle and the initiation of SnowballEarth (Tziperman et al. 2011). Specifically,they implicate biomineralized taxa and largereukaryotic cell sizes as a potential triggerfor increased sinking, complementing thehypotheses discussed above that the rifting ofRodinia, increased basin formation, weathering,and phosphorous delivery to the ocean allcould have led to increased organic carbonexport, burial, and anaerobic remineralization.Ultimately, the long term oxygenation of theoceans must have been driven by the removal ofreduced carbon from the ocean atmospheresystem into the sedimentary reservoir, andwhile increased basin formation with therifting of Rodinia provided the backdrop forthis change, organic carbon sequestrationwas likely aided by increased productivityand sinking. Thus, the record of increaseddiversification and the timing of recalcitranttests and scales documented here supportsthe hypothesis of positive feedbacks betweentectonics, evolutionary innovations and changingoceanographic and geochemical conditions.

Conclusions

Our new analysis of the Proterozoic eukar-yotic fossil record indicates that within assem-blage diversity rose through the Proterozoic,especially during the late Tonian and Edia-caran. The overall rise documented here is duein large part to an increase in the number ofhigh diversity assemblages; low diversityassemblages remain common throughout theProterozoic. Diversity is low in the Cryogeniannonglacial interlude, likely due to a combina-tion of biological factors coupled with rockrecord and sampling biases. In general, litho-logical, taphonomic, and sampling biases dopersist and can influence our view of the tempoand mode of eukaryotic fossil diversity.Despite these biases, our results indicate thateukaryotic ecosystems became more complexand diverse through the Proterozoic. In addi-tion, this re-analysis sheds light on the intricate


relationship between biotic and abiotic eventsin the Proterozoic, including changing redoxconditions, rising oxygen levels, glacialepisodes, and supercontinent breakup, and hintsat complex feedbacks between evolution of lifeand the environment. A closer assessment of theoverall trends as well as the biases inherent inthe Proterozoic fossil record also allows us tomove forwardwith a clearer picture of samplingstrategies. This analysis provides a frameworkinto which future paleontological and geobiolo-gical sampling and research can be placed,allowing for a more holistic view of ProterozoicEarth system dynamics.

Acknowledgements

We thank S. Finnegan, B. Heggeseth,B. Kotrc, J. McKay, and S. Peters for technicalassistance and feedback. This paper wasimproved by reviews from S. Xiao andJ. Huntley. Support was provided by aNational Aeronautics and Space Administra-tion Astrobiology Institute grant through theMassachusetts Institute of Technology andWilliams College.

Literature CitedAllison, C. W., and J. W. Hilgert. 1986. Scale microfossils from theEarly Cambrian of northwest Canada. Journal of Paleontology60:973–1015.

Alroy, J., M. Aberhan, D. J. Bottjer, M. Foote, F. T. Fürsich,P. J. Harries, A. J. W. Hendy, S. M. Holland, L. C. Ivany,W. Kiessling, M. A. Kosnik, C. R. Marshall, A. J. McGowan,A. I. Miller, T. D. Olszewski, M. E. Patzkowsky, S. E. Peters,L. Villier, P. J. Wagner, N. Bonuso, P. S. Borkow, B. Brenneis,M. E. Clapham, L. M. Fall, C. A. Ferguson, V. L. Hanson,A. Z. Krug, K. M. Layou, E. H. Leckey, S. Nürnberg,C. M. Powers, J. A. Sessa, C. Simpson, A. Tomašových, andC. C. Visaggi. 2008. Phanerozoic Trends in the Global Diversityof Marine Invertebrates. Science 321:97–100.

Anbar, A. D., and A.H. Knoll. 2002. Proterozoic Ocean Chemistryand Evolution: A Bioinorganic Bridge? Science 297:1137–1142.

Bambach, R. K. 1977. Species richness in marine benthic habitatsthrough the Phanerozoic. Paleobiology 3:152–167.

Berner, R. A. 1973. Phosphate removal from sea water by adsorp-tion on volcanogenic ferric oxides. Earth and Planetary ScienceLetters 18:77–86.

Boenigk, J., K. Pfandl, T. Garstecki, H. Harms, G. Novarino, andA. Chatzinotas. 2006. Evidence for Geographic Isolation andSigns of Endemism within a Protistan Morphospecies. Appliedand Environmental Microbiology 72:5159–5164.

Bosak, T., D. J. G. Lahr, S. B. Pruss, F. A. Macdonald, A. J. Gooday,L. Dalton, and E. D. Matys. 2011a. Possible early foraminiferansin post-Sturtian (716-635 Ma) cap carbonates. Geology 40:67–70.

Bosak, T., D. J. G. Lahr, S. B. Pruss, F. A. Macdonald, L. Dalton, andE. Matys. 2011b. Agglutinated tests in post-Sturtian cap

carbonates of Namibia and Mongolia. Earth and PlanetaryScience Letters 308:29–40.

Bosak, T., F. Macdonald, D. Lahr, and E. Matys. 2011c. PutativeCryogenian ciliates from Mongolia. Geology 39:1123–1126.

Bradley, D. C. 2008. Passive margins through earth history. EarthScience Reviews 91:1–26.

Butterfield, N. J. 2004. A vaucheriacean alga from the middleNeoproterozoic of Spitsbergen: implications for the evolution ofProterozoic eukaryotes and the Cambrian explosion. Paleobio-logy 30:231–252.

——. 2005. Probable proterozoic fungi. Paleobiology 31:165–182.——. 2009. Modes of pre-Ediacaran multicellularity. PrecambrianResearch 173:1–11.

——. 2000. Bangiomorpha pubescens n. gen., n. sp.: implications forthe evolution of sex, multicellularity, and the Mesoproterozoic/Neoproterozoic radiation of eukaryotes. Journal of Paleontology26:386–404.

Butterfield, N. J., A. H. Knoll, and K. Swett. 1994. Paleobiology ofthe Neoproterozoic Svanbergfjellet Formation, Spitsbergen.Lethaia 27:76–76.

Calver, C. R., K. Grey, and M. Laan. 2010. The “string of beads”fossil (Horodyskia) in the mid-Proterozoic of Tasmania.Precambrian Research 180:18–25.

Casteleyn, G., F. Leliaert, T. Backeljau, A.-E. Debeer, Y. Kotaki,L. Rhodes, N. Lundholm, K. Sabbe, andW. Vyverman. 2010. Limitsto gene flow in a cosmopolitan marine planktonic diatom. Proceed-ings of the National Academy of Sciences USA 107:12952–12957.

Chase, J. M., and M. A. Leibold. 2002. Spatial scale dictates theproductivity–biodiversity relationship. Nature 416:427–430.

Cloud, P. E. 1968. Atmospheric and Hydrospheric Evolution on thePrimitive Earth Both secular accretion and biological andgeochemical processes have affected earth's volatile envelope.Science 160:729–736.

Cohen, P. A., A. H. Knoll, and R. B. Kodner. 2009. Large spinosemicrofossils in Ediacaran rocks as resting stages of early animals.Proceedings of the National Academy of Sciences 106:6519–6524.

Cohen, P. A., and A. H. Knoll. 2012. Scale Microfossils from theMid-Neoproterozoic Fifteenmile Group, Yukon Territory.Journal of Paleontology 86:775–800.

Cohen, P. A., J. W. Schopf, N. J. Butterfield, A. B. Kudryavtsev, andF. A. Macdonald. 2011. Phosphate biomineralization inmid-Neoproterozoic protists. Geology 39:539–542.

Cohen, P. A., F.A. Macdonald, S. Pruss, E. Matys, and T. Bosak.2015. Fossils of putative marine algae from the Cryogenianglacial interlude of Mongolia. Palaios 30:238–247.

Condon, D., M. Zhu, S. Bowring, W. Wang, A. Yang, and Y. Jin.2005. U-Pb ages from the neoproterozoic DoushantuoFormation, China. Science 308:95–98.

Corsetti, F. A., S. M. Awramik, and D. Pierce. 2003. A complexmicrobiota from snowball Earth times: microfossils from theNeoproterozoic Kingston Peak Formation, Death Valley, USA.Proceedings of the National Academy of Sciences 100:4399–4404.

Creveling, J. R., D. T. Johnston, S. W. Poulton, B. Kotrc, C. März,D. P. Schrag, and A. H. Knoll. 2013. Phosphorus sources forphosphatic Cambrian carbonates. Geological Society of AmericaBulletin 126:145–163.

Dalton, T. Bosak, F. A. Macdonald, D. J. G. Lahr, and S. B. Pruss.2013. Preservational and Morphological Variability ofAssemblages of Agglutinated Eukaryotes in Cryogenian CapCarbonates of Northern Namibia. Palaios 28:67–79.

Dalziel, I. W. D. 1997. Neoproterozoic-Paleozoic geography andtectonics: Review, hypothesis, environmental speculation. Geo-logical Society of America Bulletin 109:16–42.

Dupont, C. L., A. Butcher, R. E. Valas, P. E. Bourne, andG. Caetano-Anollés. 2010. History of biological metal utilization inferredthrough phylogenomic analysis of protein structures. Proceed-ings of the National Academy of Sciences 107:10567–10572.


Ernst, R. E., M. Wingate, K. L. Buchan, and Z. X. Li. 2008. Globalrecord of 1600–700Ma Large Igneous Provinces (LIPs): implica-tions for the reconstruction of the proposed Nuna (Columbia)and Rodinia supercontinents. Precambrian Research 160:159–178.

Erwin, D. H., M. Laflamme, S. M. Tweedt, E. A. Sperling, D. Pisani,and K. J. Peterson. 2011. The Cambrian Conundrum: EarlyDivergence and Later Ecological Success in the Early History ofAnimals. Science 334:1091–1097.

Evans, D. A. D. 2013. Reconstructing pre-Pangean supercontinents.Geological Society of America Bulletin 125:1735–1751.

Fenchel, T. 1968. The ecology of marine microbenthos II. The foodof marine benthic ciliates. Ophelia 5:73–121.

Foissner, W., A. Chao, and L. A. Katz. 2007. Diversity andgeographic distribution of ciliates (Protista: Ciliophora). Biodi-versity and Conservation 17:345–363.

Halverson, G. P., B. P. Wade, M. T. Hurtgen, and K. M. Barovich.2010. Neoproterozoic chemostratigraphy. Precambrian Research182:337–350.

Halverson, G. P., G. Cox, M. T. Hurtgen, P. Sansjofre,M. Kunzmann, J. V. Strauss, and F. A Macdonald. 2014. Acontinental flood basalt driver for Neoproterozoic climate andoxygenation. Geological Society of America Abstracts withPrograms, 256–256.

Han, B.-P., T. Wang, Q.-Q. Lin, and H. J. Dumont. 2007. Carnivoryand active hunting by the planktonic testate amoeba Difflugiatuberspinifera. Hydrobiologia 596:197–201.

Heinz, P., and E. Geslin. 2012. Ecological and Biological Responseof Benthic Foraminifera Under Oxygen-Depleted Conditions:Evidence from Laboratory Approaches. Pp. 287–303 inA. V. Altenbach, J. M. Bernhard, and J. Seckbach, eds. AnoxiaVol. 21. Springer, Netherlands.

Hoffman, P. F. 1998. A Neoproterozoic Snowball Earth. Science281:1342–1346.

Huntley, J. W., and M. Kowalewski. 2007. Strong coupling ofpredation intensity and diversity in the Phanerozoic fossilrecord. Proceedings of the National Academy of Sciences104:15006–15010.

Huntley, J. W., S. Xiao, and M. Kowalewski. 2006. 1.3 Billion yearsof acritarch history: An empirical morphospace approach.Precambrian Research 144:52–68.

Javaux, E. J., A. H. Knoll, and M. R. Walter. 2004. TEM evidence foreukaryotic diversity in mid-Proterozoic oceans. Geobiology2:121–132.

Javaux, E. J., A. H. Knoll, and M. Walter. 2003. Recognizing andinterpreting the fossils of early eukaryotes. Origins of Life andEvolution of the Biosphere 33:75–94.

Johnston, D. T., S. W. Poulton, C. Dehler, S. Porter, J. Husson,D. E. Canfield, and A. H. Knoll. 2010. An emerging picture ofNeoproterozoic ocean chemistry: Insights from the Chuar Group,Grand Canyon, USA. Earth and Planetary Science Letters 290:64–73.

Johnston, D. T., S. W. Poulton, N. J. Tosca, T. O'Brien, G. P. Hal-verson, D. P. Schrag, and F. A. Macdonald. 2013. Search for anoxygenation event in the fossiliferous Ediacaran ofNorthwest Canada. Chemical Geology 362:273–286.

Kah, L. C., and J. K. Bartley. 2011. Protracted oxygenation of theProterozoic biosphere. International Geology Review 53:1424–1442.

Kaufman, A. J., A. H. Knoll, and S. M. Awramik. 1992. Biostrati-graphic and chemostratigraphic correlation of Neoproterozoicsedimentary successions: Upper Tindir Group, northwesternCanada, as a test case. Geology 20:181–185.

Knoll, A. H. 1999. Early Animal Evolution: Emerging Views fromComparative Biology and Geology. Science 284:2129–2137.

——. 2003. Biomineralization and evolutionary history. Reviews inMineralogy and Geochemistry 54:329–356.

Knoll, A. H., and E. A. Sperling. 2014. Oxygen and animals in Earthhistory. Proceedings of the National Academy of Sciences111:3907–3908.

Knoll, A. H., E. J. Javaux, D. Hewitt, and P. Cohen. 2006.Eukaryotic organisms in Proterozoic oceans. PhilosophicalTransactions of the Royal Society B: Biological Sciences 361:1023–1038.

Kowalewski, M., and P. M. Novack-Gottshall. 2010. Resamplingmethods in paleontology. Pp. 19–54 in J. Alroy, and G. Hunt, eds.Quantitative Methods in Paleobiology Vol. 16. The Paleontolo-gical Society, Chicago.

Laakso, T. A., and D. P. Schrag. 2014. Regulation of atmosphericoxygen during the Proterozoic. Earth and Planetary ScienceLetters 388:81–91.

Lenton, T. M., R. A. Boyle, S. W. Poulton, G. A. Shields-Zhou, andN. J. Butterfield. 2014. Co-evolution of eukaryotes and oceanoxygenation in the Neoproterozoic era. Nature Geoscience 7:257–265.

Li, Z. X., S. V. Bogdanova, A. S. Collins, A. Davidson, B. De Waele,R. E. Ernst, I. C. W. Fitzsimons, R. A. Fuck, D. P. Gladkochub,J. Jacobs, K. E. Karlstrom, S. Lu, L. M. Natapov, V. Pease, S. A.Pisarevsky, K. Thrane, and V. Vernikovsky. 2008. Assembly,configuration, and break-up history of Rodinia: A synthesis.Precambrian Research 160:179–210.

Li, Z. X., X. H. Li, P. D. Kinny, and J. Wang. 1999. Thebreakup of Rodinia: did it start with a mantle plumebeneath South China? Earth and Planetary Science Letters173:171–181.

Li, Z.-X., D. A. D. Evans, and G. Halverson. 2013. Neoproterozoicglaciations in a revised global palaeogeography from thebreakup of Rodinia to the assembly of Gondwanaland. Sedi-mentary Geology, 1–63.

Love, G. D., E. Grosjean, C. Stalvies, D. A. Fike, J. P. Grotzinger,A. S. Bradley, A. E. Kelly, M. Bhatia, W. Meredith, C. E. Snape,S. A. Bowring, D. J. Condon, and R. E. Summons. 2008. Fossilsteroids record the appearance of Demospongiae during theCryogenian period. Nature 457:718–721.

Lyons, T. W., C. T. Reinhard, and N. J. Planavsky. 2014. The rise ofoxygen in Earth’s earlyocean and atmosphere. Nature 506:307–315.

Macdonald, F. A., A. R. Prave, R. Petterson, E. F. Smith, S. B. Pruss,K. Oates, F. Waechter, D. Trotzuk, and A. E. Fallick. 2013. TheLaurentian record of Neoproterozoic glaciation, tectonism, andeukaryotic evolution in Death Valley, California. GeologicalSociety of America Bulletin 125:1203–1223.

Macdonald, F. A., and P. A. Cohen. 2011. Chapter 35 The Tatondukinlier, Alaska-Yukon border. Geological Society, London, Mem-oirs 36:389–396.

Macdonald, F. A., M. D. Schmitz, J. L. Crowley, C. F. Roots, D. S.Jones, A. C. Maloof, J. V. Strauss, P. A. Cohen, D. T. Johnston, andD. P. Schrag. 2010b. Calibrating the Cryogenian. Science327:1241–1243.

Maliva, R. G., A. H. Knoll, and R. Siever. 1989. Secular change inchert distribution: a reflection of evolving biological participationin the silica cycle. Palaios 4:519–532.

Maloof, A. C., C. V. Rose, R. Beach, B. M. Samuels, C. C. Calmet,D. H. Erwin, G. R. Poirier, N. Yao, and F. J. Simons. 2010. Possibleanimal-body fossils in pre-Marinoan limestones from SouthAustralia. Nature 3:653–659.

Marshall, C. R. 2006. Explaining the Cambrian “explosion” ofanimals. Annual Review of Earth and Planetary Sciences 34:355–384.

Mills, D. B., L. M. Ward, C. Jones, B. Sweeten, M. Forth,A. H. Treusch, and D. E. Canfield. 2014. Oxygen requirements ofthe earliest animals. Proceedings of the National Academy ofSciences USA 111:4168–4172.


Moczydłowska, M. 2005. Taxonomic review of some Ediacaranacritarchs from the Siberian Platform. Precambrian Research136:283–307.

Moczydłowska, M., J. W. Schopf, and S. Willman. 2009. Micro- andnano-scale ultrastructure of cell walls in Cryogenianmicrofossils:revealing their biological affinity. Lethaia 43:129–136.

Och, L. M., and G. A. Shields-Zhou. 2012. The Neoproterozoicoxygenation event: Environmental perturbations and biogeo-chemical cycling. Earth Science Reviews 110:26–57.

Parfrey, L. W., D. J. Lahr, A. H. Knoll, and L. A. Katz. 2011.Estimating the timing of early eukaryotic diversification withmultigene molecular clocks. Proceedings of the NationalAcademy of Sciences 108:13624–13629.

Partin, C. A., A. Bekker, N. J. Planavsky, C. T. Scott, B. C. Gill, C. Li,V. Podkovyrov, A. Maslov, K. O. Konhauser, S. V. Lalonde, G. D.Love, S. W. Poulton, and T. W. Lyons. 2013. Large-scale fluc-tuations in Precambrian atmospheric and oceanic oxygen levelsfrom the record of U in shales. Earth and Planetary Science Let-ters 369–370:284–293.

Peters, S. E., and N. A. Heim. 2010. The geological completeness ofpaleontological sampling in North America. Paleobiology 36:61–79.

Peterson, K. J., J. B. Lyons, K. S. Nowak, C. M. Takacs, M. J. Wargo,and M. A. McPeek. 2004. Estimating metazoan divergence timeswith a molecular clock. Proceedings of the National Academy ofSciences 101:6536–6541.

Planavsky, N. J., C. T. Reinhard, X. Wang, D. Thomson,P. McGoldrick, R. H. Rainbird, T. Johnson, W. W. Fischer, andT. W. Lyons. 2014. Low Mid-Proterozoic atmospheric oxygenlevels and the delayed rise of animals. Science 346:635–638.

Planavsky, N. J., O. J. Rouxel, A. Bekker, S. V. Lalonde,K. O. Konhauser, C. T. Reinhard, and T. W. Lyons. 2010. Theevolution of the marine phosphate reservoir. Nature 467:1088–1090.

Porter, S. 2011. The rise of predators. Geology 39:607–608.Porter, S. M. 2004. Closing the phosphatization window: testing forthe influence of taphonomic megabias on the pattern of smallshelly fossil decline. Palaios 19:178–183.

Porter, S. M., and A. H. Knoll. 2000. Testate amoebae in the Neo-proterozoic Era: evidence from vase-shaped microfossils in theChuar Group, Grand Canyon. Paleobiology 26:360–385.

Poulton, S. W., and D. E. Canfield. 2011. Ferruginous Conditions: ADominant Feature of the Ocean through Earth's History. Ele-ments 7:107–112.

——. 2006. Co-diagenesis of iron and phosphorus in hydrothermalsediments from the southern East Pacific Rise: Implications forthe evaluation of paleoseawater phosphate concentrations.Geochimica et Cosmochimica Acta 70:5883–5898.

Pruss, S. B., T. Bosak, F. A. Macdonald, M. McLane, and P. F.Hoffman. 2010. Microbial facies in a Sturtian cap carbonate, theRasthof Formation, Otavi Group, northern Namibia. Pre-cambrian Research 181:187–198.

Rhoads, D. C., and J. W. Morse. 1971. Evolutionary and ecologicsignificance of oxygen-deficient marine basins. Lethaia 4:413–428.

Riedman, L. A., S. M. Porter, G. P. Halverson, M. T. Hurtgen, andC. K. Junium. 2014. Organic-walled microfossil assemblages fromglacial and interglacial Neoproterozoic units of Australia andSvalbard. Geology 42:1011–1014.

Roger, A. J., and L. A. Hug. 2006. The origin and diversification ofeukaryotes: problems with molecular phylogenetics and mole-cular clock estimation. Philosophical Transactions of the RoyalSociety B: Biological Sciences 361:1039–1054.

Ronov, A. B. 1982. The Earth's sedimentary shell (quantitativepatterns of its structure, compositions, and evolution) The20th VI Vernadskiy Lecture, March 12, 1978. InternationalGeology Review 24:1313–1363.

Rooney, A. D., F. A. Macdonald, J. V. Strauss, F. Ö. Dudás,C. Hallmann, and D. Selby. 2014. Re-Os geochronology andcoupled Os-Sr isotope constraints on the Sturtian snowball Earth.Proceedings of the National Academy of Sciences 111:51–56.

Rooney, A. D., J. V. Strauss, A. D. Brandon, and F. A. Macdonald.2015. A Cryogenian chronology: Two long-lasting synchronousNeoproterozoic glaciations. Geology 43:459–462.

Sahoo, S. K., N. J. Planavsky, B. Kendall, X. Wang, X. Shi, C. Scott,A. D. Anbar, T. W. Lyons, and G. Jiang. 2013. Ocean oxygenationin the wake of the Marinoan glaciation. Nature 488:546–549.

Saito, M. A., D. M. Sigman, and F. M. M. Morel. 2003. Thebioinorganic chemistry of the ancient ocean: the co-evolution ofcyanobacterial metal requirements and biogeochemical cycles atthe Archean–Proterozoic boundary? Inorganica Chimica Acta356:308–318.

Sayre, R. M. 1973. Theratromyxa weberi, An Amoeba Predatory onPlant-Parasitic Nematodes. Journal of Nematology 5:259–264.

Schrag, D. P., J. A. Higgins, F. A. Macdonald, and D. T. Johnston.2013. Authigenic Carbonate and the History of the GlobalCarbon Cycle. Science 339:540–543.

Scott, C., T. W. Lyons, A. Bekker, Y. Shen, S. W. Poulton, X. Chu,and A. D. Anbar. 2008. Tracing the stepwise oxygenation of theProterozoic ocean. Nature 452:456–459.

Sharma, M., and Y. Shukla. 2009. Taxonomy and affinity of EarlyMesoproterozoic megascopic helically coiled and related fossilsfrom the Rohtas Formation, the Vindhyan Supergroup, India.Precambrian Research 173:105–122.

Sperling, E. A., C. J. Wolock, A. S. Morgan, B. C. Gill,M. Kunzmann, G. P. Halverson, F. A. Macdonald, A. H. Knoll,and D. T. Johnston. 2015. Statistical analysis of iron geochemicaldata suggests limited Late Proterozoic. Nature 523:451–454.

Sperling, E. A., G. P. Halverson, A. H. Knoll, F. A. Macdonald, andD. T. Johnston. 2013. A basin redox transect at the dawn ofanimal life. Earth and Planetary Science Letters 371:143–155.

Sperling, E. A., J. M. Robinson, D. Pisani, and K. J. Peterson. 2009.Where's the glass? Biomarkers, molecular clocks, and micro-RNAs suggest a 200-Myr missing Precambrian fossil record ofsiliceous sponge spicules. Geobiology 8:24–36.

Squire, R. J., I. H. Campbell, C. M. Allen, and C. J. L. Wilson. 2006.The Transgondwanan Supermountain: A trigger for the Cam-brian explosion. Geochimica et Cosmochimica Acta 70:A608.

Srivastava, P. 2012. Problematic fossils from the Palaeo-Neoproterozoic Vindhyan Supergroup, India. Arabian Journalof Geosciences 5:1411–1422.

Stanley, S. M. 1973. An ecological theory for the sudden origin ofmulticellular life in the late Precambrian. Proceedings of theNational Academy of Sciences 70:1486–1489.

Strauss, J. V., A. D. Rooney, F. A. Macdonald, A. D. Brandon, andA. H. Knoll. 2014. 740 Ma vase-shaped microfossils from Yukon,Canada: Implications for Neoproterozoic chronology andbiostratigraphy. Geology 42:659–662.

Turner, E. C., and B. S. Kamber. 2012. Arctic Bay Formation, BordenBasin, Nunavut (Canada): Basin evolution, black shale, anddissolved metal systematics in the Mesoproterozoic ocean.Precambrian Research 208:1–18.

Tziperman, E., I. Halevy, D. T. Johnston, A. H. Knoll, and D. P.Schrag. 2011. Biologically induced initiation of Neoproterozoicsnowball-Earth events. Proceedings of the National academy ofSciences 108:15091–15096.

Valentine, J. W., and E. M. Moores. 1970. Plate-tectonicregulation of faunal diversity and sea level: a model. Nature228:657–659.

Vermeij, G. J. 1977. The Mesozoic marine revolution: evidence fromsnails, predators and grazers. Paleobiology, 245–258.

Vidal, G., and A. H. Knoll. 1983. Proterozoic plankton. GeologicalSociety of America Memoir 161:265–277.


Vidal, G., and M. Moczydłowska-Vidal. 1997. Biodiversity, spe-ciation, and extinction trends of Proterozoic and Cambrian phy-toplankton. Paleobiology, 230–246.

Vidal, G., M. Moczydtowska, and V. A. Rudavskaya. 1993.Biostratigraphical implications of a Chuaria-Tavuia assemblageand associated acritarchs from the Neoproterozoic of Yakutia.Palaeontology 36:387–387.

Wang, X.-C., X.-H. Li, W.-X. Li, Z.-X. Li, Y. Liu, Y.-H. Yang,X.-R. Liang, andX.-L. Tu. 2008. The Bikoubasalts in the northwesternYangtze block, South China: Remnants of 820–810 Ma continentalflood basalts? Geological Society of America Bulletin 120:1478–1492.

Wang, X.-C., Z.-X. Li, X.-H. Li, Q.-L. Li, and Q.-R. Zhang.2011. Geochemical and Hf–Nd isotope data of Nanhua rift sedi-mentary and volcaniclastic rocks indicate a Neoproterozoiccontinental flood basalt provenance. Lithos 127:427–440.

Wingate, M. T., I. H. Campbell, W. Compston, and G. M. Gibson.1998. Ion microprobe U–Pb ages for Neoproterozoicbasaltic magmatism in south-central Australia andimplications for the breakup of Rodinia. Precambrian Research87:135–159.

Xiao, S., and L. Dong. 2006. On the morphological and ecologicalhistory of Proterozoic macroalgae. Pp. 57–90 in S. Xiao, and A.J.Kaufman, eds. Neoproterozoic Geobiology and Paleobiology.Springer, Netherlands.

Yin, L., M. Zhu, A. H. Knoll, X. Yuan, J. Zhang, and J. Hu. 2007.Doushantuo embryos preserved inside diapause egg cysts. Nat-ure 446:661–663.

Zhou, C., R. Tucker, S. Xiao, Z. Peng, X. Yuan, and Z. Chen. 2004.New constraints on the ages of Neoproterozoic glaciations insouth China. Geology 32:437–440.


Documents

The Proterozoic Record of Eukaryotes Phoebe A. Cohen and … · 2015-09-15 · biases inherent in the Proterozoic fossil record, including issues of rock record availability, poor