Evolution amp Development 2019e12312 wileyonlinelibrarycomjournalede copy 2019 Wiley Periodicals Inc | 1 of 15httpsdoiorg101111ede12312

Received 28 March 2019 | Revised 5 August 2019 | Accepted 6 August 2019

DOI 101111ede12312

RE S EARCH

Developmental bias in the fossil record

Illiam S C Jackson

Department of Biology Lund UniversitySweden

CorrespondenceIlliam S C JacksonLund University SwedenEmail illiamjacksongmailcom

Funding informationKnut och Alice Wallenbergs StiftelseGrantAward Number KAW 20120155

Abstract

The role of developmental bias and plasticity in evolution is a central research

interest in evolutionary biology Studies of these concepts and related

processes are usually conducted on extant systems and have seen limited

investigation in the fossil record Here I identify plasticity‐led evolution (PLE)

as a form of developmental bias accessible through scrutiny of paleontological

material I summarize the process of PLE and describe it in terms of the

environmentally mediated accumulation and release of cryptic genetic

variation Given this structure I then predict its manifestation in the fossil

record discuss its similarity to quantum evolution and punctuated equili-

brium and argue that these describe macroevolutionary patterns concordant

with PLE Finally I suggest methods and directions towards providing

evidence of PLE in the fossil record and conclude that such endeavors are

likely to be highly rewarding

KEYWORD S

cryptic genetic variation developmental bias plasticity‐led evolution

1 | INTRODUCTION

Developmental bias describes the manner in whichcertain changes to development are more accessible toevolution than others (Arthur 2004) A key generator ofdevelopmental bias and integral to understanding itsrole in evolution is developmental plasticity whichrefers to an organismrsquos ability to adjust its phenotype inresponse to environmental conditions (Laland et al2015 Schwab Casasa amp Moczek 2019) Developmentalbias framed as a form of constraint is an importanttopic studied through the fossil record but the biasimposed by plasticity has not received similar attentionInvestigating plasticity‐induced developmental bias(ldquoplasticity‐led evolutionrdquo PLE Levis amp Pfennig 2016)in paleontological data sets necessitates identifying itstraces in the fossil record

In this paper I outline the process of PLE and corerelated concepts and discuss how it relates to develop-mental bias Thereafter I predict what kind of patterns

we might expect to result from PLE in the fossil recordanchor these patterns in historical work and suggestmethods and future directions for examining PLE usingfossil data

2 | CONSTRUCTIVEDEVELOPMENT

Central to understanding the process of PLE is theconceptualization of development as constructive ratherthan programmed (Laland et al 2015) Development hasbeen traditionally viewed as a programmed processcaptured in the ldquogenes as blueprintrdquo metaphor (Lalandet al 2015 Moczek 2012 Noble 2015 Pigliucci 2010)perhaps best exemplified by Dawkins (1976) in TheSelfish Gene where genes are described as ldquosafe insidegigantic lumbering robots sealed off from the outsideworld communicating with it by tortuous indirect routesand manipulating it by remote controlrdquo This view is

starkly at odds with the framing of development asconstructive capable of ldquo[shaping] its own developmen-tal trajectory by constantly responding to and alteringinternal and external statesrdquo (Laland et al 2015)Informed by the growing field of evo‐devo and con-temporary understanding of concepts such as develop-mental plasticity the relationship between environmentand development is increasingly understood as morecomplex than a unidirectional programmed processwhich naturally has ramifications for how we understandevolution (Brakefield 2011 Moczek 2015 2015 Muumlller2007 West‐Eberhard 2003)

For example the gene regulatory network structure ofdevelopment (Carroll 2005 Davidson 2006 Muumlller2007 Carroll 2008 Erwin amp Davidson 2009 UllerMoczek Watson Brakefield amp Laland 2018) facilitatesthe generation of phenotypic variation by reducing thenumber of regulatory changes required to achieveevolutionary change (Gerhart amp Kirschner 2007Wagner 2011) Given that these regulatory systems cantransition between genetic and environmental control(Iwasaki Tsuda amp Kawata 2013 Schneider amp Meyer2017 Schwander amp Leimar 2011 West‐Eberhard 2003)it follows that environmental triggers through theirpresence or absence can act to suppress developmentalpathways shielding them from natural selection andliberating them to accumulate change without malus tothe organism (Kuumlttner et al 2014 Moczek 2008Rutherford 2000 Schlichting amp Smith 2002 ZhengPayne amp Wagner 2019) Such unexpressed variation isknown as cryptic genetic variation (CGV) and can beunderstood as a form of evolutionary capacitance thatmay play an important role in evolution (Gibson ampDworkin 2004 Levis amp Pfennig 2016 McGuigan amp Sgrograve2009 Moczek 2008 Paaby amp Rockman 2014 Schlicht-ing 2008 Schwab et al 2019)

3 | THE ROLE OF CRYPTICGENETIC VARIATION INPLASTICITY ‐LED EVOLUTION

CGV may but need not necessarily constitute anadaptive trait Rather it can be understood as randomlyaccumulated variation on a developmental systemallowed to persist due to its nonexpression (Kuumlttneret al 2014 Moczek 2008 Rutherford 2000 Schlichtingamp Smith 2002 Zheng et al 2019)Given that this cryptic variation may be sensitive toenvironmental triggers it acts as a source of innovationthat can be episodically induced and presented toselection (Gibson amp Dworkin 2004 Levis amp Pfennig2016 McGuigan amp Sgrograve 2009 Moczek 2008 Paaby amp

Rockman 2014 Schlichting 2008 Schwab et al 2019Zheng et al 2019) It is this form of plasticity notadaptive plasticity that plays the central role of drivingthe process of PLE as described below and discussedthroughout this paper In other words the plasticityleading evolution in this model of PLE is not an evolvedresponse but rather the periodical accumulation andplastic release of CGV sensu Landersquos (2009) ldquotransientevolution of increased plasticityrdquo

If development indeed operates in this constructivemanner then what kind of consequent patterns would weanticipate A growing body of work is exploring theimplications of constructive development and develop-mental plasticity in extant systems (Levis amp Pfennig2019 van Bergen et al 2017) but research from apaleontological perspective has so far been limited Ifcryptic genetic variation drives the process of plasticity‐led evolution then what kind of traces would we expect todetect in the fossil record

4 | PATTERNS AND PROCESSESIN THE FOSSIL RECORD

The fossil record is the history book of life preservingpatterns of evolutionary change through time Theserecorded patterns are regarded as macroevolutionarythat is broad‐scale at the level of species and above(Gould 2002 Myers amp Saupe 2013) This is the classicalunderstanding of macroevolution as a counterpart tomicroevolution which describes processes governingevolutionary change below a species level (Dobzhansky1937 Simpson 1944) Although macroevolutionarypatterns are generally assumed to be generated bymicroevolutionary processes over time (see discussionin Erwin 2017) the fossil record is more commonlyused to map out evolutionary patterns at the level ofspecies and above than to infer microevolutionaryprocesses But see for example Organ Copper andHieronymus (2015) and Jablonski and Shubin (2015) foran overview of work in such a context This is due inpart to the historical conceptual dichotomy betweenmicro‐ and macroevolution but also to the kinds of dataand methods available for paleontological studies towhich I will return further on

Obviously in order for microevolutionary processes tobe detectable in the fossil record they must necessarily betheoretically capable of generating a morphologicalsignature that is relevant in a macroevolutionary time-frame To determine what such a signature of PLE mightlook like I first outline the process before going on todraw up strategies for identifying it

2 of 15 | JACKSON

5 | PLASTICITY ‐LEDEVOLUTION

PLE also known as the plasticity‐first hypothesisdescribes a mode of evolution in which the environmentplays a pivotal role in both inducing and subsequentlyselecting for phenotypic variation (Levis amp Pfennig 2016West‐Eberhard 2003 2005) It has been investigated inextant organisms such as sticklebacks (Wund ValenaWood amp Baker 2012) house finches (Badyaev 2009)and spadefoot toads (Levis amp Pfennig 2019) yet how wemight approach this process in a paleontological contextremains unknown Here follows a summary of PLE alsoillustrated in Figure 1

A population of organisms well‐adapted to itsenvironment is subject to the process of stabilizing

selection which selects not only for a specific pheno-typic optimum but also selects against variation aroundthis optimum (Figure 1a Schmalhausen 1949) Apartfrom the removal of unfit variants this is achieved byreducing developmental sensitivity to any perturbationincluding environmental conditions (Gibson amp Wagner2000 Pertoldi Bundgaard Loeschcke amp Barker 2014Waddington 1942 Wagner Booth amp Bagheri‐Chai-chian 1997) When development is canalized in thismanner CGV may accumulate in shielded parts of thegenome (Kuumlttner et al 2014 Moczek 2008 Rutherford2000 Schlichting amp Smith 2002 Zheng et al 2019)generating evolutionary capacitance (Figure 1b Gibsonamp Dworkin 2004 Levis amp Pfennig 2016 McGuigan ampSgrograve 2009 Paaby amp Rockman 2014 Schlichting 2008Schwab et al 2019)

yg

olo

hpr

oM

Selective

gradient

Mo

rph

olo

gy

Mo

rph

olo

gy

Selective

gradient

yg

olo

hpr

oM

Morphology

Morphology

Morphology

Morphology

(a) (b)

(c)(d)

FIGURE 1 Illustration of the process of plasticity‐led evolution in terms of phenotype stress and cryptic genetic variation (CGV)Ellipses represent populations in morphospace The shading intensity of the ellipses gives their relative level of CGV (a) Stabilizing selectionacting on a population selects both for a phenotypic optimum as well as reduced variation around that optimum (b) Phenotypic variation isreduced through both the removal of unfit variants as well as decreasing sensitivity to environmental conditions that is canalization ofdevelopment This shields previously expressed but now hidden developmental pathways from natural selection such that CGV that isevolutionary capacitance can accumulate (c) An environmental stress is encountered that both induces as well as selects for variation CGVis revealed by this environmental trigger and phenotypic accommodation occurs generating increasing phenotypic variability This can beunderstood as the expenditure of the previously accumulated evolutionary capacitance (d) Selection acts on the uncovered variation byselecting for the new environmental optimum and genetically accommodating such change that was beneficial thus restarting this processby undergoing stabilizing selection again and recanalizing development once more

JACKSON | 3 of 15

Novel or extreme environmental conditions mayreveal this capacitance (Badyaev 2005 Flatt 2005) byexpressing novel developmental pathways or by down-stream effects of mutual adjustment of different parts ofthe phenotype a process known as phenotypic accom-modation (Figure 1c Badyaev 2009 Laland et al 2015Schwab et al 2019 West‐Eberhard 2005) Environ-mental stress can thus be understood to simultaneouslyinduce and select for novel phenotypes a process thatwould facilitate rapid evolutionary change (Gibson ampDworkin 2004 Levis amp Pfennig 2016 McGuigan ampSgrograve 2009 Paaby amp Rockman 2014 Schlichting 2008Schwab et al 2019)

Due to the internal architecture of developmentalsystems which generates integrated modules (GoswamiBinder Meachen amp OrsquoKeefe 2015 Schlichting 1989Wagner amp Zhang 2011) random genetic mutations maystill lead to coordinated phenotypic responses (Jablonski2017 Moczek et al 2011 Uller et al 2018 West‐Eberhard 2003) For this reason the effects of thepreviously cryptic genetic variation may in fact causedirected phenotypic change biased towards functionallyintegrated phenotypes (Gerhart amp Kirschner 2007Masel 2006 Wagner 2011 Watson amp Szathmaacutery 2016Watson Wagner Pavlicev Weinreich amp Mills 2014)Taken together and in the context of PLE this impliesthat accumulated CGV revealed by environmental stressmay be biased towards experimental functionality

Initially this constitutes a pulse of phenotypic changewithout genotypic change (Noble 2015 West‐Eberhard2003) However when CGV is thus revealed its pheno-typic manifestation nevertheless becomes visible tonatural selection which can select for or against it Moreimportantly selection indirectly acts on the develop-mental system itself rendering its regulation either moreor less sensitive to environmental conditions (Moczek2008 Schlichting amp Wund 2014 West‐Eberhard 2005Wund 2012) This subsequent process is known asgenetic accommodation (Lande 2009 West‐Eberhard2003) If nature selects strongly for a novel phenotypethat is a newly induced variant that is preadapted to itsenvironment then the process of stabilizing selectionstarts over and development can begin to canalize oncemore (Figure 1d) If one such novel phenotype issubjected to very powerful selection it may lose itsplasticity entirely and is then said to have beengenetically assimilated (Pigliucci Murren amp Schlichting2006 Waddington 1961 West‐Eberhard 2003)

Given that the modular structure of gene regulatorynetworks generates patterns in which parts of develop-ment interact with the environment whereas othersdo not (Iwasaki et al 2013 Schneider amp Meyer 2017)those parts that do interact with and are intermittently

mediated by environmental factors become the centersof evolutionary change and evolution can proceedpredominantly by modification of environmentally regu-lated parts of the developmental system with consequentphenotypic diversification (Schneider amp Meyer 2017Uller et al 2018 Wund 2012) Thus this processrepeated in isolated populations is likely to directevolution along particular trajectories

6 | PLASTICITY ‐LEDEVOLUTION AS A MODE

The cyclical accumulation and release of CGV duringperiods of stabilizing selection and environmental stressrespectively thus describes the operation of PLE This issimilar to Badyaevrsquos (2009) conceptualization of ldquoevolu-tionary diversifications and maintenance of adaptationsprobably operating in alternation between the emer-gence of novel developmental variation and stabilizationof locally appropriate organismndashenvironment associa-tions by natural selectionrdquo and represents an excitinghypothetical mode of evolution Indeed the role ofgenetic assimilation in evolution that is the adoption ofgenetic management of initially environmentally in-duced phenotypes continues to demand attention(Ehrenreich amp Pfennig 2016 Schneider amp Meyer2017 Schwab et al 2019)

7 | QUANTUM EVOLUTION

As a mode of evolution this model of PLE constitutes apattern of periods of stasis during which capacitance isaccrued followed by rapid bursts of change as thiscapacitance is expended In a macroevolutionary contextthis is not without historical analogues Indeed thedistinction between stasis and burst is captured in thecontrast between Simpsonrsquos (1944) phyletic and quantumevolution (Figure 2) two of his three modes of evolutionthe third being speciation He defines quantum evolutionas the ldquorelatively rapid shift of a biotic population indisequilibrium to an equilibrium distinctly unlike anancestral conditionrdquo while phyletic evolution is theldquosustained directional (but not necessarily rectilinear)shiftrdquo more commonly observed in the fossil record

Simpson (1944) considered these two modes ofevolution as qualitatively different to his mind theyfulfill different functions and presumably operate bydifferent mechanisms Phyletic evolution manages theadaptation of species to their adaptive zone whereasquantum evolution is a much more radical processcapable of generating rapid changes and accommodating

4 of 15 | JACKSON

jumps across inadaptive zones to new adaptive zones Inother words quantum evolution was a suggested solutionto the contemporary problem of how a population mightcross the valleys of Wrightrsquos adaptive landscapes (Wright1932) given that evolution proceeds through the gradualaccumulation of small changes (cf Zheng et al 2019) Tothis end Simpson (1944) writes that ldquopreadaptation [is] anecessary element [of quantum evolution and is] anothermarked distinction between quantum evolution andphyletic evolution which is mostly adaptive (andpostadaptive)rdquo Furthermore quantum evolution pro-vided a biological explanation for the long‐observeddiscontinuities in the fossil record they were not merelyartifacts of incompleteness but may in fact be indicativeof extremely rapid phenotypic transitions

According to Simpson (1944) quantum evolutionprogressed through three phases

ldquo(1) An inadaptive phase in which the groupin question loses the equilibrium of itsancestors or collaterals

(2) A preadaptive phase in which there isgreat selection pressure and the group movestoward a new equilibrium

(3) An adaptive phase in which the newequilibrium is reachedrdquo

There is a striking concordance between these phasesand (a) environmental induction of CGV release that isphenotypic accommodation (b) corresponding environ-mental selection for novel variants that is geneticaccommodation and potential assimilation and (c)stabilizing selection that is canalization of development

Yet at the time the best candidate for a mechanismby which quantum evolution might proceed was genetic

drift During the ldquohardeningrdquo of the modern synthesisas evolutionary explanations became increasingly re-stricted to selection acting on gradual allelic change(Gould 1982) the role of genetic drift seemed lesssignificant leading to Simpson changing his views onquantum evolution and ceasing to see it is a differentkind of evolution to phyletic evolution (Gould 1986) Helater wrote that quantum evolution was ldquonot a differentsort of evolution from phyletic evolution or even adistinctly different element of the total phylogeneticpattern It is a special more or less extreme and limitingcase of phyletic evolutionrdquo (Simpson 1953) In otherwords quantum evolution became simply phyleticevolution sped up and the norm mode of evolutionconsequently became phyletic gradualism (Eldredge ampGould 1972 Gould 1980)

8 | PUNCTUATED EQUILIBRIUM

It was against this backdrop that Eldredge and Gouldformulated their theory of punctuated equilibrium(Eldredge amp Gould 1972 Gould amp Eldredge 19771993) Similar to Simpsonrsquos (1944) quantum evolutionit posits that discontinuities in the fossil record need notnecessarily be indicative of incompleteness Ratherphenotypic evolution may be as intermittently rapidthat is punctuated as fossil data suggest They argue thatldquothe remarkable stasis exhibited by most species duringmillions of years is ignored (as no data)rdquo (Gould ampEldredge 1993) In other words stasis is taken as a lackof evolution and contrasted with directional phenotypicchange through time The assumption that discontinu-ities are artifacts of incompleteness implies that phyleticgradualism is synonymous with evolution

Gould and Eldredge (1993) write that we insteadought to ldquoregard stasis and discontinuity as an

FIGURE 2 Simpsonrsquos (1944) comparison of phyletic and quantum evolution Phyletic evolution is the selective tracking of a shiftingadaptive zone through time with subsequent directional phenotypic gradualism as a result Quantum evolution on the other hand is a rapidevolutionary transition between adaptive zones Note the preadaptive variation manifest in the ancestral population immediately beforequantum evolution

JACKSON | 5 of 15

expression of how evolution works when translatedinto geological timerdquo and argue that phenotypicevolution is more closely tied to speciation thanpreviously imagined Indeed studies on patterns ofchange in the fossil record have confirmed that stasis isa far more common observation than phyletic gradu-alism (Hunt 2004 2007 Hunt amp Rabosky 2014) withstasis in this sense more accurately regarded asdirectionless or nonaccumulating morphological fluc-tuations rather than true stagnance (Gould 1982Gould 2002 Voje 2016) Furthermore adaptivephenotypic change may occur so rapidly that it maypass by the record (Hunt 2010) Yet while the patterndescribed by punctuated equilibrium is thereforebroadly supported by observation the mechanism bywhich this pattern arises particularly the rapidtransitions of form remains unclear (Futuyma 2015)Around the time of its inception and early receptionthe theory of punctuated equilibrium was met withstrong critique (eg Charlesworth Lande amp Slatkin1982 Gingerich 1984 Kirkpatrick 1982) generating adebate that continues to this day (eg Lieberman ampEldredge 2014 Pennell Harmon amp Uyeda 2014)

9 | FUNDAMENTAL DEBATE

It seems that the same point of contention that ledSimpson to dilute his concept of quantum evolution alsoframes the debate on the theory of punctuated equili-brium as well as the controversy surrounding phenotypicand genetic accommodation namely the tension be-tween the understanding of development as programmedor constructive (Laland et al 2015) itself a manifestationof the differing perspectives of the gene‐centered andorganism‐centered views of evolution (Baedke 2018Nicholson 2014) Clearly the same currents of thoughtare manifest in each of these instances yet the contentionremains unresolved

It is tempting to argue that Badyaevrsquos (2009)alternating ldquoemergence of novel developmental varia-tionrdquo and ldquostabilization of locally appropriate organis-mndashenvironment associations by natural selectionrdquo arequalitatively different processes the former approxi-mated by quantum evolution punctuated equilibriaand plasticity‐led evolution and the latter by phyleticevolution stasis and gradualism Indeed this would bean approach informed by the distinction betweenorigins and distributional explanations (Erwin 2017)Yet this does not make sense in terms of our under-standing of development

From a developmental perspective gradualism andPLE are not qualitatively different processes Both are

governed by a regulatory network that facilitates varia-tion and buffers genetic and environmental fluctuationstowards functionality (Uller et al 2018) The keydifference is simply the phenotypic accommodation ofPLE which may be exaggerated by the release ofcapacitance in the form of CGV This facilitates thedistinctive rapidity described by quantum evolution(Simpson 1944) and punctuated equilibrium (Eldredgeamp Gould 1972) Thus the difference is quantitative notqualitative a matter of thresholds and amounts ratherthan kinds

10 | MACROEVOLUTIONARYPATTERNS OF PLASTICITY ‐LEDEVOLUTION

PLE as the evolutionary process described abovepredicts in my view in short two major features inthe fossil record (a) stabilizing selection duringperiods of environmental stability leading to decreas-ing phenotypic variation but increasing evolutionarycapacitance in the form of cryptic genetic variationand (b) episodes of environmentally induced pheno-typic plasticity facilitating rapid bursts of evolution asthis capacitance is expended

To be clear in this framework plasticity‐led evolu-tion (PLE) that is the accumulation and release ofcryptic genetic variation (CGV) is a putative generatorof those patterns in the fossil record previouslydescribed as quantum evolution (Simpson 1944) andpunctuated equilibrium (Eldredge amp Gould 1972)Periods of phyletic evolution or stasis are understoodas stabilizing selection acting on an organism Quantumevolution and the rapid shifts of punctuated equilibriaare hypothesized to be facilitated by phenotypic andgenetic accommodation This is not to say that evolutioncannot occur through phyletic gradualism that isstepwise accumulation of genetic change nor to saythat the fossil record is not incomplete in parts I onlymean to say that PLE provides an alternative explana-tion for the phenomena mentioned above While thetheory of punctuated equilibrium is associated withspeciation and quantum evolution is traditionallydiscussed in terms of anagenetic evolution despiteSimpson (1944) himself writing that it ldquomay be involvedin either speciation or phyletic evolutionrdquo and thatldquocertain patterns within those modes intergrade withquantum evolutionrdquo this distinction holds little sig-nificance for this model of PLE the environmentallymediated accumulation and release of CGV can occur ineither instance and whether or not speciation occurs isinconsequential

6 of 15 | JACKSON

11 | EVIDENCE OF PLASTICITY‐LEDEVOLUTION IN THE FOSSIL RECORD

When searching for evidence of PLE in the fossil recordit is insufficient to simply point to patterns of stasis ordiscontinuity As Gould and Eldredge (1993) point outldquoExamples of stasis alone [hellip] and simple abruptreplacement although conforming to expectations ofpunctuated equilibrium are not direct evidence for ourmechanism for stasis might just be a lull in anageneticgradualism [hellip] and replacement might represent rapidtransformation without branching or migration of adistant (phyletic or geographic) relative rather thanevolution in siturdquo Likewise discontinuities may indeedhave geological rather than biological explanationscaused by erosion of fossil‐bearing strata or a hiatus ina sedimentary deposition for example Furthermore theprocesses of phenotypic and genetic accommodation arechallenging to study in the fossil record As Pigliucciand Murren (2003) caution us ldquogiven the hypothesisthat [genetic] assimilation can occur within a fewgenerations it ironically may be too fast for us tocatchrdquo Indeed the rapidity of phenotypic transitionsreported by Hunt (2010) points to a similar problemThese challenges present a problem to the study of PLEin the fossil record

Beyond these challenges a further difficulty is thatthe central macroevolutionary prediction of PLE issuperficially identical to that of phyletic gradualismPLE predicts that novel environments lead to evolu-tionary change mediated by phenotypic and geneticaccommodation Phyletic gradualism predicts that novelenvironments lead to evolutionary change mediated bynatural selection on gradually accumulating mutationsThus the observation of phenotypic change accompa-nying environmental change is insufficient to supportPLE over phyletic gradualism without access to geneticmaterial and consequently the underlying processes weare left unable to determine which explanation is morelikely In general this is a difficulty faced wheninvestigating the fossil record in search for support forPLE macroevolutionary patterns have historically pro-ven an inadequate source of discrimination betweengenetic variation and plasticity (Hopkins 2014 Schoch2014 Webster 2019)

While the fact that the process of PLE is coherent withthe structure of the fossil record should not be ignored asinsignificant it remains the case that to provide definitivesupport for PLE as a mode of evolution we need to directour attention towards identifying evidence that distin-guishes it from phyletic gradualism Approaches towardsdifferentiating these processes rely on the study ofprocesses instead of patterns In other words they focus

on providing support for the underlying mechanisms ofPLE rather than the broad‐scale outcome

12 | ENVIRONMENTAL STRESSAND PHENOTYPIC VARIABILITY

Specifically the key feature distinguishing PLE fromphyletic gradualism is the environmentally mediateddevelopmental plasticity that facilitates rapid phenotypicchange While such change can be measured in thedifference between mean morphologies this does little tohelp separate hypotheses of PLE and phyletic gradualismseeing as both predict shifting phenotypic means as aresult of environmental change A stronger indicator ofreduced canalization and subsequent increased develop-mental plasticity is a measure of the morphologicalvariability induced by environmental stress (Hoffmann ampHercus 2000)

The relationship between environmental stress andmorphology is the subject of the Plus ccedila Change model(Sheldon 1997) In short it seeks to explain theobservation that morphological stasis appears the mostcommon response to fluctuations in environmentalconditions in the fossil record Drawing on a deep‐timeperspective it argues that long‐term generalists are thespecies most likely to survive such fluctuations andconsequently predicts gradualistic phenotypic changeunder stable or near‐stable environmental conditionsand phenotypic stasis during periods of strong environ-mental fluctuations

In its general sense this model seemingly suggests theopposite of PLE to manifest in the fossil record Yet itcontains the caveat that upon reaching a critical thresh-old of stress rapid phyletic evolution may be the result(Figure 3) The parallels between this prediction and theprocess of PLE outlined above are readily apparent thismodel suggests that in a strongly fluctuating environmentnatural selection favors those organisms that sufficientlycanalize their development to best track an underlyingphenotypic optimum yet when an extreme environmentis encountered developmental canalization is overcomeand previously cryptic variation is released

The prediction that environmental stressors candecrease developmental stability by overcoming cana-lization and releasing CGV implies that phenotypicvariability can be regarded as a proxy for develop-mental instability (Klingenberg 2019 Webster 2019Willmore Young amp Richtsmeier 2007) Thus thephenotypendashenvironment association of interest for thestudy of PLE in the fossil record is that betweenenvironmental conditions and phenotypic variabilityrather than phenotypic variation

JACKSON | 7 of 15

This distinction is key for separating the predictions ofPLE and phyletic gradualism Variation that is morpho-logical differences between assemblages associated withenvironmental change is insufficient to support PLE forthe reasons discussed above However an associationbetween environmental change and variability that isthe variance in form within assemblages informs usabout their developmental stability and thus permitsinvestigation into the unique prediction of PLE thatenvironmental stress may overcome canalization andinduce the release of CGV with consequent expression ofnovel variants upon which selection can act Theconsequence of this process is an increase in phenotypicvariability coupled with a shift in phenotypic meanassociated with environmental stress a pattern notpredicted by phyletic gradualism

Below I provide two examples of approaches towardsproviding evidence of PLE in the fossil record that usethis association Before discussing them however I firstpresent their necessary methodological framework

13 | FRAMEWORK OF ANALYSIS

The focal point for these investigations is the relation-ship between environmental factors and morphology offossil assemblages Morphology is essentially the onlybiological information available in the fossil record andit is therefore in general of paramount importance tocapture and study morphology in as high resolution aspossible and with careful consideration of potentialtaphonomic biases such as disarticulation of hardparts or later structural deformation (Webster 2019)This is particularly true however when studyingintraspecific patterns of variation such as thoseI will present below Many modern phylogeneticcladistic analyses rely on discretizing variation intocharacters which strongly limits their ability to studyintraspecific variation An alternative approach toanalyzing morphological variation is through the useof morphometric methods which quantify shapes intosets of measurements that can be mathematicallyanalyzed (MacLeod 2017) Such methods includelandmark analyses (Bookstein 1997) eigenshape ana-lysis (MacLeod 2002) and elliptical Fourier analysis(Kuhl amp Giardina 1982) These numerical approachesto describing variation in form translate morphologyinto continuous data and are consequently more suitedto the study of the subtle variation exhibited intraspe-cifically than the use of discrete characters and indeedare a prerequisite to examining fine‐scale patterns ofmorphological variability

While straightforward use of statistical variance isoften employed in studies on developmental instability(eg Imasheva Loeschcke Zhivotovsky amp Lazebny1997) novel methods for quantifying variance specificallystemming from developmental plasticity have seen recentdevelopment (Pertoldi amp Kristensen 2015 Pertoldi et al2014) Coupling shifts in variability to shifts in pheno-typic mean is the key to identifying developmentalplasticity in the fossil record as I will discuss belowThe putative confounding effect of time‐averaging onmorphological variance in fossil specimens has not beensupported by empirical studies Rather it seems thatpaleontological and neontological samples have compar-able levels of variability (Kidwell amp Flessa 1995 BushPowell Arnold Bert amp Daley 2002 Hunt 2004 Hunt2004 Hunt Bell amp Travis 2008 Webster 2019)

To approach the environmental aspect of PLE it isnecessary to identify indications of environmental con-ditions in fossil assemblages Patterns of intraspecificvariation in fossil species across a geographic andputative environmental range are by no means uncom-mon (eg Hopkins amp Webster 2009 Jackson amp Budd2017) Such patterns have been studied through the

FIGURE 3 Sheldonrsquos (1997) Plus ca Change model Stronglyfluctuating environments are predicted to increase selection forstasis until a threshold is reached and a punctuation sensuEldredge and Gould (1972) is generated This model can beinterpreted as predicting selection for increased canalization influctuating environments until a threshold is reached at whichpoint cryptic genetic variation is expressed and phenotypicaccommodation occurs

8 of 15 | JACKSON

assessment of associations between morphology andvarious environmental gradients (eg Cisne ChandleeRabe amp Cohen 1982 Scarponi amp Kowalewski 2004Webber amp Hunda 2007) Indeed a variety of geochemicalproxies are available for the investigation and approx-imation of depositional environments (Webster 2019)such as using trace metals as a proxy for paleoredox andpaleoproductivity (Tribovillard Algeo Lyons amp Riboul-leau 2006) for example However care must be takenwhen interpreting such proxies to minimize environ-mental variation generated by taphonomic processessuch as diagenetic alterations of the rock (Webster 2019)

The relationship between environmental stress anddevelopmental plasticity by proxy of morphologicalvariability is used in the following two approachestowards providing evidence of PLE as a mode ofevolution in the fossil record

14 | EVIDENCE OF PHENOTYPICAND GENETIC ACCOMMODATION

As mentioned above both phyletic gradualism and PLEpredict a shift in morphology following exposure tonovel environments albeit resulting from differentprocesses making it difficult to uncover traces defini-tively of PLE Gradual adaptive evolution wouldmanifest as an assemblage shifting towards a newphenotypic optimum in response to environmentalperturbation Likewise PLE would present a similarpattern yet with the environment both inducing as wellas selecting for novel variants

However PLE predicts a pattern of increasingvariability as the expression of CGV is induced byenvironmental stress which distinguishes it fromphyletic gradualism In this mode of evolution pheno-typic accommodation increases variability such that thephenotype subsequently genetically accommodated or

canalized lies outside its ancestral normal range (Figure4 cf Figure 1) While developmental instability isknown to occur during stress including on the rangemargins of populations (Bridle amp Vines 2007 RiesebergArcher amp Wayne 1999) such dynamics do not predictthe induction of novel variation outside the ancestralrange Consequently shifts in mean morphologycoupled with strongly increasing variability beyond anancestrally normal range can be used as an indicator ofPLE (Figure 4)

This means that the study of morphological varia-bility across environmental gradients offers an oppor-tunity to test specifically for evidence of the phenotypicand genetic accommodation inherent to PLE as de-scribed above

Naturally demonstrating this kind of intraspecificpattern requires a fossil organism or lineage with a highlyresolved stratigraphy sufficient abundance and knownpresence across temporal andor environmental gradients(Schoch 2014) Such research is not uncommon (egHopkins amp Webster 2009 Schoch 2014 Webber ampHunda 2007) yet morphological variability is rarely thefocal point and phenotypic accommodation rarer still

However the Cambrian trilobite‐like arthropod Ag-nostus pisiformis has been demonstrated to exhibitdifferentiated patterns of variation and variability acrossa gradient of dysoxic stress (Jackson amp Budd 2017)implying that environmental conditions may indeed beinducing as well as selecting for phenotypic variants

Further investigation in this framework is needed toproperly test the fossil record for evidence of PLEPatterns of phenotypic variation and chiefly variabilityuncovered by fine‐scale studies of fossil morphology (egDai Zhang Peng amp Yao 2017 Esteve 2012 NeubauerHarzhauser amp Kroh 2013) ought to be considered withinthe framework of PLE Likewise research exploringputative canalization of fossil organisms (eg PimientoTang Zamora Klug amp Saacutenchez‐Villagra 2018 Webster

Mor

phol

ogy

Time

Mor

phol

ogy

Time

(a) (b) EnvironmentalOptimum

Population

FIGURE 4 Illustration of two hypothetical responses of an evolutionary lineage to an environmental perturbation (a) Natural selectioncauses an adaptive shift of morphology towards a new environmental optimum in line with the predictions of phyletic gradualism (b) Theenvironmentally mediated release of cryptic genetic variation increases phenotypic variability facilitating rapid adaptive evolution towards anew optimum as per the process of plasticity‐led evolution Stabilizing selection subsequently leads to canalization around this newoptimum

JACKSON | 9 of 15

2015) could explicitly test for indications of PLE havingoccurred

15 | EVIDENCE OFDEVELOPMENTALLY BIASEDEVOLUTIONARY TRAJECTORIES

Adaptive radiations represent periods of rapid evolu-tion in which ancestral forms of a lineage diversify inenvironment morphology functionality and so forthWhen such patterns feature repeated evolutionarysequences such as in the case of the Anolis lizards ofthe Caribbean (Losos Jackman Larson de Queiroz ampRodriacuteguez‐Schettino 1998) they afford us the oppor-tunity to examine the fossil record for evidence ofanother component of PLE namely whether or notrepeated patterns of phenotypic evolution are some-times better explained by developmental bias thanconvergent selection At the heart of this issue lies therole of the environment in adaptive radiationsTraditionally the environment is seen as facilitatingsuch episodes by providing ecological opportunity todiversifying clades (Stroud amp Losos 2016) However asoutlined above the environment may also act as aninducer of novel variants by interacting with develop-mental processes Erwin (2017) describes these per-spectives as emphasizing ldquoenvironmental pullrdquo andldquodevelopmental pushrdquo respectively

West‐Eberhardrsquos (2003) flexible‐stem model sug-gests that plasticity mediated by the regulatorystructure of an ancestral populationrsquos developmentwill bias repeated independent diversificationsalong similar evolutionary trajectories via the pro-cesses of phenotypic and genetic accommodationThis represents an alternative or complementaryexplanation as to why independently derived lineages

of an ancestral population repeatedly evolve similarphenotypes such as the ldquoecomorphsrdquo of the Anolislizards mentioned above (Losos et al 1998 Williams1972) While this model has been explored in aneontological context (Gibert 2017 Parsons et al2016 Schneider amp Meyer 2017 Wund Baker ClancyGolub amp Foster 2008) it has not yet seen attention ina paleontological setting

Although the identification of adaptive radiations inthe fossil record is not a straightforward endeavor(Lieberman 2012) essentially any sequence of repeateddivergence from an ancestral stock population is suitablefor investigating the flexible‐stem informed model ofparallelism Similarly to the discussion on phenotypicand genetic accommodation above the focal point ofresearch in such systems should be on the variability ofmorphology rather than variation relative environmen-tal gradients If it is indeed phenotypic accommodationthat is the environmentally mediated release of CGVthat directs these repeated evolutionary trajectories weshould predict an increasing degree of phenotypicvariability to coincide with these divergences (Figure 5)This coupling of increasing variability and shiftingvariation is not conventionally anticipated in a patternof repeated parallel divergences that is convergentselection as explained above

The positive identification of such patterns in thefossil record would thus provide distinctive support forthe underlying processes of PLE both in terms of thesuggested environmentally mediated release of CGV aswell as the hypothesized bias exercised by the regulatorynature of development Research towards uncoveringsuch support should be directed towards reassessing thephenotypendashenvironment associations of studied fossiladaptive radiations (eg Abe amp Lieberman 2012 NeigeDera amp Dommergues 2013 Neubauer et al 2013) in thecontext of PLE

Mor

phol

ogy

Time

EnvironmentalOptimum

Ancestral Population

Diverging Population

FIGURE 5 Repeated divergences from an ancestral or stock population The bias native to the ancestral developmental network directsmorphological change along similar evolutionary trajectories as the successive independently diverging populations encounter the newenvironment An increase in morphological variability coupled with such parallelism of phenotypic mean change is predicted by the releaseof cryptic genetic variation in plasticity‐led evolution but not by the conventional process of convergent evolution resulting from similar andrepeated selective pressures

10 of 15 | JACKSON

16 | CONCLUDING REMARKS

Viewing development as constructive is a promisingconceptual framework for enriching our understandingof the processes and patterns of evolution It provides thescaffolding for properly addressing issues of centralimportance such as the role of developmental biases indirecting evolutionary change However the majority ofefforts towards this end are conducted on extantorganisms with minimal reference to paleontologicaldata This is unfortunate as the fossil record is afantastically rich source of information that is yet to befully explored in this intriguing and stimulating contextWhile microevolutionary concepts such as developmentalbias are challenging to approach in the fossil record Iargue that it is manifest in paleontological materials inthe form of macroevolutionary patterns concordant withthe predictions of plasticity‐led evolution Indeed suchpredictions recapitulate earlier theories of quantumevolution and punctuated equilibrium

These predictions are specifically an evolutionarymodel consisting of two major parts (a) stabilizingselection during periods of environmental stabilityleading to a reduction in phenotypic variation but anincrease in canalization CGV and hence evolutionarycapacitance and (b) episodes of environmentally inducedplasticity facilitating rapid evolution via the expression ofpreviously cryptic variation

Research aiming to uncover support for PLE in thefossil record should be directed towards the environ-mentndashphenotype association of fossil assemblages Spe-cifically towards patterns of correlation between envir-onmental stress and phenotypic variability as a proxy fordevelopmental instability These may be manifest alongsingle lineages or throughout adaptive radiations Thefossil record offers enormous potential to provide datathat when viewed through the lens of PLE providesbiologically meaningful information about evolution interms of both patterns and processes

ACKNOWLEDGMENTS

I would like to thank Armin Moczek for inviting me tocontribute to this special issue on developmental bias inevolution My thanks to two anonymous reviewers whoprovided thoughtful constructive and immensely helpfulcomments that greatly improved this paper I am gratefulto Graham Budd for our many discussions on this topicover the years I am thankful to the Uller group at LundUniversity for formative discussions particularly TobiasUller who also provided valuable comments on an earlydraft of this paper My thoughts presented here weregreatly informed by the SFI workshop ldquoDevelopmental

Bias and Evolutionrdquo and I direct my gratitude towards itsorganizers and participants This study was funded by theKnut and Alice Wallenberg Foundation grant to TobiasUller (KAW 20120155)

CONFLICT OF INTEREST

The author declares that there is no conflict of interest

ORCID

Illiam S C Jackson httporcidorg0000-0002-7948-2860

REFERENCES

Abe F R amp Lieberman B S (2012) Quantifying morphologicalchange during an evolutionary radiation of Devonian trilobitesPaleobiology 38 292ndash307 httpsdoiorg101666100471

Arthur W (2004) The effect of development on the directionof evolution Toward a twenty‐first century consensusEvolution amp Development 6(4) 282ndash288 httpsdoiorg101111j1525‐142X200404033x

Badyaev A V (2005) Stress‐induced variation in evolution Frombehavioural plasticity to genetic assimilation Proceedings of theRoyal Society B Biological Sciences 272 877ndash886 httpsdoiorg101098rspb20043045

Badyaev A V (2009) Evolutionary significance of phenotypicaccommodation in novel environments An empirical test of theBaldwin effect Philosophical Transactions of the Royal SocietyB Biological Sciences 364 1125ndash1141 httpsdoiorg101098rstb20080285

Baedke J (2018) Organism where art thou Old and new challengesfor organism‐centered biology Journal of the History of Biology52 293ndash324 httpsdoiorg101007s10739‐018‐9549‐4

Bookstein F L (1997) Morphometric tools for landmark dataGeometry and biology Cambridge Cambridge University Press

Brakefield P M (2011) Evo‐devo and accounting for Darwinsendless forms Philosophical Transactions of the Royal Society BBiological Sciences 366(1574) 2069ndash2075 httpsdoiorg101098rstb20110007

Bridle J R amp Vines T H (2007) Limits to evolution at rangemargins When and why does adaptation fail Trends in Ecologyamp Evolution 22(3) 140ndash147 httpsdoiorg101016jtree200611002

Bush A M Powell M G Arnold W S Bert T M amp Daley G M(2002) Time‐averaging evolution and morphologic variationPaleobiology 28 9ndash25 httpsdoiorg1016660094‐8373(2002)0283C0009TAEAMV3E20CO2

Carroll S B (2005) Evolution at two levels On genes and form PLOSBiology 3 e245 httpsdoiorg101371journalpbio0030245

Carroll S B (2008) Evo‐devo and an expanding evolutionarysynthesis A genetic theory of morphological evolution Cell134 25ndash36 httpsdoiorg101016jcell200806030

Charlesworth B Lande R amp Slatkin M (1982) A neo‐Darwiniancommentary on macroevolution Evolution 36 474ndash498 httpsdoiorg1023072408095

JACKSON | 11 of 15

5 | PLASTICITY ‐LEDEVOLUTION

PLE also known as the plasticity‐first hypothesisdescribes a mode of evolution in which the environmentplays a pivotal role in both inducing and subsequentlyselecting for phenotypic variation (Levis amp Pfennig 2016West‐Eberhard 2003 2005) It has been investigated inextant organisms such as sticklebacks (Wund ValenaWood amp Baker 2012) house finches (Badyaev 2009)and spadefoot toads (Levis amp Pfennig 2019) yet how wemight approach this process in a paleontological contextremains unknown Here follows a summary of PLE alsoillustrated in Figure 1

A population of organisms well‐adapted to itsenvironment is subject to the process of stabilizing

selection which selects not only for a specific pheno-typic optimum but also selects against variation aroundthis optimum (Figure 1a Schmalhausen 1949) Apartfrom the removal of unfit variants this is achieved byreducing developmental sensitivity to any perturbationincluding environmental conditions (Gibson amp Wagner2000 Pertoldi Bundgaard Loeschcke amp Barker 2014Waddington 1942 Wagner Booth amp Bagheri‐Chai-chian 1997) When development is canalized in thismanner CGV may accumulate in shielded parts of thegenome (Kuumlttner et al 2014 Moczek 2008 Rutherford2000 Schlichting amp Smith 2002 Zheng et al 2019)generating evolutionary capacitance (Figure 1b Gibsonamp Dworkin 2004 Levis amp Pfennig 2016 McGuigan ampSgrograve 2009 Paaby amp Rockman 2014 Schlichting 2008Schwab et al 2019)

yg

olo

hpr

oM

Selective

gradient

Mo

rph

olo

gy

Mo

rph

olo

gy

Selective

gradient

yg

olo

hpr

oM

Morphology

Morphology

Morphology

Morphology

(a) (b)

(c)(d)

FIGURE 1 Illustration of the process of plasticity‐led evolution in terms of phenotype stress and cryptic genetic variation (CGV)Ellipses represent populations in morphospace The shading intensity of the ellipses gives their relative level of CGV (a) Stabilizing selectionacting on a population selects both for a phenotypic optimum as well as reduced variation around that optimum (b) Phenotypic variation isreduced through both the removal of unfit variants as well as decreasing sensitivity to environmental conditions that is canalization ofdevelopment This shields previously expressed but now hidden developmental pathways from natural selection such that CGV that isevolutionary capacitance can accumulate (c) An environmental stress is encountered that both induces as well as selects for variation CGVis revealed by this environmental trigger and phenotypic accommodation occurs generating increasing phenotypic variability This can beunderstood as the expenditure of the previously accumulated evolutionary capacitance (d) Selection acts on the uncovered variation byselecting for the new environmental optimum and genetically accommodating such change that was beneficial thus restarting this processby undergoing stabilizing selection again and recanalizing development once more

JACKSON | 3 of 15

Novel or extreme environmental conditions mayreveal this capacitance (Badyaev 2005 Flatt 2005) byexpressing novel developmental pathways or by down-stream effects of mutual adjustment of different parts ofthe phenotype a process known as phenotypic accom-modation (Figure 1c Badyaev 2009 Laland et al 2015Schwab et al 2019 West‐Eberhard 2005) Environ-mental stress can thus be understood to simultaneouslyinduce and select for novel phenotypes a process thatwould facilitate rapid evolutionary change (Gibson ampDworkin 2004 Levis amp Pfennig 2016 McGuigan ampSgrograve 2009 Paaby amp Rockman 2014 Schlichting 2008Schwab et al 2019)

Due to the internal architecture of developmentalsystems which generates integrated modules (GoswamiBinder Meachen amp OrsquoKeefe 2015 Schlichting 1989Wagner amp Zhang 2011) random genetic mutations maystill lead to coordinated phenotypic responses (Jablonski2017 Moczek et al 2011 Uller et al 2018 West‐Eberhard 2003) For this reason the effects of thepreviously cryptic genetic variation may in fact causedirected phenotypic change biased towards functionallyintegrated phenotypes (Gerhart amp Kirschner 2007Masel 2006 Wagner 2011 Watson amp Szathmaacutery 2016Watson Wagner Pavlicev Weinreich amp Mills 2014)Taken together and in the context of PLE this impliesthat accumulated CGV revealed by environmental stressmay be biased towards experimental functionality

Initially this constitutes a pulse of phenotypic changewithout genotypic change (Noble 2015 West‐Eberhard2003) However when CGV is thus revealed its pheno-typic manifestation nevertheless becomes visible tonatural selection which can select for or against it Moreimportantly selection indirectly acts on the develop-mental system itself rendering its regulation either moreor less sensitive to environmental conditions (Moczek2008 Schlichting amp Wund 2014 West‐Eberhard 2005Wund 2012) This subsequent process is known asgenetic accommodation (Lande 2009 West‐Eberhard2003) If nature selects strongly for a novel phenotypethat is a newly induced variant that is preadapted to itsenvironment then the process of stabilizing selectionstarts over and development can begin to canalize oncemore (Figure 1d) If one such novel phenotype issubjected to very powerful selection it may lose itsplasticity entirely and is then said to have beengenetically assimilated (Pigliucci Murren amp Schlichting2006 Waddington 1961 West‐Eberhard 2003)

Given that the modular structure of gene regulatorynetworks generates patterns in which parts of develop-ment interact with the environment whereas othersdo not (Iwasaki et al 2013 Schneider amp Meyer 2017)those parts that do interact with and are intermittently

mediated by environmental factors become the centersof evolutionary change and evolution can proceedpredominantly by modification of environmentally regu-lated parts of the developmental system with consequentphenotypic diversification (Schneider amp Meyer 2017Uller et al 2018 Wund 2012) Thus this processrepeated in isolated populations is likely to directevolution along particular trajectories

6 | PLASTICITY ‐LEDEVOLUTION AS A MODE

The cyclical accumulation and release of CGV duringperiods of stabilizing selection and environmental stressrespectively thus describes the operation of PLE This issimilar to Badyaevrsquos (2009) conceptualization of ldquoevolu-tionary diversifications and maintenance of adaptationsprobably operating in alternation between the emer-gence of novel developmental variation and stabilizationof locally appropriate organismndashenvironment associa-tions by natural selectionrdquo and represents an excitinghypothetical mode of evolution Indeed the role ofgenetic assimilation in evolution that is the adoption ofgenetic management of initially environmentally in-duced phenotypes continues to demand attention(Ehrenreich amp Pfennig 2016 Schneider amp Meyer2017 Schwab et al 2019)

7 | QUANTUM EVOLUTION

As a mode of evolution this model of PLE constitutes apattern of periods of stasis during which capacitance isaccrued followed by rapid bursts of change as thiscapacitance is expended In a macroevolutionary contextthis is not without historical analogues Indeed thedistinction between stasis and burst is captured in thecontrast between Simpsonrsquos (1944) phyletic and quantumevolution (Figure 2) two of his three modes of evolutionthe third being speciation He defines quantum evolutionas the ldquorelatively rapid shift of a biotic population indisequilibrium to an equilibrium distinctly unlike anancestral conditionrdquo while phyletic evolution is theldquosustained directional (but not necessarily rectilinear)shiftrdquo more commonly observed in the fossil record

Simpson (1944) considered these two modes ofevolution as qualitatively different to his mind theyfulfill different functions and presumably operate bydifferent mechanisms Phyletic evolution manages theadaptation of species to their adaptive zone whereasquantum evolution is a much more radical processcapable of generating rapid changes and accommodating

4 of 15 | JACKSON

jumps across inadaptive zones to new adaptive zones Inother words quantum evolution was a suggested solutionto the contemporary problem of how a population mightcross the valleys of Wrightrsquos adaptive landscapes (Wright1932) given that evolution proceeds through the gradualaccumulation of small changes (cf Zheng et al 2019) Tothis end Simpson (1944) writes that ldquopreadaptation [is] anecessary element [of quantum evolution and is] anothermarked distinction between quantum evolution andphyletic evolution which is mostly adaptive (andpostadaptive)rdquo Furthermore quantum evolution pro-vided a biological explanation for the long‐observeddiscontinuities in the fossil record they were not merelyartifacts of incompleteness but may in fact be indicativeof extremely rapid phenotypic transitions

According to Simpson (1944) quantum evolutionprogressed through three phases

ldquo(1) An inadaptive phase in which the groupin question loses the equilibrium of itsancestors or collaterals

(2) A preadaptive phase in which there isgreat selection pressure and the group movestoward a new equilibrium

(3) An adaptive phase in which the newequilibrium is reachedrdquo

There is a striking concordance between these phasesand (a) environmental induction of CGV release that isphenotypic accommodation (b) corresponding environ-mental selection for novel variants that is geneticaccommodation and potential assimilation and (c)stabilizing selection that is canalization of development

Yet at the time the best candidate for a mechanismby which quantum evolution might proceed was genetic

drift During the ldquohardeningrdquo of the modern synthesisas evolutionary explanations became increasingly re-stricted to selection acting on gradual allelic change(Gould 1982) the role of genetic drift seemed lesssignificant leading to Simpson changing his views onquantum evolution and ceasing to see it is a differentkind of evolution to phyletic evolution (Gould 1986) Helater wrote that quantum evolution was ldquonot a differentsort of evolution from phyletic evolution or even adistinctly different element of the total phylogeneticpattern It is a special more or less extreme and limitingcase of phyletic evolutionrdquo (Simpson 1953) In otherwords quantum evolution became simply phyleticevolution sped up and the norm mode of evolutionconsequently became phyletic gradualism (Eldredge ampGould 1972 Gould 1980)

8 | PUNCTUATED EQUILIBRIUM

It was against this backdrop that Eldredge and Gouldformulated their theory of punctuated equilibrium(Eldredge amp Gould 1972 Gould amp Eldredge 19771993) Similar to Simpsonrsquos (1944) quantum evolutionit posits that discontinuities in the fossil record need notnecessarily be indicative of incompleteness Ratherphenotypic evolution may be as intermittently rapidthat is punctuated as fossil data suggest They argue thatldquothe remarkable stasis exhibited by most species duringmillions of years is ignored (as no data)rdquo (Gould ampEldredge 1993) In other words stasis is taken as a lackof evolution and contrasted with directional phenotypicchange through time The assumption that discontinu-ities are artifacts of incompleteness implies that phyleticgradualism is synonymous with evolution

Gould and Eldredge (1993) write that we insteadought to ldquoregard stasis and discontinuity as an

FIGURE 2 Simpsonrsquos (1944) comparison of phyletic and quantum evolution Phyletic evolution is the selective tracking of a shiftingadaptive zone through time with subsequent directional phenotypic gradualism as a result Quantum evolution on the other hand is a rapidevolutionary transition between adaptive zones Note the preadaptive variation manifest in the ancestral population immediately beforequantum evolution

JACKSON | 5 of 15

expression of how evolution works when translatedinto geological timerdquo and argue that phenotypicevolution is more closely tied to speciation thanpreviously imagined Indeed studies on patterns ofchange in the fossil record have confirmed that stasis isa far more common observation than phyletic gradu-alism (Hunt 2004 2007 Hunt amp Rabosky 2014) withstasis in this sense more accurately regarded asdirectionless or nonaccumulating morphological fluc-tuations rather than true stagnance (Gould 1982Gould 2002 Voje 2016) Furthermore adaptivephenotypic change may occur so rapidly that it maypass by the record (Hunt 2010) Yet while the patterndescribed by punctuated equilibrium is thereforebroadly supported by observation the mechanism bywhich this pattern arises particularly the rapidtransitions of form remains unclear (Futuyma 2015)Around the time of its inception and early receptionthe theory of punctuated equilibrium was met withstrong critique (eg Charlesworth Lande amp Slatkin1982 Gingerich 1984 Kirkpatrick 1982) generating adebate that continues to this day (eg Lieberman ampEldredge 2014 Pennell Harmon amp Uyeda 2014)

9 | FUNDAMENTAL DEBATE

It seems that the same point of contention that ledSimpson to dilute his concept of quantum evolution alsoframes the debate on the theory of punctuated equili-brium as well as the controversy surrounding phenotypicand genetic accommodation namely the tension be-tween the understanding of development as programmedor constructive (Laland et al 2015) itself a manifestationof the differing perspectives of the gene‐centered andorganism‐centered views of evolution (Baedke 2018Nicholson 2014) Clearly the same currents of thoughtare manifest in each of these instances yet the contentionremains unresolved

It is tempting to argue that Badyaevrsquos (2009)alternating ldquoemergence of novel developmental varia-tionrdquo and ldquostabilization of locally appropriate organis-mndashenvironment associations by natural selectionrdquo arequalitatively different processes the former approxi-mated by quantum evolution punctuated equilibriaand plasticity‐led evolution and the latter by phyleticevolution stasis and gradualism Indeed this would bean approach informed by the distinction betweenorigins and distributional explanations (Erwin 2017)Yet this does not make sense in terms of our under-standing of development

From a developmental perspective gradualism andPLE are not qualitatively different processes Both are

governed by a regulatory network that facilitates varia-tion and buffers genetic and environmental fluctuationstowards functionality (Uller et al 2018) The keydifference is simply the phenotypic accommodation ofPLE which may be exaggerated by the release ofcapacitance in the form of CGV This facilitates thedistinctive rapidity described by quantum evolution(Simpson 1944) and punctuated equilibrium (Eldredgeamp Gould 1972) Thus the difference is quantitative notqualitative a matter of thresholds and amounts ratherthan kinds

10 | MACROEVOLUTIONARYPATTERNS OF PLASTICITY ‐LEDEVOLUTION

PLE as the evolutionary process described abovepredicts in my view in short two major features inthe fossil record (a) stabilizing selection duringperiods of environmental stability leading to decreas-ing phenotypic variation but increasing evolutionarycapacitance in the form of cryptic genetic variationand (b) episodes of environmentally induced pheno-typic plasticity facilitating rapid bursts of evolution asthis capacitance is expended

To be clear in this framework plasticity‐led evolu-tion (PLE) that is the accumulation and release ofcryptic genetic variation (CGV) is a putative generatorof those patterns in the fossil record previouslydescribed as quantum evolution (Simpson 1944) andpunctuated equilibrium (Eldredge amp Gould 1972)Periods of phyletic evolution or stasis are understoodas stabilizing selection acting on an organism Quantumevolution and the rapid shifts of punctuated equilibriaare hypothesized to be facilitated by phenotypic andgenetic accommodation This is not to say that evolutioncannot occur through phyletic gradualism that isstepwise accumulation of genetic change nor to saythat the fossil record is not incomplete in parts I onlymean to say that PLE provides an alternative explana-tion for the phenomena mentioned above While thetheory of punctuated equilibrium is associated withspeciation and quantum evolution is traditionallydiscussed in terms of anagenetic evolution despiteSimpson (1944) himself writing that it ldquomay be involvedin either speciation or phyletic evolutionrdquo and thatldquocertain patterns within those modes intergrade withquantum evolutionrdquo this distinction holds little sig-nificance for this model of PLE the environmentallymediated accumulation and release of CGV can occur ineither instance and whether or not speciation occurs isinconsequential

6 of 15 | JACKSON

11 | EVIDENCE OF PLASTICITY‐LEDEVOLUTION IN THE FOSSIL RECORD

When searching for evidence of PLE in the fossil recordit is insufficient to simply point to patterns of stasis ordiscontinuity As Gould and Eldredge (1993) point outldquoExamples of stasis alone [hellip] and simple abruptreplacement although conforming to expectations ofpunctuated equilibrium are not direct evidence for ourmechanism for stasis might just be a lull in anageneticgradualism [hellip] and replacement might represent rapidtransformation without branching or migration of adistant (phyletic or geographic) relative rather thanevolution in siturdquo Likewise discontinuities may indeedhave geological rather than biological explanationscaused by erosion of fossil‐bearing strata or a hiatus ina sedimentary deposition for example Furthermore theprocesses of phenotypic and genetic accommodation arechallenging to study in the fossil record As Pigliucciand Murren (2003) caution us ldquogiven the hypothesisthat [genetic] assimilation can occur within a fewgenerations it ironically may be too fast for us tocatchrdquo Indeed the rapidity of phenotypic transitionsreported by Hunt (2010) points to a similar problemThese challenges present a problem to the study of PLEin the fossil record

Beyond these challenges a further difficulty is thatthe central macroevolutionary prediction of PLE issuperficially identical to that of phyletic gradualismPLE predicts that novel environments lead to evolu-tionary change mediated by phenotypic and geneticaccommodation Phyletic gradualism predicts that novelenvironments lead to evolutionary change mediated bynatural selection on gradually accumulating mutationsThus the observation of phenotypic change accompa-nying environmental change is insufficient to supportPLE over phyletic gradualism without access to geneticmaterial and consequently the underlying processes weare left unable to determine which explanation is morelikely In general this is a difficulty faced wheninvestigating the fossil record in search for support forPLE macroevolutionary patterns have historically pro-ven an inadequate source of discrimination betweengenetic variation and plasticity (Hopkins 2014 Schoch2014 Webster 2019)

While the fact that the process of PLE is coherent withthe structure of the fossil record should not be ignored asinsignificant it remains the case that to provide definitivesupport for PLE as a mode of evolution we need to directour attention towards identifying evidence that distin-guishes it from phyletic gradualism Approaches towardsdifferentiating these processes rely on the study ofprocesses instead of patterns In other words they focus

on providing support for the underlying mechanisms ofPLE rather than the broad‐scale outcome

12 | ENVIRONMENTAL STRESSAND PHENOTYPIC VARIABILITY

Specifically the key feature distinguishing PLE fromphyletic gradualism is the environmentally mediateddevelopmental plasticity that facilitates rapid phenotypicchange While such change can be measured in thedifference between mean morphologies this does little tohelp separate hypotheses of PLE and phyletic gradualismseeing as both predict shifting phenotypic means as aresult of environmental change A stronger indicator ofreduced canalization and subsequent increased develop-mental plasticity is a measure of the morphologicalvariability induced by environmental stress (Hoffmann ampHercus 2000)

The relationship between environmental stress andmorphology is the subject of the Plus ccedila Change model(Sheldon 1997) In short it seeks to explain theobservation that morphological stasis appears the mostcommon response to fluctuations in environmentalconditions in the fossil record Drawing on a deep‐timeperspective it argues that long‐term generalists are thespecies most likely to survive such fluctuations andconsequently predicts gradualistic phenotypic changeunder stable or near‐stable environmental conditionsand phenotypic stasis during periods of strong environ-mental fluctuations

In its general sense this model seemingly suggests theopposite of PLE to manifest in the fossil record Yet itcontains the caveat that upon reaching a critical thresh-old of stress rapid phyletic evolution may be the result(Figure 3) The parallels between this prediction and theprocess of PLE outlined above are readily apparent thismodel suggests that in a strongly fluctuating environmentnatural selection favors those organisms that sufficientlycanalize their development to best track an underlyingphenotypic optimum yet when an extreme environmentis encountered developmental canalization is overcomeand previously cryptic variation is released

The prediction that environmental stressors candecrease developmental stability by overcoming cana-lization and releasing CGV implies that phenotypicvariability can be regarded as a proxy for develop-mental instability (Klingenberg 2019 Webster 2019Willmore Young amp Richtsmeier 2007) Thus thephenotypendashenvironment association of interest for thestudy of PLE in the fossil record is that betweenenvironmental conditions and phenotypic variabilityrather than phenotypic variation

JACKSON | 7 of 15

This distinction is key for separating the predictions ofPLE and phyletic gradualism Variation that is morpho-logical differences between assemblages associated withenvironmental change is insufficient to support PLE forthe reasons discussed above However an associationbetween environmental change and variability that isthe variance in form within assemblages informs usabout their developmental stability and thus permitsinvestigation into the unique prediction of PLE thatenvironmental stress may overcome canalization andinduce the release of CGV with consequent expression ofnovel variants upon which selection can act Theconsequence of this process is an increase in phenotypicvariability coupled with a shift in phenotypic meanassociated with environmental stress a pattern notpredicted by phyletic gradualism

Below I provide two examples of approaches towardsproviding evidence of PLE in the fossil record that usethis association Before discussing them however I firstpresent their necessary methodological framework

13 | FRAMEWORK OF ANALYSIS

The focal point for these investigations is the relation-ship between environmental factors and morphology offossil assemblages Morphology is essentially the onlybiological information available in the fossil record andit is therefore in general of paramount importance tocapture and study morphology in as high resolution aspossible and with careful consideration of potentialtaphonomic biases such as disarticulation of hardparts or later structural deformation (Webster 2019)This is particularly true however when studyingintraspecific patterns of variation such as thoseI will present below Many modern phylogeneticcladistic analyses rely on discretizing variation intocharacters which strongly limits their ability to studyintraspecific variation An alternative approach toanalyzing morphological variation is through the useof morphometric methods which quantify shapes intosets of measurements that can be mathematicallyanalyzed (MacLeod 2017) Such methods includelandmark analyses (Bookstein 1997) eigenshape ana-lysis (MacLeod 2002) and elliptical Fourier analysis(Kuhl amp Giardina 1982) These numerical approachesto describing variation in form translate morphologyinto continuous data and are consequently more suitedto the study of the subtle variation exhibited intraspe-cifically than the use of discrete characters and indeedare a prerequisite to examining fine‐scale patterns ofmorphological variability

While straightforward use of statistical variance isoften employed in studies on developmental instability(eg Imasheva Loeschcke Zhivotovsky amp Lazebny1997) novel methods for quantifying variance specificallystemming from developmental plasticity have seen recentdevelopment (Pertoldi amp Kristensen 2015 Pertoldi et al2014) Coupling shifts in variability to shifts in pheno-typic mean is the key to identifying developmentalplasticity in the fossil record as I will discuss belowThe putative confounding effect of time‐averaging onmorphological variance in fossil specimens has not beensupported by empirical studies Rather it seems thatpaleontological and neontological samples have compar-able levels of variability (Kidwell amp Flessa 1995 BushPowell Arnold Bert amp Daley 2002 Hunt 2004 Hunt2004 Hunt Bell amp Travis 2008 Webster 2019)

To approach the environmental aspect of PLE it isnecessary to identify indications of environmental con-ditions in fossil assemblages Patterns of intraspecificvariation in fossil species across a geographic andputative environmental range are by no means uncom-mon (eg Hopkins amp Webster 2009 Jackson amp Budd2017) Such patterns have been studied through the

FIGURE 3 Sheldonrsquos (1997) Plus ca Change model Stronglyfluctuating environments are predicted to increase selection forstasis until a threshold is reached and a punctuation sensuEldredge and Gould (1972) is generated This model can beinterpreted as predicting selection for increased canalization influctuating environments until a threshold is reached at whichpoint cryptic genetic variation is expressed and phenotypicaccommodation occurs

8 of 15 | JACKSON

assessment of associations between morphology andvarious environmental gradients (eg Cisne ChandleeRabe amp Cohen 1982 Scarponi amp Kowalewski 2004Webber amp Hunda 2007) Indeed a variety of geochemicalproxies are available for the investigation and approx-imation of depositional environments (Webster 2019)such as using trace metals as a proxy for paleoredox andpaleoproductivity (Tribovillard Algeo Lyons amp Riboul-leau 2006) for example However care must be takenwhen interpreting such proxies to minimize environ-mental variation generated by taphonomic processessuch as diagenetic alterations of the rock (Webster 2019)

The relationship between environmental stress anddevelopmental plasticity by proxy of morphologicalvariability is used in the following two approachestowards providing evidence of PLE as a mode ofevolution in the fossil record

14 | EVIDENCE OF PHENOTYPICAND GENETIC ACCOMMODATION

As mentioned above both phyletic gradualism and PLEpredict a shift in morphology following exposure tonovel environments albeit resulting from differentprocesses making it difficult to uncover traces defini-tively of PLE Gradual adaptive evolution wouldmanifest as an assemblage shifting towards a newphenotypic optimum in response to environmentalperturbation Likewise PLE would present a similarpattern yet with the environment both inducing as wellas selecting for novel variants

However PLE predicts a pattern of increasingvariability as the expression of CGV is induced byenvironmental stress which distinguishes it fromphyletic gradualism In this mode of evolution pheno-typic accommodation increases variability such that thephenotype subsequently genetically accommodated or

canalized lies outside its ancestral normal range (Figure4 cf Figure 1) While developmental instability isknown to occur during stress including on the rangemargins of populations (Bridle amp Vines 2007 RiesebergArcher amp Wayne 1999) such dynamics do not predictthe induction of novel variation outside the ancestralrange Consequently shifts in mean morphologycoupled with strongly increasing variability beyond anancestrally normal range can be used as an indicator ofPLE (Figure 4)

This means that the study of morphological varia-bility across environmental gradients offers an oppor-tunity to test specifically for evidence of the phenotypicand genetic accommodation inherent to PLE as de-scribed above

Naturally demonstrating this kind of intraspecificpattern requires a fossil organism or lineage with a highlyresolved stratigraphy sufficient abundance and knownpresence across temporal andor environmental gradients(Schoch 2014) Such research is not uncommon (egHopkins amp Webster 2009 Schoch 2014 Webber ampHunda 2007) yet morphological variability is rarely thefocal point and phenotypic accommodation rarer still

However the Cambrian trilobite‐like arthropod Ag-nostus pisiformis has been demonstrated to exhibitdifferentiated patterns of variation and variability acrossa gradient of dysoxic stress (Jackson amp Budd 2017)implying that environmental conditions may indeed beinducing as well as selecting for phenotypic variants

Further investigation in this framework is needed toproperly test the fossil record for evidence of PLEPatterns of phenotypic variation and chiefly variabilityuncovered by fine‐scale studies of fossil morphology (egDai Zhang Peng amp Yao 2017 Esteve 2012 NeubauerHarzhauser amp Kroh 2013) ought to be considered withinthe framework of PLE Likewise research exploringputative canalization of fossil organisms (eg PimientoTang Zamora Klug amp Saacutenchez‐Villagra 2018 Webster

Mor

phol

ogy

Time

Mor

phol

ogy

Time

(a) (b) EnvironmentalOptimum

Population

FIGURE 4 Illustration of two hypothetical responses of an evolutionary lineage to an environmental perturbation (a) Natural selectioncauses an adaptive shift of morphology towards a new environmental optimum in line with the predictions of phyletic gradualism (b) Theenvironmentally mediated release of cryptic genetic variation increases phenotypic variability facilitating rapid adaptive evolution towards anew optimum as per the process of plasticity‐led evolution Stabilizing selection subsequently leads to canalization around this newoptimum

JACKSON | 9 of 15

2015) could explicitly test for indications of PLE havingoccurred

15 | EVIDENCE OFDEVELOPMENTALLY BIASEDEVOLUTIONARY TRAJECTORIES

Adaptive radiations represent periods of rapid evolu-tion in which ancestral forms of a lineage diversify inenvironment morphology functionality and so forthWhen such patterns feature repeated evolutionarysequences such as in the case of the Anolis lizards ofthe Caribbean (Losos Jackman Larson de Queiroz ampRodriacuteguez‐Schettino 1998) they afford us the oppor-tunity to examine the fossil record for evidence ofanother component of PLE namely whether or notrepeated patterns of phenotypic evolution are some-times better explained by developmental bias thanconvergent selection At the heart of this issue lies therole of the environment in adaptive radiationsTraditionally the environment is seen as facilitatingsuch episodes by providing ecological opportunity todiversifying clades (Stroud amp Losos 2016) However asoutlined above the environment may also act as aninducer of novel variants by interacting with develop-mental processes Erwin (2017) describes these per-spectives as emphasizing ldquoenvironmental pullrdquo andldquodevelopmental pushrdquo respectively

West‐Eberhardrsquos (2003) flexible‐stem model sug-gests that plasticity mediated by the regulatorystructure of an ancestral populationrsquos developmentwill bias repeated independent diversificationsalong similar evolutionary trajectories via the pro-cesses of phenotypic and genetic accommodationThis represents an alternative or complementaryexplanation as to why independently derived lineages

of an ancestral population repeatedly evolve similarphenotypes such as the ldquoecomorphsrdquo of the Anolislizards mentioned above (Losos et al 1998 Williams1972) While this model has been explored in aneontological context (Gibert 2017 Parsons et al2016 Schneider amp Meyer 2017 Wund Baker ClancyGolub amp Foster 2008) it has not yet seen attention ina paleontological setting

Although the identification of adaptive radiations inthe fossil record is not a straightforward endeavor(Lieberman 2012) essentially any sequence of repeateddivergence from an ancestral stock population is suitablefor investigating the flexible‐stem informed model ofparallelism Similarly to the discussion on phenotypicand genetic accommodation above the focal point ofresearch in such systems should be on the variability ofmorphology rather than variation relative environmen-tal gradients If it is indeed phenotypic accommodationthat is the environmentally mediated release of CGVthat directs these repeated evolutionary trajectories weshould predict an increasing degree of phenotypicvariability to coincide with these divergences (Figure 5)This coupling of increasing variability and shiftingvariation is not conventionally anticipated in a patternof repeated parallel divergences that is convergentselection as explained above

The positive identification of such patterns in thefossil record would thus provide distinctive support forthe underlying processes of PLE both in terms of thesuggested environmentally mediated release of CGV aswell as the hypothesized bias exercised by the regulatorynature of development Research towards uncoveringsuch support should be directed towards reassessing thephenotypendashenvironment associations of studied fossiladaptive radiations (eg Abe amp Lieberman 2012 NeigeDera amp Dommergues 2013 Neubauer et al 2013) in thecontext of PLE

Mor

phol

ogy

Time

EnvironmentalOptimum

Ancestral Population

Diverging Population

FIGURE 5 Repeated divergences from an ancestral or stock population The bias native to the ancestral developmental network directsmorphological change along similar evolutionary trajectories as the successive independently diverging populations encounter the newenvironment An increase in morphological variability coupled with such parallelism of phenotypic mean change is predicted by the releaseof cryptic genetic variation in plasticity‐led evolution but not by the conventional process of convergent evolution resulting from similar andrepeated selective pressures

10 of 15 | JACKSON

16 | CONCLUDING REMARKS

Viewing development as constructive is a promisingconceptual framework for enriching our understandingof the processes and patterns of evolution It provides thescaffolding for properly addressing issues of centralimportance such as the role of developmental biases indirecting evolutionary change However the majority ofefforts towards this end are conducted on extantorganisms with minimal reference to paleontologicaldata This is unfortunate as the fossil record is afantastically rich source of information that is yet to befully explored in this intriguing and stimulating contextWhile microevolutionary concepts such as developmentalbias are challenging to approach in the fossil record Iargue that it is manifest in paleontological materials inthe form of macroevolutionary patterns concordant withthe predictions of plasticity‐led evolution Indeed suchpredictions recapitulate earlier theories of quantumevolution and punctuated equilibrium

These predictions are specifically an evolutionarymodel consisting of two major parts (a) stabilizingselection during periods of environmental stabilityleading to a reduction in phenotypic variation but anincrease in canalization CGV and hence evolutionarycapacitance and (b) episodes of environmentally inducedplasticity facilitating rapid evolution via the expression ofpreviously cryptic variation

Research aiming to uncover support for PLE in thefossil record should be directed towards the environ-mentndashphenotype association of fossil assemblages Spe-cifically towards patterns of correlation between envir-onmental stress and phenotypic variability as a proxy fordevelopmental instability These may be manifest alongsingle lineages or throughout adaptive radiations Thefossil record offers enormous potential to provide datathat when viewed through the lens of PLE providesbiologically meaningful information about evolution interms of both patterns and processes

ACKNOWLEDGMENTS

I would like to thank Armin Moczek for inviting me tocontribute to this special issue on developmental bias inevolution My thanks to two anonymous reviewers whoprovided thoughtful constructive and immensely helpfulcomments that greatly improved this paper I am gratefulto Graham Budd for our many discussions on this topicover the years I am thankful to the Uller group at LundUniversity for formative discussions particularly TobiasUller who also provided valuable comments on an earlydraft of this paper My thoughts presented here weregreatly informed by the SFI workshop ldquoDevelopmental

Bias and Evolutionrdquo and I direct my gratitude towards itsorganizers and participants This study was funded by theKnut and Alice Wallenberg Foundation grant to TobiasUller (KAW 20120155)

CONFLICT OF INTEREST

The author declares that there is no conflict of interest

ORCID

Illiam S C Jackson httporcidorg0000-0002-7948-2860

REFERENCES

Abe F R amp Lieberman B S (2012) Quantifying morphologicalchange during an evolutionary radiation of Devonian trilobitesPaleobiology 38 292ndash307 httpsdoiorg101666100471

Arthur W (2004) The effect of development on the directionof evolution Toward a twenty‐first century consensusEvolution amp Development 6(4) 282ndash288 httpsdoiorg101111j1525‐142X200404033x

Badyaev A V (2005) Stress‐induced variation in evolution Frombehavioural plasticity to genetic assimilation Proceedings of theRoyal Society B Biological Sciences 272 877ndash886 httpsdoiorg101098rspb20043045

Badyaev A V (2009) Evolutionary significance of phenotypicaccommodation in novel environments An empirical test of theBaldwin effect Philosophical Transactions of the Royal SocietyB Biological Sciences 364 1125ndash1141 httpsdoiorg101098rstb20080285

Baedke J (2018) Organism where art thou Old and new challengesfor organism‐centered biology Journal of the History of Biology52 293ndash324 httpsdoiorg101007s10739‐018‐9549‐4

Bookstein F L (1997) Morphometric tools for landmark dataGeometry and biology Cambridge Cambridge University Press

Brakefield P M (2011) Evo‐devo and accounting for Darwinsendless forms Philosophical Transactions of the Royal Society BBiological Sciences 366(1574) 2069ndash2075 httpsdoiorg101098rstb20110007

Bridle J R amp Vines T H (2007) Limits to evolution at rangemargins When and why does adaptation fail Trends in Ecologyamp Evolution 22(3) 140ndash147 httpsdoiorg101016jtree200611002

Bush A M Powell M G Arnold W S Bert T M amp Daley G M(2002) Time‐averaging evolution and morphologic variationPaleobiology 28 9ndash25 httpsdoiorg1016660094‐8373(2002)0283C0009TAEAMV3E20CO2

Carroll S B (2005) Evolution at two levels On genes and form PLOSBiology 3 e245 httpsdoiorg101371journalpbio0030245

Carroll S B (2008) Evo‐devo and an expanding evolutionarysynthesis A genetic theory of morphological evolution Cell134 25ndash36 httpsdoiorg101016jcell200806030

Charlesworth B Lande R amp Slatkin M (1982) A neo‐Darwiniancommentary on macroevolution Evolution 36 474ndash498 httpsdoiorg1023072408095

JACKSON | 11 of 15

Novel or extreme environmental conditions mayreveal this capacitance (Badyaev 2005 Flatt 2005) byexpressing novel developmental pathways or by down-stream effects of mutual adjustment of different parts ofthe phenotype a process known as phenotypic accom-modation (Figure 1c Badyaev 2009 Laland et al 2015Schwab et al 2019 West‐Eberhard 2005) Environ-mental stress can thus be understood to simultaneouslyinduce and select for novel phenotypes a process thatwould facilitate rapid evolutionary change (Gibson ampDworkin 2004 Levis amp Pfennig 2016 McGuigan ampSgrograve 2009 Paaby amp Rockman 2014 Schlichting 2008Schwab et al 2019)

Due to the internal architecture of developmentalsystems which generates integrated modules (GoswamiBinder Meachen amp OrsquoKeefe 2015 Schlichting 1989Wagner amp Zhang 2011) random genetic mutations maystill lead to coordinated phenotypic responses (Jablonski2017 Moczek et al 2011 Uller et al 2018 West‐Eberhard 2003) For this reason the effects of thepreviously cryptic genetic variation may in fact causedirected phenotypic change biased towards functionallyintegrated phenotypes (Gerhart amp Kirschner 2007Masel 2006 Wagner 2011 Watson amp Szathmaacutery 2016Watson Wagner Pavlicev Weinreich amp Mills 2014)Taken together and in the context of PLE this impliesthat accumulated CGV revealed by environmental stressmay be biased towards experimental functionality

Initially this constitutes a pulse of phenotypic changewithout genotypic change (Noble 2015 West‐Eberhard2003) However when CGV is thus revealed its pheno-typic manifestation nevertheless becomes visible tonatural selection which can select for or against it Moreimportantly selection indirectly acts on the develop-mental system itself rendering its regulation either moreor less sensitive to environmental conditions (Moczek2008 Schlichting amp Wund 2014 West‐Eberhard 2005Wund 2012) This subsequent process is known asgenetic accommodation (Lande 2009 West‐Eberhard2003) If nature selects strongly for a novel phenotypethat is a newly induced variant that is preadapted to itsenvironment then the process of stabilizing selectionstarts over and development can begin to canalize oncemore (Figure 1d) If one such novel phenotype issubjected to very powerful selection it may lose itsplasticity entirely and is then said to have beengenetically assimilated (Pigliucci Murren amp Schlichting2006 Waddington 1961 West‐Eberhard 2003)

Given that the modular structure of gene regulatorynetworks generates patterns in which parts of develop-ment interact with the environment whereas othersdo not (Iwasaki et al 2013 Schneider amp Meyer 2017)those parts that do interact with and are intermittently

mediated by environmental factors become the centersof evolutionary change and evolution can proceedpredominantly by modification of environmentally regu-lated parts of the developmental system with consequentphenotypic diversification (Schneider amp Meyer 2017Uller et al 2018 Wund 2012) Thus this processrepeated in isolated populations is likely to directevolution along particular trajectories

6 | PLASTICITY ‐LEDEVOLUTION AS A MODE

The cyclical accumulation and release of CGV duringperiods of stabilizing selection and environmental stressrespectively thus describes the operation of PLE This issimilar to Badyaevrsquos (2009) conceptualization of ldquoevolu-tionary diversifications and maintenance of adaptationsprobably operating in alternation between the emer-gence of novel developmental variation and stabilizationof locally appropriate organismndashenvironment associa-tions by natural selectionrdquo and represents an excitinghypothetical mode of evolution Indeed the role ofgenetic assimilation in evolution that is the adoption ofgenetic management of initially environmentally in-duced phenotypes continues to demand attention(Ehrenreich amp Pfennig 2016 Schneider amp Meyer2017 Schwab et al 2019)

7 | QUANTUM EVOLUTION

As a mode of evolution this model of PLE constitutes apattern of periods of stasis during which capacitance isaccrued followed by rapid bursts of change as thiscapacitance is expended In a macroevolutionary contextthis is not without historical analogues Indeed thedistinction between stasis and burst is captured in thecontrast between Simpsonrsquos (1944) phyletic and quantumevolution (Figure 2) two of his three modes of evolutionthe third being speciation He defines quantum evolutionas the ldquorelatively rapid shift of a biotic population indisequilibrium to an equilibrium distinctly unlike anancestral conditionrdquo while phyletic evolution is theldquosustained directional (but not necessarily rectilinear)shiftrdquo more commonly observed in the fossil record

Simpson (1944) considered these two modes ofevolution as qualitatively different to his mind theyfulfill different functions and presumably operate bydifferent mechanisms Phyletic evolution manages theadaptation of species to their adaptive zone whereasquantum evolution is a much more radical processcapable of generating rapid changes and accommodating

4 of 15 | JACKSON

jumps across inadaptive zones to new adaptive zones Inother words quantum evolution was a suggested solutionto the contemporary problem of how a population mightcross the valleys of Wrightrsquos adaptive landscapes (Wright1932) given that evolution proceeds through the gradualaccumulation of small changes (cf Zheng et al 2019) Tothis end Simpson (1944) writes that ldquopreadaptation [is] anecessary element [of quantum evolution and is] anothermarked distinction between quantum evolution andphyletic evolution which is mostly adaptive (andpostadaptive)rdquo Furthermore quantum evolution pro-vided a biological explanation for the long‐observeddiscontinuities in the fossil record they were not merelyartifacts of incompleteness but may in fact be indicativeof extremely rapid phenotypic transitions

According to Simpson (1944) quantum evolutionprogressed through three phases

ldquo(1) An inadaptive phase in which the groupin question loses the equilibrium of itsancestors or collaterals

(2) A preadaptive phase in which there isgreat selection pressure and the group movestoward a new equilibrium

(3) An adaptive phase in which the newequilibrium is reachedrdquo

There is a striking concordance between these phasesand (a) environmental induction of CGV release that isphenotypic accommodation (b) corresponding environ-mental selection for novel variants that is geneticaccommodation and potential assimilation and (c)stabilizing selection that is canalization of development

Yet at the time the best candidate for a mechanismby which quantum evolution might proceed was genetic

drift During the ldquohardeningrdquo of the modern synthesisas evolutionary explanations became increasingly re-stricted to selection acting on gradual allelic change(Gould 1982) the role of genetic drift seemed lesssignificant leading to Simpson changing his views onquantum evolution and ceasing to see it is a differentkind of evolution to phyletic evolution (Gould 1986) Helater wrote that quantum evolution was ldquonot a differentsort of evolution from phyletic evolution or even adistinctly different element of the total phylogeneticpattern It is a special more or less extreme and limitingcase of phyletic evolutionrdquo (Simpson 1953) In otherwords quantum evolution became simply phyleticevolution sped up and the norm mode of evolutionconsequently became phyletic gradualism (Eldredge ampGould 1972 Gould 1980)

8 | PUNCTUATED EQUILIBRIUM

It was against this backdrop that Eldredge and Gouldformulated their theory of punctuated equilibrium(Eldredge amp Gould 1972 Gould amp Eldredge 19771993) Similar to Simpsonrsquos (1944) quantum evolutionit posits that discontinuities in the fossil record need notnecessarily be indicative of incompleteness Ratherphenotypic evolution may be as intermittently rapidthat is punctuated as fossil data suggest They argue thatldquothe remarkable stasis exhibited by most species duringmillions of years is ignored (as no data)rdquo (Gould ampEldredge 1993) In other words stasis is taken as a lackof evolution and contrasted with directional phenotypicchange through time The assumption that discontinu-ities are artifacts of incompleteness implies that phyleticgradualism is synonymous with evolution

Gould and Eldredge (1993) write that we insteadought to ldquoregard stasis and discontinuity as an

FIGURE 2 Simpsonrsquos (1944) comparison of phyletic and quantum evolution Phyletic evolution is the selective tracking of a shiftingadaptive zone through time with subsequent directional phenotypic gradualism as a result Quantum evolution on the other hand is a rapidevolutionary transition between adaptive zones Note the preadaptive variation manifest in the ancestral population immediately beforequantum evolution

JACKSON | 5 of 15

expression of how evolution works when translatedinto geological timerdquo and argue that phenotypicevolution is more closely tied to speciation thanpreviously imagined Indeed studies on patterns ofchange in the fossil record have confirmed that stasis isa far more common observation than phyletic gradu-alism (Hunt 2004 2007 Hunt amp Rabosky 2014) withstasis in this sense more accurately regarded asdirectionless or nonaccumulating morphological fluc-tuations rather than true stagnance (Gould 1982Gould 2002 Voje 2016) Furthermore adaptivephenotypic change may occur so rapidly that it maypass by the record (Hunt 2010) Yet while the patterndescribed by punctuated equilibrium is thereforebroadly supported by observation the mechanism bywhich this pattern arises particularly the rapidtransitions of form remains unclear (Futuyma 2015)Around the time of its inception and early receptionthe theory of punctuated equilibrium was met withstrong critique (eg Charlesworth Lande amp Slatkin1982 Gingerich 1984 Kirkpatrick 1982) generating adebate that continues to this day (eg Lieberman ampEldredge 2014 Pennell Harmon amp Uyeda 2014)

9 | FUNDAMENTAL DEBATE

It seems that the same point of contention that ledSimpson to dilute his concept of quantum evolution alsoframes the debate on the theory of punctuated equili-brium as well as the controversy surrounding phenotypicand genetic accommodation namely the tension be-tween the understanding of development as programmedor constructive (Laland et al 2015) itself a manifestationof the differing perspectives of the gene‐centered andorganism‐centered views of evolution (Baedke 2018Nicholson 2014) Clearly the same currents of thoughtare manifest in each of these instances yet the contentionremains unresolved

It is tempting to argue that Badyaevrsquos (2009)alternating ldquoemergence of novel developmental varia-tionrdquo and ldquostabilization of locally appropriate organis-mndashenvironment associations by natural selectionrdquo arequalitatively different processes the former approxi-mated by quantum evolution punctuated equilibriaand plasticity‐led evolution and the latter by phyleticevolution stasis and gradualism Indeed this would bean approach informed by the distinction betweenorigins and distributional explanations (Erwin 2017)Yet this does not make sense in terms of our under-standing of development

From a developmental perspective gradualism andPLE are not qualitatively different processes Both are

governed by a regulatory network that facilitates varia-tion and buffers genetic and environmental fluctuationstowards functionality (Uller et al 2018) The keydifference is simply the phenotypic accommodation ofPLE which may be exaggerated by the release ofcapacitance in the form of CGV This facilitates thedistinctive rapidity described by quantum evolution(Simpson 1944) and punctuated equilibrium (Eldredgeamp Gould 1972) Thus the difference is quantitative notqualitative a matter of thresholds and amounts ratherthan kinds

10 | MACROEVOLUTIONARYPATTERNS OF PLASTICITY ‐LEDEVOLUTION

PLE as the evolutionary process described abovepredicts in my view in short two major features inthe fossil record (a) stabilizing selection duringperiods of environmental stability leading to decreas-ing phenotypic variation but increasing evolutionarycapacitance in the form of cryptic genetic variationand (b) episodes of environmentally induced pheno-typic plasticity facilitating rapid bursts of evolution asthis capacitance is expended

To be clear in this framework plasticity‐led evolu-tion (PLE) that is the accumulation and release ofcryptic genetic variation (CGV) is a putative generatorof those patterns in the fossil record previouslydescribed as quantum evolution (Simpson 1944) andpunctuated equilibrium (Eldredge amp Gould 1972)Periods of phyletic evolution or stasis are understoodas stabilizing selection acting on an organism Quantumevolution and the rapid shifts of punctuated equilibriaare hypothesized to be facilitated by phenotypic andgenetic accommodation This is not to say that evolutioncannot occur through phyletic gradualism that isstepwise accumulation of genetic change nor to saythat the fossil record is not incomplete in parts I onlymean to say that PLE provides an alternative explana-tion for the phenomena mentioned above While thetheory of punctuated equilibrium is associated withspeciation and quantum evolution is traditionallydiscussed in terms of anagenetic evolution despiteSimpson (1944) himself writing that it ldquomay be involvedin either speciation or phyletic evolutionrdquo and thatldquocertain patterns within those modes intergrade withquantum evolutionrdquo this distinction holds little sig-nificance for this model of PLE the environmentallymediated accumulation and release of CGV can occur ineither instance and whether or not speciation occurs isinconsequential

6 of 15 | JACKSON

11 | EVIDENCE OF PLASTICITY‐LEDEVOLUTION IN THE FOSSIL RECORD

When searching for evidence of PLE in the fossil recordit is insufficient to simply point to patterns of stasis ordiscontinuity As Gould and Eldredge (1993) point outldquoExamples of stasis alone [hellip] and simple abruptreplacement although conforming to expectations ofpunctuated equilibrium are not direct evidence for ourmechanism for stasis might just be a lull in anageneticgradualism [hellip] and replacement might represent rapidtransformation without branching or migration of adistant (phyletic or geographic) relative rather thanevolution in siturdquo Likewise discontinuities may indeedhave geological rather than biological explanationscaused by erosion of fossil‐bearing strata or a hiatus ina sedimentary deposition for example Furthermore theprocesses of phenotypic and genetic accommodation arechallenging to study in the fossil record As Pigliucciand Murren (2003) caution us ldquogiven the hypothesisthat [genetic] assimilation can occur within a fewgenerations it ironically may be too fast for us tocatchrdquo Indeed the rapidity of phenotypic transitionsreported by Hunt (2010) points to a similar problemThese challenges present a problem to the study of PLEin the fossil record

Beyond these challenges a further difficulty is thatthe central macroevolutionary prediction of PLE issuperficially identical to that of phyletic gradualismPLE predicts that novel environments lead to evolu-tionary change mediated by phenotypic and geneticaccommodation Phyletic gradualism predicts that novelenvironments lead to evolutionary change mediated bynatural selection on gradually accumulating mutationsThus the observation of phenotypic change accompa-nying environmental change is insufficient to supportPLE over phyletic gradualism without access to geneticmaterial and consequently the underlying processes weare left unable to determine which explanation is morelikely In general this is a difficulty faced wheninvestigating the fossil record in search for support forPLE macroevolutionary patterns have historically pro-ven an inadequate source of discrimination betweengenetic variation and plasticity (Hopkins 2014 Schoch2014 Webster 2019)

While the fact that the process of PLE is coherent withthe structure of the fossil record should not be ignored asinsignificant it remains the case that to provide definitivesupport for PLE as a mode of evolution we need to directour attention towards identifying evidence that distin-guishes it from phyletic gradualism Approaches towardsdifferentiating these processes rely on the study ofprocesses instead of patterns In other words they focus

on providing support for the underlying mechanisms ofPLE rather than the broad‐scale outcome

12 | ENVIRONMENTAL STRESSAND PHENOTYPIC VARIABILITY

Specifically the key feature distinguishing PLE fromphyletic gradualism is the environmentally mediateddevelopmental plasticity that facilitates rapid phenotypicchange While such change can be measured in thedifference between mean morphologies this does little tohelp separate hypotheses of PLE and phyletic gradualismseeing as both predict shifting phenotypic means as aresult of environmental change A stronger indicator ofreduced canalization and subsequent increased develop-mental plasticity is a measure of the morphologicalvariability induced by environmental stress (Hoffmann ampHercus 2000)

The relationship between environmental stress andmorphology is the subject of the Plus ccedila Change model(Sheldon 1997) In short it seeks to explain theobservation that morphological stasis appears the mostcommon response to fluctuations in environmentalconditions in the fossil record Drawing on a deep‐timeperspective it argues that long‐term generalists are thespecies most likely to survive such fluctuations andconsequently predicts gradualistic phenotypic changeunder stable or near‐stable environmental conditionsand phenotypic stasis during periods of strong environ-mental fluctuations

In its general sense this model seemingly suggests theopposite of PLE to manifest in the fossil record Yet itcontains the caveat that upon reaching a critical thresh-old of stress rapid phyletic evolution may be the result(Figure 3) The parallels between this prediction and theprocess of PLE outlined above are readily apparent thismodel suggests that in a strongly fluctuating environmentnatural selection favors those organisms that sufficientlycanalize their development to best track an underlyingphenotypic optimum yet when an extreme environmentis encountered developmental canalization is overcomeand previously cryptic variation is released

The prediction that environmental stressors candecrease developmental stability by overcoming cana-lization and releasing CGV implies that phenotypicvariability can be regarded as a proxy for develop-mental instability (Klingenberg 2019 Webster 2019Willmore Young amp Richtsmeier 2007) Thus thephenotypendashenvironment association of interest for thestudy of PLE in the fossil record is that betweenenvironmental conditions and phenotypic variabilityrather than phenotypic variation

JACKSON | 7 of 15

This distinction is key for separating the predictions ofPLE and phyletic gradualism Variation that is morpho-logical differences between assemblages associated withenvironmental change is insufficient to support PLE forthe reasons discussed above However an associationbetween environmental change and variability that isthe variance in form within assemblages informs usabout their developmental stability and thus permitsinvestigation into the unique prediction of PLE thatenvironmental stress may overcome canalization andinduce the release of CGV with consequent expression ofnovel variants upon which selection can act Theconsequence of this process is an increase in phenotypicvariability coupled with a shift in phenotypic meanassociated with environmental stress a pattern notpredicted by phyletic gradualism

Below I provide two examples of approaches towardsproviding evidence of PLE in the fossil record that usethis association Before discussing them however I firstpresent their necessary methodological framework

13 | FRAMEWORK OF ANALYSIS

The focal point for these investigations is the relation-ship between environmental factors and morphology offossil assemblages Morphology is essentially the onlybiological information available in the fossil record andit is therefore in general of paramount importance tocapture and study morphology in as high resolution aspossible and with careful consideration of potentialtaphonomic biases such as disarticulation of hardparts or later structural deformation (Webster 2019)This is particularly true however when studyingintraspecific patterns of variation such as thoseI will present below Many modern phylogeneticcladistic analyses rely on discretizing variation intocharacters which strongly limits their ability to studyintraspecific variation An alternative approach toanalyzing morphological variation is through the useof morphometric methods which quantify shapes intosets of measurements that can be mathematicallyanalyzed (MacLeod 2017) Such methods includelandmark analyses (Bookstein 1997) eigenshape ana-lysis (MacLeod 2002) and elliptical Fourier analysis(Kuhl amp Giardina 1982) These numerical approachesto describing variation in form translate morphologyinto continuous data and are consequently more suitedto the study of the subtle variation exhibited intraspe-cifically than the use of discrete characters and indeedare a prerequisite to examining fine‐scale patterns ofmorphological variability

While straightforward use of statistical variance isoften employed in studies on developmental instability(eg Imasheva Loeschcke Zhivotovsky amp Lazebny1997) novel methods for quantifying variance specificallystemming from developmental plasticity have seen recentdevelopment (Pertoldi amp Kristensen 2015 Pertoldi et al2014) Coupling shifts in variability to shifts in pheno-typic mean is the key to identifying developmentalplasticity in the fossil record as I will discuss belowThe putative confounding effect of time‐averaging onmorphological variance in fossil specimens has not beensupported by empirical studies Rather it seems thatpaleontological and neontological samples have compar-able levels of variability (Kidwell amp Flessa 1995 BushPowell Arnold Bert amp Daley 2002 Hunt 2004 Hunt2004 Hunt Bell amp Travis 2008 Webster 2019)

To approach the environmental aspect of PLE it isnecessary to identify indications of environmental con-ditions in fossil assemblages Patterns of intraspecificvariation in fossil species across a geographic andputative environmental range are by no means uncom-mon (eg Hopkins amp Webster 2009 Jackson amp Budd2017) Such patterns have been studied through the

FIGURE 3 Sheldonrsquos (1997) Plus ca Change model Stronglyfluctuating environments are predicted to increase selection forstasis until a threshold is reached and a punctuation sensuEldredge and Gould (1972) is generated This model can beinterpreted as predicting selection for increased canalization influctuating environments until a threshold is reached at whichpoint cryptic genetic variation is expressed and phenotypicaccommodation occurs

8 of 15 | JACKSON

assessment of associations between morphology andvarious environmental gradients (eg Cisne ChandleeRabe amp Cohen 1982 Scarponi amp Kowalewski 2004Webber amp Hunda 2007) Indeed a variety of geochemicalproxies are available for the investigation and approx-imation of depositional environments (Webster 2019)such as using trace metals as a proxy for paleoredox andpaleoproductivity (Tribovillard Algeo Lyons amp Riboul-leau 2006) for example However care must be takenwhen interpreting such proxies to minimize environ-mental variation generated by taphonomic processessuch as diagenetic alterations of the rock (Webster 2019)

The relationship between environmental stress anddevelopmental plasticity by proxy of morphologicalvariability is used in the following two approachestowards providing evidence of PLE as a mode ofevolution in the fossil record

14 | EVIDENCE OF PHENOTYPICAND GENETIC ACCOMMODATION

As mentioned above both phyletic gradualism and PLEpredict a shift in morphology following exposure tonovel environments albeit resulting from differentprocesses making it difficult to uncover traces defini-tively of PLE Gradual adaptive evolution wouldmanifest as an assemblage shifting towards a newphenotypic optimum in response to environmentalperturbation Likewise PLE would present a similarpattern yet with the environment both inducing as wellas selecting for novel variants

However PLE predicts a pattern of increasingvariability as the expression of CGV is induced byenvironmental stress which distinguishes it fromphyletic gradualism In this mode of evolution pheno-typic accommodation increases variability such that thephenotype subsequently genetically accommodated or

canalized lies outside its ancestral normal range (Figure4 cf Figure 1) While developmental instability isknown to occur during stress including on the rangemargins of populations (Bridle amp Vines 2007 RiesebergArcher amp Wayne 1999) such dynamics do not predictthe induction of novel variation outside the ancestralrange Consequently shifts in mean morphologycoupled with strongly increasing variability beyond anancestrally normal range can be used as an indicator ofPLE (Figure 4)

This means that the study of morphological varia-bility across environmental gradients offers an oppor-tunity to test specifically for evidence of the phenotypicand genetic accommodation inherent to PLE as de-scribed above

Naturally demonstrating this kind of intraspecificpattern requires a fossil organism or lineage with a highlyresolved stratigraphy sufficient abundance and knownpresence across temporal andor environmental gradients(Schoch 2014) Such research is not uncommon (egHopkins amp Webster 2009 Schoch 2014 Webber ampHunda 2007) yet morphological variability is rarely thefocal point and phenotypic accommodation rarer still

However the Cambrian trilobite‐like arthropod Ag-nostus pisiformis has been demonstrated to exhibitdifferentiated patterns of variation and variability acrossa gradient of dysoxic stress (Jackson amp Budd 2017)implying that environmental conditions may indeed beinducing as well as selecting for phenotypic variants

Further investigation in this framework is needed toproperly test the fossil record for evidence of PLEPatterns of phenotypic variation and chiefly variabilityuncovered by fine‐scale studies of fossil morphology (egDai Zhang Peng amp Yao 2017 Esteve 2012 NeubauerHarzhauser amp Kroh 2013) ought to be considered withinthe framework of PLE Likewise research exploringputative canalization of fossil organisms (eg PimientoTang Zamora Klug amp Saacutenchez‐Villagra 2018 Webster

Mor

phol

ogy

Time

Mor

phol

ogy

Time

(a) (b) EnvironmentalOptimum

Population

FIGURE 4 Illustration of two hypothetical responses of an evolutionary lineage to an environmental perturbation (a) Natural selectioncauses an adaptive shift of morphology towards a new environmental optimum in line with the predictions of phyletic gradualism (b) Theenvironmentally mediated release of cryptic genetic variation increases phenotypic variability facilitating rapid adaptive evolution towards anew optimum as per the process of plasticity‐led evolution Stabilizing selection subsequently leads to canalization around this newoptimum

JACKSON | 9 of 15

2015) could explicitly test for indications of PLE havingoccurred

15 | EVIDENCE OFDEVELOPMENTALLY BIASEDEVOLUTIONARY TRAJECTORIES

Adaptive radiations represent periods of rapid evolu-tion in which ancestral forms of a lineage diversify inenvironment morphology functionality and so forthWhen such patterns feature repeated evolutionarysequences such as in the case of the Anolis lizards ofthe Caribbean (Losos Jackman Larson de Queiroz ampRodriacuteguez‐Schettino 1998) they afford us the oppor-tunity to examine the fossil record for evidence ofanother component of PLE namely whether or notrepeated patterns of phenotypic evolution are some-times better explained by developmental bias thanconvergent selection At the heart of this issue lies therole of the environment in adaptive radiationsTraditionally the environment is seen as facilitatingsuch episodes by providing ecological opportunity todiversifying clades (Stroud amp Losos 2016) However asoutlined above the environment may also act as aninducer of novel variants by interacting with develop-mental processes Erwin (2017) describes these per-spectives as emphasizing ldquoenvironmental pullrdquo andldquodevelopmental pushrdquo respectively

West‐Eberhardrsquos (2003) flexible‐stem model sug-gests that plasticity mediated by the regulatorystructure of an ancestral populationrsquos developmentwill bias repeated independent diversificationsalong similar evolutionary trajectories via the pro-cesses of phenotypic and genetic accommodationThis represents an alternative or complementaryexplanation as to why independently derived lineages

of an ancestral population repeatedly evolve similarphenotypes such as the ldquoecomorphsrdquo of the Anolislizards mentioned above (Losos et al 1998 Williams1972) While this model has been explored in aneontological context (Gibert 2017 Parsons et al2016 Schneider amp Meyer 2017 Wund Baker ClancyGolub amp Foster 2008) it has not yet seen attention ina paleontological setting

Although the identification of adaptive radiations inthe fossil record is not a straightforward endeavor(Lieberman 2012) essentially any sequence of repeateddivergence from an ancestral stock population is suitablefor investigating the flexible‐stem informed model ofparallelism Similarly to the discussion on phenotypicand genetic accommodation above the focal point ofresearch in such systems should be on the variability ofmorphology rather than variation relative environmen-tal gradients If it is indeed phenotypic accommodationthat is the environmentally mediated release of CGVthat directs these repeated evolutionary trajectories weshould predict an increasing degree of phenotypicvariability to coincide with these divergences (Figure 5)This coupling of increasing variability and shiftingvariation is not conventionally anticipated in a patternof repeated parallel divergences that is convergentselection as explained above

The positive identification of such patterns in thefossil record would thus provide distinctive support forthe underlying processes of PLE both in terms of thesuggested environmentally mediated release of CGV aswell as the hypothesized bias exercised by the regulatorynature of development Research towards uncoveringsuch support should be directed towards reassessing thephenotypendashenvironment associations of studied fossiladaptive radiations (eg Abe amp Lieberman 2012 NeigeDera amp Dommergues 2013 Neubauer et al 2013) in thecontext of PLE

Mor

phol

ogy

Time

EnvironmentalOptimum

Ancestral Population

Diverging Population

FIGURE 5 Repeated divergences from an ancestral or stock population The bias native to the ancestral developmental network directsmorphological change along similar evolutionary trajectories as the successive independently diverging populations encounter the newenvironment An increase in morphological variability coupled with such parallelism of phenotypic mean change is predicted by the releaseof cryptic genetic variation in plasticity‐led evolution but not by the conventional process of convergent evolution resulting from similar andrepeated selective pressures

10 of 15 | JACKSON

16 | CONCLUDING REMARKS

Viewing development as constructive is a promisingconceptual framework for enriching our understandingof the processes and patterns of evolution It provides thescaffolding for properly addressing issues of centralimportance such as the role of developmental biases indirecting evolutionary change However the majority ofefforts towards this end are conducted on extantorganisms with minimal reference to paleontologicaldata This is unfortunate as the fossil record is afantastically rich source of information that is yet to befully explored in this intriguing and stimulating contextWhile microevolutionary concepts such as developmentalbias are challenging to approach in the fossil record Iargue that it is manifest in paleontological materials inthe form of macroevolutionary patterns concordant withthe predictions of plasticity‐led evolution Indeed suchpredictions recapitulate earlier theories of quantumevolution and punctuated equilibrium

These predictions are specifically an evolutionarymodel consisting of two major parts (a) stabilizingselection during periods of environmental stabilityleading to a reduction in phenotypic variation but anincrease in canalization CGV and hence evolutionarycapacitance and (b) episodes of environmentally inducedplasticity facilitating rapid evolution via the expression ofpreviously cryptic variation

Research aiming to uncover support for PLE in thefossil record should be directed towards the environ-mentndashphenotype association of fossil assemblages Spe-cifically towards patterns of correlation between envir-onmental stress and phenotypic variability as a proxy fordevelopmental instability These may be manifest alongsingle lineages or throughout adaptive radiations Thefossil record offers enormous potential to provide datathat when viewed through the lens of PLE providesbiologically meaningful information about evolution interms of both patterns and processes

ACKNOWLEDGMENTS

I would like to thank Armin Moczek for inviting me tocontribute to this special issue on developmental bias inevolution My thanks to two anonymous reviewers whoprovided thoughtful constructive and immensely helpfulcomments that greatly improved this paper I am gratefulto Graham Budd for our many discussions on this topicover the years I am thankful to the Uller group at LundUniversity for formative discussions particularly TobiasUller who also provided valuable comments on an earlydraft of this paper My thoughts presented here weregreatly informed by the SFI workshop ldquoDevelopmental

Bias and Evolutionrdquo and I direct my gratitude towards itsorganizers and participants This study was funded by theKnut and Alice Wallenberg Foundation grant to TobiasUller (KAW 20120155)

CONFLICT OF INTEREST

The author declares that there is no conflict of interest

ORCID

Illiam S C Jackson httporcidorg0000-0002-7948-2860

REFERENCES

Abe F R amp Lieberman B S (2012) Quantifying morphologicalchange during an evolutionary radiation of Devonian trilobitesPaleobiology 38 292ndash307 httpsdoiorg101666100471

Arthur W (2004) The effect of development on the directionof evolution Toward a twenty‐first century consensusEvolution amp Development 6(4) 282ndash288 httpsdoiorg101111j1525‐142X200404033x

Badyaev A V (2005) Stress‐induced variation in evolution Frombehavioural plasticity to genetic assimilation Proceedings of theRoyal Society B Biological Sciences 272 877ndash886 httpsdoiorg101098rspb20043045

Badyaev A V (2009) Evolutionary significance of phenotypicaccommodation in novel environments An empirical test of theBaldwin effect Philosophical Transactions of the Royal SocietyB Biological Sciences 364 1125ndash1141 httpsdoiorg101098rstb20080285

Baedke J (2018) Organism where art thou Old and new challengesfor organism‐centered biology Journal of the History of Biology52 293ndash324 httpsdoiorg101007s10739‐018‐9549‐4

Bookstein F L (1997) Morphometric tools for landmark dataGeometry and biology Cambridge Cambridge University Press

Brakefield P M (2011) Evo‐devo and accounting for Darwinsendless forms Philosophical Transactions of the Royal Society BBiological Sciences 366(1574) 2069ndash2075 httpsdoiorg101098rstb20110007

Bridle J R amp Vines T H (2007) Limits to evolution at rangemargins When and why does adaptation fail Trends in Ecologyamp Evolution 22(3) 140ndash147 httpsdoiorg101016jtree200611002

Bush A M Powell M G Arnold W S Bert T M amp Daley G M(2002) Time‐averaging evolution and morphologic variationPaleobiology 28 9ndash25 httpsdoiorg1016660094‐8373(2002)0283C0009TAEAMV3E20CO2

Carroll S B (2005) Evolution at two levels On genes and form PLOSBiology 3 e245 httpsdoiorg101371journalpbio0030245

Carroll S B (2008) Evo‐devo and an expanding evolutionarysynthesis A genetic theory of morphological evolution Cell134 25ndash36 httpsdoiorg101016jcell200806030

Charlesworth B Lande R amp Slatkin M (1982) A neo‐Darwiniancommentary on macroevolution Evolution 36 474ndash498 httpsdoiorg1023072408095

JACKSON | 11 of 15

jumps across inadaptive zones to new adaptive zones Inother words quantum evolution was a suggested solutionto the contemporary problem of how a population mightcross the valleys of Wrightrsquos adaptive landscapes (Wright1932) given that evolution proceeds through the gradualaccumulation of small changes (cf Zheng et al 2019) Tothis end Simpson (1944) writes that ldquopreadaptation [is] anecessary element [of quantum evolution and is] anothermarked distinction between quantum evolution andphyletic evolution which is mostly adaptive (andpostadaptive)rdquo Furthermore quantum evolution pro-vided a biological explanation for the long‐observeddiscontinuities in the fossil record they were not merelyartifacts of incompleteness but may in fact be indicativeof extremely rapid phenotypic transitions

According to Simpson (1944) quantum evolutionprogressed through three phases

ldquo(1) An inadaptive phase in which the groupin question loses the equilibrium of itsancestors or collaterals

(2) A preadaptive phase in which there isgreat selection pressure and the group movestoward a new equilibrium

(3) An adaptive phase in which the newequilibrium is reachedrdquo

There is a striking concordance between these phasesand (a) environmental induction of CGV release that isphenotypic accommodation (b) corresponding environ-mental selection for novel variants that is geneticaccommodation and potential assimilation and (c)stabilizing selection that is canalization of development

Yet at the time the best candidate for a mechanismby which quantum evolution might proceed was genetic

drift During the ldquohardeningrdquo of the modern synthesisas evolutionary explanations became increasingly re-stricted to selection acting on gradual allelic change(Gould 1982) the role of genetic drift seemed lesssignificant leading to Simpson changing his views onquantum evolution and ceasing to see it is a differentkind of evolution to phyletic evolution (Gould 1986) Helater wrote that quantum evolution was ldquonot a differentsort of evolution from phyletic evolution or even adistinctly different element of the total phylogeneticpattern It is a special more or less extreme and limitingcase of phyletic evolutionrdquo (Simpson 1953) In otherwords quantum evolution became simply phyleticevolution sped up and the norm mode of evolutionconsequently became phyletic gradualism (Eldredge ampGould 1972 Gould 1980)

8 | PUNCTUATED EQUILIBRIUM

It was against this backdrop that Eldredge and Gouldformulated their theory of punctuated equilibrium(Eldredge amp Gould 1972 Gould amp Eldredge 19771993) Similar to Simpsonrsquos (1944) quantum evolutionit posits that discontinuities in the fossil record need notnecessarily be indicative of incompleteness Ratherphenotypic evolution may be as intermittently rapidthat is punctuated as fossil data suggest They argue thatldquothe remarkable stasis exhibited by most species duringmillions of years is ignored (as no data)rdquo (Gould ampEldredge 1993) In other words stasis is taken as a lackof evolution and contrasted with directional phenotypicchange through time The assumption that discontinu-ities are artifacts of incompleteness implies that phyleticgradualism is synonymous with evolution

Gould and Eldredge (1993) write that we insteadought to ldquoregard stasis and discontinuity as an

FIGURE 2 Simpsonrsquos (1944) comparison of phyletic and quantum evolution Phyletic evolution is the selective tracking of a shiftingadaptive zone through time with subsequent directional phenotypic gradualism as a result Quantum evolution on the other hand is a rapidevolutionary transition between adaptive zones Note the preadaptive variation manifest in the ancestral population immediately beforequantum evolution

JACKSON | 5 of 15

expression of how evolution works when translatedinto geological timerdquo and argue that phenotypicevolution is more closely tied to speciation thanpreviously imagined Indeed studies on patterns ofchange in the fossil record have confirmed that stasis isa far more common observation than phyletic gradu-alism (Hunt 2004 2007 Hunt amp Rabosky 2014) withstasis in this sense more accurately regarded asdirectionless or nonaccumulating morphological fluc-tuations rather than true stagnance (Gould 1982Gould 2002 Voje 2016) Furthermore adaptivephenotypic change may occur so rapidly that it maypass by the record (Hunt 2010) Yet while the patterndescribed by punctuated equilibrium is thereforebroadly supported by observation the mechanism bywhich this pattern arises particularly the rapidtransitions of form remains unclear (Futuyma 2015)Around the time of its inception and early receptionthe theory of punctuated equilibrium was met withstrong critique (eg Charlesworth Lande amp Slatkin1982 Gingerich 1984 Kirkpatrick 1982) generating adebate that continues to this day (eg Lieberman ampEldredge 2014 Pennell Harmon amp Uyeda 2014)

9 | FUNDAMENTAL DEBATE

It seems that the same point of contention that ledSimpson to dilute his concept of quantum evolution alsoframes the debate on the theory of punctuated equili-brium as well as the controversy surrounding phenotypicand genetic accommodation namely the tension be-tween the understanding of development as programmedor constructive (Laland et al 2015) itself a manifestationof the differing perspectives of the gene‐centered andorganism‐centered views of evolution (Baedke 2018Nicholson 2014) Clearly the same currents of thoughtare manifest in each of these instances yet the contentionremains unresolved

It is tempting to argue that Badyaevrsquos (2009)alternating ldquoemergence of novel developmental varia-tionrdquo and ldquostabilization of locally appropriate organis-mndashenvironment associations by natural selectionrdquo arequalitatively different processes the former approxi-mated by quantum evolution punctuated equilibriaand plasticity‐led evolution and the latter by phyleticevolution stasis and gradualism Indeed this would bean approach informed by the distinction betweenorigins and distributional explanations (Erwin 2017)Yet this does not make sense in terms of our under-standing of development

From a developmental perspective gradualism andPLE are not qualitatively different processes Both are

governed by a regulatory network that facilitates varia-tion and buffers genetic and environmental fluctuationstowards functionality (Uller et al 2018) The keydifference is simply the phenotypic accommodation ofPLE which may be exaggerated by the release ofcapacitance in the form of CGV This facilitates thedistinctive rapidity described by quantum evolution(Simpson 1944) and punctuated equilibrium (Eldredgeamp Gould 1972) Thus the difference is quantitative notqualitative a matter of thresholds and amounts ratherthan kinds

10 | MACROEVOLUTIONARYPATTERNS OF PLASTICITY ‐LEDEVOLUTION

PLE as the evolutionary process described abovepredicts in my view in short two major features inthe fossil record (a) stabilizing selection duringperiods of environmental stability leading to decreas-ing phenotypic variation but increasing evolutionarycapacitance in the form of cryptic genetic variationand (b) episodes of environmentally induced pheno-typic plasticity facilitating rapid bursts of evolution asthis capacitance is expended

To be clear in this framework plasticity‐led evolu-tion (PLE) that is the accumulation and release ofcryptic genetic variation (CGV) is a putative generatorof those patterns in the fossil record previouslydescribed as quantum evolution (Simpson 1944) andpunctuated equilibrium (Eldredge amp Gould 1972)Periods of phyletic evolution or stasis are understoodas stabilizing selection acting on an organism Quantumevolution and the rapid shifts of punctuated equilibriaare hypothesized to be facilitated by phenotypic andgenetic accommodation This is not to say that evolutioncannot occur through phyletic gradualism that isstepwise accumulation of genetic change nor to saythat the fossil record is not incomplete in parts I onlymean to say that PLE provides an alternative explana-tion for the phenomena mentioned above While thetheory of punctuated equilibrium is associated withspeciation and quantum evolution is traditionallydiscussed in terms of anagenetic evolution despiteSimpson (1944) himself writing that it ldquomay be involvedin either speciation or phyletic evolutionrdquo and thatldquocertain patterns within those modes intergrade withquantum evolutionrdquo this distinction holds little sig-nificance for this model of PLE the environmentallymediated accumulation and release of CGV can occur ineither instance and whether or not speciation occurs isinconsequential

6 of 15 | JACKSON

11 | EVIDENCE OF PLASTICITY‐LEDEVOLUTION IN THE FOSSIL RECORD

When searching for evidence of PLE in the fossil recordit is insufficient to simply point to patterns of stasis ordiscontinuity As Gould and Eldredge (1993) point outldquoExamples of stasis alone [hellip] and simple abruptreplacement although conforming to expectations ofpunctuated equilibrium are not direct evidence for ourmechanism for stasis might just be a lull in anageneticgradualism [hellip] and replacement might represent rapidtransformation without branching or migration of adistant (phyletic or geographic) relative rather thanevolution in siturdquo Likewise discontinuities may indeedhave geological rather than biological explanationscaused by erosion of fossil‐bearing strata or a hiatus ina sedimentary deposition for example Furthermore theprocesses of phenotypic and genetic accommodation arechallenging to study in the fossil record As Pigliucciand Murren (2003) caution us ldquogiven the hypothesisthat [genetic] assimilation can occur within a fewgenerations it ironically may be too fast for us tocatchrdquo Indeed the rapidity of phenotypic transitionsreported by Hunt (2010) points to a similar problemThese challenges present a problem to the study of PLEin the fossil record

Beyond these challenges a further difficulty is thatthe central macroevolutionary prediction of PLE issuperficially identical to that of phyletic gradualismPLE predicts that novel environments lead to evolu-tionary change mediated by phenotypic and geneticaccommodation Phyletic gradualism predicts that novelenvironments lead to evolutionary change mediated bynatural selection on gradually accumulating mutationsThus the observation of phenotypic change accompa-nying environmental change is insufficient to supportPLE over phyletic gradualism without access to geneticmaterial and consequently the underlying processes weare left unable to determine which explanation is morelikely In general this is a difficulty faced wheninvestigating the fossil record in search for support forPLE macroevolutionary patterns have historically pro-ven an inadequate source of discrimination betweengenetic variation and plasticity (Hopkins 2014 Schoch2014 Webster 2019)

While the fact that the process of PLE is coherent withthe structure of the fossil record should not be ignored asinsignificant it remains the case that to provide definitivesupport for PLE as a mode of evolution we need to directour attention towards identifying evidence that distin-guishes it from phyletic gradualism Approaches towardsdifferentiating these processes rely on the study ofprocesses instead of patterns In other words they focus

on providing support for the underlying mechanisms ofPLE rather than the broad‐scale outcome

12 | ENVIRONMENTAL STRESSAND PHENOTYPIC VARIABILITY

Specifically the key feature distinguishing PLE fromphyletic gradualism is the environmentally mediateddevelopmental plasticity that facilitates rapid phenotypicchange While such change can be measured in thedifference between mean morphologies this does little tohelp separate hypotheses of PLE and phyletic gradualismseeing as both predict shifting phenotypic means as aresult of environmental change A stronger indicator ofreduced canalization and subsequent increased develop-mental plasticity is a measure of the morphologicalvariability induced by environmental stress (Hoffmann ampHercus 2000)

The relationship between environmental stress andmorphology is the subject of the Plus ccedila Change model(Sheldon 1997) In short it seeks to explain theobservation that morphological stasis appears the mostcommon response to fluctuations in environmentalconditions in the fossil record Drawing on a deep‐timeperspective it argues that long‐term generalists are thespecies most likely to survive such fluctuations andconsequently predicts gradualistic phenotypic changeunder stable or near‐stable environmental conditionsand phenotypic stasis during periods of strong environ-mental fluctuations

In its general sense this model seemingly suggests theopposite of PLE to manifest in the fossil record Yet itcontains the caveat that upon reaching a critical thresh-old of stress rapid phyletic evolution may be the result(Figure 3) The parallels between this prediction and theprocess of PLE outlined above are readily apparent thismodel suggests that in a strongly fluctuating environmentnatural selection favors those organisms that sufficientlycanalize their development to best track an underlyingphenotypic optimum yet when an extreme environmentis encountered developmental canalization is overcomeand previously cryptic variation is released

The prediction that environmental stressors candecrease developmental stability by overcoming cana-lization and releasing CGV implies that phenotypicvariability can be regarded as a proxy for develop-mental instability (Klingenberg 2019 Webster 2019Willmore Young amp Richtsmeier 2007) Thus thephenotypendashenvironment association of interest for thestudy of PLE in the fossil record is that betweenenvironmental conditions and phenotypic variabilityrather than phenotypic variation

JACKSON | 7 of 15

This distinction is key for separating the predictions ofPLE and phyletic gradualism Variation that is morpho-logical differences between assemblages associated withenvironmental change is insufficient to support PLE forthe reasons discussed above However an associationbetween environmental change and variability that isthe variance in form within assemblages informs usabout their developmental stability and thus permitsinvestigation into the unique prediction of PLE thatenvironmental stress may overcome canalization andinduce the release of CGV with consequent expression ofnovel variants upon which selection can act Theconsequence of this process is an increase in phenotypicvariability coupled with a shift in phenotypic meanassociated with environmental stress a pattern notpredicted by phyletic gradualism

Below I provide two examples of approaches towardsproviding evidence of PLE in the fossil record that usethis association Before discussing them however I firstpresent their necessary methodological framework

13 | FRAMEWORK OF ANALYSIS

The focal point for these investigations is the relation-ship between environmental factors and morphology offossil assemblages Morphology is essentially the onlybiological information available in the fossil record andit is therefore in general of paramount importance tocapture and study morphology in as high resolution aspossible and with careful consideration of potentialtaphonomic biases such as disarticulation of hardparts or later structural deformation (Webster 2019)This is particularly true however when studyingintraspecific patterns of variation such as thoseI will present below Many modern phylogeneticcladistic analyses rely on discretizing variation intocharacters which strongly limits their ability to studyintraspecific variation An alternative approach toanalyzing morphological variation is through the useof morphometric methods which quantify shapes intosets of measurements that can be mathematicallyanalyzed (MacLeod 2017) Such methods includelandmark analyses (Bookstein 1997) eigenshape ana-lysis (MacLeod 2002) and elliptical Fourier analysis(Kuhl amp Giardina 1982) These numerical approachesto describing variation in form translate morphologyinto continuous data and are consequently more suitedto the study of the subtle variation exhibited intraspe-cifically than the use of discrete characters and indeedare a prerequisite to examining fine‐scale patterns ofmorphological variability

While straightforward use of statistical variance isoften employed in studies on developmental instability(eg Imasheva Loeschcke Zhivotovsky amp Lazebny1997) novel methods for quantifying variance specificallystemming from developmental plasticity have seen recentdevelopment (Pertoldi amp Kristensen 2015 Pertoldi et al2014) Coupling shifts in variability to shifts in pheno-typic mean is the key to identifying developmentalplasticity in the fossil record as I will discuss belowThe putative confounding effect of time‐averaging onmorphological variance in fossil specimens has not beensupported by empirical studies Rather it seems thatpaleontological and neontological samples have compar-able levels of variability (Kidwell amp Flessa 1995 BushPowell Arnold Bert amp Daley 2002 Hunt 2004 Hunt2004 Hunt Bell amp Travis 2008 Webster 2019)

To approach the environmental aspect of PLE it isnecessary to identify indications of environmental con-ditions in fossil assemblages Patterns of intraspecificvariation in fossil species across a geographic andputative environmental range are by no means uncom-mon (eg Hopkins amp Webster 2009 Jackson amp Budd2017) Such patterns have been studied through the

FIGURE 3 Sheldonrsquos (1997) Plus ca Change model Stronglyfluctuating environments are predicted to increase selection forstasis until a threshold is reached and a punctuation sensuEldredge and Gould (1972) is generated This model can beinterpreted as predicting selection for increased canalization influctuating environments until a threshold is reached at whichpoint cryptic genetic variation is expressed and phenotypicaccommodation occurs

8 of 15 | JACKSON

assessment of associations between morphology andvarious environmental gradients (eg Cisne ChandleeRabe amp Cohen 1982 Scarponi amp Kowalewski 2004Webber amp Hunda 2007) Indeed a variety of geochemicalproxies are available for the investigation and approx-imation of depositional environments (Webster 2019)such as using trace metals as a proxy for paleoredox andpaleoproductivity (Tribovillard Algeo Lyons amp Riboul-leau 2006) for example However care must be takenwhen interpreting such proxies to minimize environ-mental variation generated by taphonomic processessuch as diagenetic alterations of the rock (Webster 2019)

The relationship between environmental stress anddevelopmental plasticity by proxy of morphologicalvariability is used in the following two approachestowards providing evidence of PLE as a mode ofevolution in the fossil record

14 | EVIDENCE OF PHENOTYPICAND GENETIC ACCOMMODATION

As mentioned above both phyletic gradualism and PLEpredict a shift in morphology following exposure tonovel environments albeit resulting from differentprocesses making it difficult to uncover traces defini-tively of PLE Gradual adaptive evolution wouldmanifest as an assemblage shifting towards a newphenotypic optimum in response to environmentalperturbation Likewise PLE would present a similarpattern yet with the environment both inducing as wellas selecting for novel variants

However PLE predicts a pattern of increasingvariability as the expression of CGV is induced byenvironmental stress which distinguishes it fromphyletic gradualism In this mode of evolution pheno-typic accommodation increases variability such that thephenotype subsequently genetically accommodated or

canalized lies outside its ancestral normal range (Figure4 cf Figure 1) While developmental instability isknown to occur during stress including on the rangemargins of populations (Bridle amp Vines 2007 RiesebergArcher amp Wayne 1999) such dynamics do not predictthe induction of novel variation outside the ancestralrange Consequently shifts in mean morphologycoupled with strongly increasing variability beyond anancestrally normal range can be used as an indicator ofPLE (Figure 4)

This means that the study of morphological varia-bility across environmental gradients offers an oppor-tunity to test specifically for evidence of the phenotypicand genetic accommodation inherent to PLE as de-scribed above

Naturally demonstrating this kind of intraspecificpattern requires a fossil organism or lineage with a highlyresolved stratigraphy sufficient abundance and knownpresence across temporal andor environmental gradients(Schoch 2014) Such research is not uncommon (egHopkins amp Webster 2009 Schoch 2014 Webber ampHunda 2007) yet morphological variability is rarely thefocal point and phenotypic accommodation rarer still

However the Cambrian trilobite‐like arthropod Ag-nostus pisiformis has been demonstrated to exhibitdifferentiated patterns of variation and variability acrossa gradient of dysoxic stress (Jackson amp Budd 2017)implying that environmental conditions may indeed beinducing as well as selecting for phenotypic variants

Further investigation in this framework is needed toproperly test the fossil record for evidence of PLEPatterns of phenotypic variation and chiefly variabilityuncovered by fine‐scale studies of fossil morphology (egDai Zhang Peng amp Yao 2017 Esteve 2012 NeubauerHarzhauser amp Kroh 2013) ought to be considered withinthe framework of PLE Likewise research exploringputative canalization of fossil organisms (eg PimientoTang Zamora Klug amp Saacutenchez‐Villagra 2018 Webster

Mor

phol

ogy

Time

Mor

phol

ogy

Time

(a) (b) EnvironmentalOptimum

Population

FIGURE 4 Illustration of two hypothetical responses of an evolutionary lineage to an environmental perturbation (a) Natural selectioncauses an adaptive shift of morphology towards a new environmental optimum in line with the predictions of phyletic gradualism (b) Theenvironmentally mediated release of cryptic genetic variation increases phenotypic variability facilitating rapid adaptive evolution towards anew optimum as per the process of plasticity‐led evolution Stabilizing selection subsequently leads to canalization around this newoptimum

JACKSON | 9 of 15

2015) could explicitly test for indications of PLE havingoccurred

15 | EVIDENCE OFDEVELOPMENTALLY BIASEDEVOLUTIONARY TRAJECTORIES

Adaptive radiations represent periods of rapid evolu-tion in which ancestral forms of a lineage diversify inenvironment morphology functionality and so forthWhen such patterns feature repeated evolutionarysequences such as in the case of the Anolis lizards ofthe Caribbean (Losos Jackman Larson de Queiroz ampRodriacuteguez‐Schettino 1998) they afford us the oppor-tunity to examine the fossil record for evidence ofanother component of PLE namely whether or notrepeated patterns of phenotypic evolution are some-times better explained by developmental bias thanconvergent selection At the heart of this issue lies therole of the environment in adaptive radiationsTraditionally the environment is seen as facilitatingsuch episodes by providing ecological opportunity todiversifying clades (Stroud amp Losos 2016) However asoutlined above the environment may also act as aninducer of novel variants by interacting with develop-mental processes Erwin (2017) describes these per-spectives as emphasizing ldquoenvironmental pullrdquo andldquodevelopmental pushrdquo respectively

West‐Eberhardrsquos (2003) flexible‐stem model sug-gests that plasticity mediated by the regulatorystructure of an ancestral populationrsquos developmentwill bias repeated independent diversificationsalong similar evolutionary trajectories via the pro-cesses of phenotypic and genetic accommodationThis represents an alternative or complementaryexplanation as to why independently derived lineages

of an ancestral population repeatedly evolve similarphenotypes such as the ldquoecomorphsrdquo of the Anolislizards mentioned above (Losos et al 1998 Williams1972) While this model has been explored in aneontological context (Gibert 2017 Parsons et al2016 Schneider amp Meyer 2017 Wund Baker ClancyGolub amp Foster 2008) it has not yet seen attention ina paleontological setting

Although the identification of adaptive radiations inthe fossil record is not a straightforward endeavor(Lieberman 2012) essentially any sequence of repeateddivergence from an ancestral stock population is suitablefor investigating the flexible‐stem informed model ofparallelism Similarly to the discussion on phenotypicand genetic accommodation above the focal point ofresearch in such systems should be on the variability ofmorphology rather than variation relative environmen-tal gradients If it is indeed phenotypic accommodationthat is the environmentally mediated release of CGVthat directs these repeated evolutionary trajectories weshould predict an increasing degree of phenotypicvariability to coincide with these divergences (Figure 5)This coupling of increasing variability and shiftingvariation is not conventionally anticipated in a patternof repeated parallel divergences that is convergentselection as explained above

The positive identification of such patterns in thefossil record would thus provide distinctive support forthe underlying processes of PLE both in terms of thesuggested environmentally mediated release of CGV aswell as the hypothesized bias exercised by the regulatorynature of development Research towards uncoveringsuch support should be directed towards reassessing thephenotypendashenvironment associations of studied fossiladaptive radiations (eg Abe amp Lieberman 2012 NeigeDera amp Dommergues 2013 Neubauer et al 2013) in thecontext of PLE

Mor

phol

ogy

Time

EnvironmentalOptimum

Ancestral Population

Diverging Population

FIGURE 5 Repeated divergences from an ancestral or stock population The bias native to the ancestral developmental network directsmorphological change along similar evolutionary trajectories as the successive independently diverging populations encounter the newenvironment An increase in morphological variability coupled with such parallelism of phenotypic mean change is predicted by the releaseof cryptic genetic variation in plasticity‐led evolution but not by the conventional process of convergent evolution resulting from similar andrepeated selective pressures

10 of 15 | JACKSON

16 | CONCLUDING REMARKS

Viewing development as constructive is a promisingconceptual framework for enriching our understandingof the processes and patterns of evolution It provides thescaffolding for properly addressing issues of centralimportance such as the role of developmental biases indirecting evolutionary change However the majority ofefforts towards this end are conducted on extantorganisms with minimal reference to paleontologicaldata This is unfortunate as the fossil record is afantastically rich source of information that is yet to befully explored in this intriguing and stimulating contextWhile microevolutionary concepts such as developmentalbias are challenging to approach in the fossil record Iargue that it is manifest in paleontological materials inthe form of macroevolutionary patterns concordant withthe predictions of plasticity‐led evolution Indeed suchpredictions recapitulate earlier theories of quantumevolution and punctuated equilibrium

These predictions are specifically an evolutionarymodel consisting of two major parts (a) stabilizingselection during periods of environmental stabilityleading to a reduction in phenotypic variation but anincrease in canalization CGV and hence evolutionarycapacitance and (b) episodes of environmentally inducedplasticity facilitating rapid evolution via the expression ofpreviously cryptic variation

Research aiming to uncover support for PLE in thefossil record should be directed towards the environ-mentndashphenotype association of fossil assemblages Spe-cifically towards patterns of correlation between envir-onmental stress and phenotypic variability as a proxy fordevelopmental instability These may be manifest alongsingle lineages or throughout adaptive radiations Thefossil record offers enormous potential to provide datathat when viewed through the lens of PLE providesbiologically meaningful information about evolution interms of both patterns and processes

ACKNOWLEDGMENTS

I would like to thank Armin Moczek for inviting me tocontribute to this special issue on developmental bias inevolution My thanks to two anonymous reviewers whoprovided thoughtful constructive and immensely helpfulcomments that greatly improved this paper I am gratefulto Graham Budd for our many discussions on this topicover the years I am thankful to the Uller group at LundUniversity for formative discussions particularly TobiasUller who also provided valuable comments on an earlydraft of this paper My thoughts presented here weregreatly informed by the SFI workshop ldquoDevelopmental

Bias and Evolutionrdquo and I direct my gratitude towards itsorganizers and participants This study was funded by theKnut and Alice Wallenberg Foundation grant to TobiasUller (KAW 20120155)

CONFLICT OF INTEREST

The author declares that there is no conflict of interest

ORCID

Illiam S C Jackson httporcidorg0000-0002-7948-2860

REFERENCES

Abe F R amp Lieberman B S (2012) Quantifying morphologicalchange during an evolutionary radiation of Devonian trilobitesPaleobiology 38 292ndash307 httpsdoiorg101666100471

Arthur W (2004) The effect of development on the directionof evolution Toward a twenty‐first century consensusEvolution amp Development 6(4) 282ndash288 httpsdoiorg101111j1525‐142X200404033x

Badyaev A V (2005) Stress‐induced variation in evolution Frombehavioural plasticity to genetic assimilation Proceedings of theRoyal Society B Biological Sciences 272 877ndash886 httpsdoiorg101098rspb20043045

Badyaev A V (2009) Evolutionary significance of phenotypicaccommodation in novel environments An empirical test of theBaldwin effect Philosophical Transactions of the Royal SocietyB Biological Sciences 364 1125ndash1141 httpsdoiorg101098rstb20080285

Baedke J (2018) Organism where art thou Old and new challengesfor organism‐centered biology Journal of the History of Biology52 293ndash324 httpsdoiorg101007s10739‐018‐9549‐4

Bookstein F L (1997) Morphometric tools for landmark dataGeometry and biology Cambridge Cambridge University Press

Brakefield P M (2011) Evo‐devo and accounting for Darwinsendless forms Philosophical Transactions of the Royal Society BBiological Sciences 366(1574) 2069ndash2075 httpsdoiorg101098rstb20110007

Bridle J R amp Vines T H (2007) Limits to evolution at rangemargins When and why does adaptation fail Trends in Ecologyamp Evolution 22(3) 140ndash147 httpsdoiorg101016jtree200611002

Bush A M Powell M G Arnold W S Bert T M amp Daley G M(2002) Time‐averaging evolution and morphologic variationPaleobiology 28 9ndash25 httpsdoiorg1016660094‐8373(2002)0283C0009TAEAMV3E20CO2

Carroll S B (2005) Evolution at two levels On genes and form PLOSBiology 3 e245 httpsdoiorg101371journalpbio0030245

Carroll S B (2008) Evo‐devo and an expanding evolutionarysynthesis A genetic theory of morphological evolution Cell134 25ndash36 httpsdoiorg101016jcell200806030

Charlesworth B Lande R amp Slatkin M (1982) A neo‐Darwiniancommentary on macroevolution Evolution 36 474ndash498 httpsdoiorg1023072408095

JACKSON | 11 of 15

expression of how evolution works when translatedinto geological timerdquo and argue that phenotypicevolution is more closely tied to speciation thanpreviously imagined Indeed studies on patterns ofchange in the fossil record have confirmed that stasis isa far more common observation than phyletic gradu-alism (Hunt 2004 2007 Hunt amp Rabosky 2014) withstasis in this sense more accurately regarded asdirectionless or nonaccumulating morphological fluc-tuations rather than true stagnance (Gould 1982Gould 2002 Voje 2016) Furthermore adaptivephenotypic change may occur so rapidly that it maypass by the record (Hunt 2010) Yet while the patterndescribed by punctuated equilibrium is thereforebroadly supported by observation the mechanism bywhich this pattern arises particularly the rapidtransitions of form remains unclear (Futuyma 2015)Around the time of its inception and early receptionthe theory of punctuated equilibrium was met withstrong critique (eg Charlesworth Lande amp Slatkin1982 Gingerich 1984 Kirkpatrick 1982) generating adebate that continues to this day (eg Lieberman ampEldredge 2014 Pennell Harmon amp Uyeda 2014)

9 | FUNDAMENTAL DEBATE

It seems that the same point of contention that ledSimpson to dilute his concept of quantum evolution alsoframes the debate on the theory of punctuated equili-brium as well as the controversy surrounding phenotypicand genetic accommodation namely the tension be-tween the understanding of development as programmedor constructive (Laland et al 2015) itself a manifestationof the differing perspectives of the gene‐centered andorganism‐centered views of evolution (Baedke 2018Nicholson 2014) Clearly the same currents of thoughtare manifest in each of these instances yet the contentionremains unresolved

It is tempting to argue that Badyaevrsquos (2009)alternating ldquoemergence of novel developmental varia-tionrdquo and ldquostabilization of locally appropriate organis-mndashenvironment associations by natural selectionrdquo arequalitatively different processes the former approxi-mated by quantum evolution punctuated equilibriaand plasticity‐led evolution and the latter by phyleticevolution stasis and gradualism Indeed this would bean approach informed by the distinction betweenorigins and distributional explanations (Erwin 2017)Yet this does not make sense in terms of our under-standing of development

From a developmental perspective gradualism andPLE are not qualitatively different processes Both are

governed by a regulatory network that facilitates varia-tion and buffers genetic and environmental fluctuationstowards functionality (Uller et al 2018) The keydifference is simply the phenotypic accommodation ofPLE which may be exaggerated by the release ofcapacitance in the form of CGV This facilitates thedistinctive rapidity described by quantum evolution(Simpson 1944) and punctuated equilibrium (Eldredgeamp Gould 1972) Thus the difference is quantitative notqualitative a matter of thresholds and amounts ratherthan kinds

10 | MACROEVOLUTIONARYPATTERNS OF PLASTICITY ‐LEDEVOLUTION

PLE as the evolutionary process described abovepredicts in my view in short two major features inthe fossil record (a) stabilizing selection duringperiods of environmental stability leading to decreas-ing phenotypic variation but increasing evolutionarycapacitance in the form of cryptic genetic variationand (b) episodes of environmentally induced pheno-typic plasticity facilitating rapid bursts of evolution asthis capacitance is expended

To be clear in this framework plasticity‐led evolu-tion (PLE) that is the accumulation and release ofcryptic genetic variation (CGV) is a putative generatorof those patterns in the fossil record previouslydescribed as quantum evolution (Simpson 1944) andpunctuated equilibrium (Eldredge amp Gould 1972)Periods of phyletic evolution or stasis are understoodas stabilizing selection acting on an organism Quantumevolution and the rapid shifts of punctuated equilibriaare hypothesized to be facilitated by phenotypic andgenetic accommodation This is not to say that evolutioncannot occur through phyletic gradualism that isstepwise accumulation of genetic change nor to saythat the fossil record is not incomplete in parts I onlymean to say that PLE provides an alternative explana-tion for the phenomena mentioned above While thetheory of punctuated equilibrium is associated withspeciation and quantum evolution is traditionallydiscussed in terms of anagenetic evolution despiteSimpson (1944) himself writing that it ldquomay be involvedin either speciation or phyletic evolutionrdquo and thatldquocertain patterns within those modes intergrade withquantum evolutionrdquo this distinction holds little sig-nificance for this model of PLE the environmentallymediated accumulation and release of CGV can occur ineither instance and whether or not speciation occurs isinconsequential

6 of 15 | JACKSON

11 | EVIDENCE OF PLASTICITY‐LEDEVOLUTION IN THE FOSSIL RECORD

When searching for evidence of PLE in the fossil recordit is insufficient to simply point to patterns of stasis ordiscontinuity As Gould and Eldredge (1993) point outldquoExamples of stasis alone [hellip] and simple abruptreplacement although conforming to expectations ofpunctuated equilibrium are not direct evidence for ourmechanism for stasis might just be a lull in anageneticgradualism [hellip] and replacement might represent rapidtransformation without branching or migration of adistant (phyletic or geographic) relative rather thanevolution in siturdquo Likewise discontinuities may indeedhave geological rather than biological explanationscaused by erosion of fossil‐bearing strata or a hiatus ina sedimentary deposition for example Furthermore theprocesses of phenotypic and genetic accommodation arechallenging to study in the fossil record As Pigliucciand Murren (2003) caution us ldquogiven the hypothesisthat [genetic] assimilation can occur within a fewgenerations it ironically may be too fast for us tocatchrdquo Indeed the rapidity of phenotypic transitionsreported by Hunt (2010) points to a similar problemThese challenges present a problem to the study of PLEin the fossil record

Beyond these challenges a further difficulty is thatthe central macroevolutionary prediction of PLE issuperficially identical to that of phyletic gradualismPLE predicts that novel environments lead to evolu-tionary change mediated by phenotypic and geneticaccommodation Phyletic gradualism predicts that novelenvironments lead to evolutionary change mediated bynatural selection on gradually accumulating mutationsThus the observation of phenotypic change accompa-nying environmental change is insufficient to supportPLE over phyletic gradualism without access to geneticmaterial and consequently the underlying processes weare left unable to determine which explanation is morelikely In general this is a difficulty faced wheninvestigating the fossil record in search for support forPLE macroevolutionary patterns have historically pro-ven an inadequate source of discrimination betweengenetic variation and plasticity (Hopkins 2014 Schoch2014 Webster 2019)

While the fact that the process of PLE is coherent withthe structure of the fossil record should not be ignored asinsignificant it remains the case that to provide definitivesupport for PLE as a mode of evolution we need to directour attention towards identifying evidence that distin-guishes it from phyletic gradualism Approaches towardsdifferentiating these processes rely on the study ofprocesses instead of patterns In other words they focus

on providing support for the underlying mechanisms ofPLE rather than the broad‐scale outcome

12 | ENVIRONMENTAL STRESSAND PHENOTYPIC VARIABILITY

Specifically the key feature distinguishing PLE fromphyletic gradualism is the environmentally mediateddevelopmental plasticity that facilitates rapid phenotypicchange While such change can be measured in thedifference between mean morphologies this does little tohelp separate hypotheses of PLE and phyletic gradualismseeing as both predict shifting phenotypic means as aresult of environmental change A stronger indicator ofreduced canalization and subsequent increased develop-mental plasticity is a measure of the morphologicalvariability induced by environmental stress (Hoffmann ampHercus 2000)

The relationship between environmental stress andmorphology is the subject of the Plus ccedila Change model(Sheldon 1997) In short it seeks to explain theobservation that morphological stasis appears the mostcommon response to fluctuations in environmentalconditions in the fossil record Drawing on a deep‐timeperspective it argues that long‐term generalists are thespecies most likely to survive such fluctuations andconsequently predicts gradualistic phenotypic changeunder stable or near‐stable environmental conditionsand phenotypic stasis during periods of strong environ-mental fluctuations

In its general sense this model seemingly suggests theopposite of PLE to manifest in the fossil record Yet itcontains the caveat that upon reaching a critical thresh-old of stress rapid phyletic evolution may be the result(Figure 3) The parallels between this prediction and theprocess of PLE outlined above are readily apparent thismodel suggests that in a strongly fluctuating environmentnatural selection favors those organisms that sufficientlycanalize their development to best track an underlyingphenotypic optimum yet when an extreme environmentis encountered developmental canalization is overcomeand previously cryptic variation is released

The prediction that environmental stressors candecrease developmental stability by overcoming cana-lization and releasing CGV implies that phenotypicvariability can be regarded as a proxy for develop-mental instability (Klingenberg 2019 Webster 2019Willmore Young amp Richtsmeier 2007) Thus thephenotypendashenvironment association of interest for thestudy of PLE in the fossil record is that betweenenvironmental conditions and phenotypic variabilityrather than phenotypic variation

JACKSON | 7 of 15

This distinction is key for separating the predictions ofPLE and phyletic gradualism Variation that is morpho-logical differences between assemblages associated withenvironmental change is insufficient to support PLE forthe reasons discussed above However an associationbetween environmental change and variability that isthe variance in form within assemblages informs usabout their developmental stability and thus permitsinvestigation into the unique prediction of PLE thatenvironmental stress may overcome canalization andinduce the release of CGV with consequent expression ofnovel variants upon which selection can act Theconsequence of this process is an increase in phenotypicvariability coupled with a shift in phenotypic meanassociated with environmental stress a pattern notpredicted by phyletic gradualism

Below I provide two examples of approaches towardsproviding evidence of PLE in the fossil record that usethis association Before discussing them however I firstpresent their necessary methodological framework

13 | FRAMEWORK OF ANALYSIS

The focal point for these investigations is the relation-ship between environmental factors and morphology offossil assemblages Morphology is essentially the onlybiological information available in the fossil record andit is therefore in general of paramount importance tocapture and study morphology in as high resolution aspossible and with careful consideration of potentialtaphonomic biases such as disarticulation of hardparts or later structural deformation (Webster 2019)This is particularly true however when studyingintraspecific patterns of variation such as thoseI will present below Many modern phylogeneticcladistic analyses rely on discretizing variation intocharacters which strongly limits their ability to studyintraspecific variation An alternative approach toanalyzing morphological variation is through the useof morphometric methods which quantify shapes intosets of measurements that can be mathematicallyanalyzed (MacLeod 2017) Such methods includelandmark analyses (Bookstein 1997) eigenshape ana-lysis (MacLeod 2002) and elliptical Fourier analysis(Kuhl amp Giardina 1982) These numerical approachesto describing variation in form translate morphologyinto continuous data and are consequently more suitedto the study of the subtle variation exhibited intraspe-cifically than the use of discrete characters and indeedare a prerequisite to examining fine‐scale patterns ofmorphological variability

While straightforward use of statistical variance isoften employed in studies on developmental instability(eg Imasheva Loeschcke Zhivotovsky amp Lazebny1997) novel methods for quantifying variance specificallystemming from developmental plasticity have seen recentdevelopment (Pertoldi amp Kristensen 2015 Pertoldi et al2014) Coupling shifts in variability to shifts in pheno-typic mean is the key to identifying developmentalplasticity in the fossil record as I will discuss belowThe putative confounding effect of time‐averaging onmorphological variance in fossil specimens has not beensupported by empirical studies Rather it seems thatpaleontological and neontological samples have compar-able levels of variability (Kidwell amp Flessa 1995 BushPowell Arnold Bert amp Daley 2002 Hunt 2004 Hunt2004 Hunt Bell amp Travis 2008 Webster 2019)

To approach the environmental aspect of PLE it isnecessary to identify indications of environmental con-ditions in fossil assemblages Patterns of intraspecificvariation in fossil species across a geographic andputative environmental range are by no means uncom-mon (eg Hopkins amp Webster 2009 Jackson amp Budd2017) Such patterns have been studied through the

FIGURE 3 Sheldonrsquos (1997) Plus ca Change model Stronglyfluctuating environments are predicted to increase selection forstasis until a threshold is reached and a punctuation sensuEldredge and Gould (1972) is generated This model can beinterpreted as predicting selection for increased canalization influctuating environments until a threshold is reached at whichpoint cryptic genetic variation is expressed and phenotypicaccommodation occurs

8 of 15 | JACKSON

assessment of associations between morphology andvarious environmental gradients (eg Cisne ChandleeRabe amp Cohen 1982 Scarponi amp Kowalewski 2004Webber amp Hunda 2007) Indeed a variety of geochemicalproxies are available for the investigation and approx-imation of depositional environments (Webster 2019)such as using trace metals as a proxy for paleoredox andpaleoproductivity (Tribovillard Algeo Lyons amp Riboul-leau 2006) for example However care must be takenwhen interpreting such proxies to minimize environ-mental variation generated by taphonomic processessuch as diagenetic alterations of the rock (Webster 2019)

The relationship between environmental stress anddevelopmental plasticity by proxy of morphologicalvariability is used in the following two approachestowards providing evidence of PLE as a mode ofevolution in the fossil record

14 | EVIDENCE OF PHENOTYPICAND GENETIC ACCOMMODATION

As mentioned above both phyletic gradualism and PLEpredict a shift in morphology following exposure tonovel environments albeit resulting from differentprocesses making it difficult to uncover traces defini-tively of PLE Gradual adaptive evolution wouldmanifest as an assemblage shifting towards a newphenotypic optimum in response to environmentalperturbation Likewise PLE would present a similarpattern yet with the environment both inducing as wellas selecting for novel variants

However PLE predicts a pattern of increasingvariability as the expression of CGV is induced byenvironmental stress which distinguishes it fromphyletic gradualism In this mode of evolution pheno-typic accommodation increases variability such that thephenotype subsequently genetically accommodated or

canalized lies outside its ancestral normal range (Figure4 cf Figure 1) While developmental instability isknown to occur during stress including on the rangemargins of populations (Bridle amp Vines 2007 RiesebergArcher amp Wayne 1999) such dynamics do not predictthe induction of novel variation outside the ancestralrange Consequently shifts in mean morphologycoupled with strongly increasing variability beyond anancestrally normal range can be used as an indicator ofPLE (Figure 4)

This means that the study of morphological varia-bility across environmental gradients offers an oppor-tunity to test specifically for evidence of the phenotypicand genetic accommodation inherent to PLE as de-scribed above

Naturally demonstrating this kind of intraspecificpattern requires a fossil organism or lineage with a highlyresolved stratigraphy sufficient abundance and knownpresence across temporal andor environmental gradients(Schoch 2014) Such research is not uncommon (egHopkins amp Webster 2009 Schoch 2014 Webber ampHunda 2007) yet morphological variability is rarely thefocal point and phenotypic accommodation rarer still

However the Cambrian trilobite‐like arthropod Ag-nostus pisiformis has been demonstrated to exhibitdifferentiated patterns of variation and variability acrossa gradient of dysoxic stress (Jackson amp Budd 2017)implying that environmental conditions may indeed beinducing as well as selecting for phenotypic variants

Further investigation in this framework is needed toproperly test the fossil record for evidence of PLEPatterns of phenotypic variation and chiefly variabilityuncovered by fine‐scale studies of fossil morphology (egDai Zhang Peng amp Yao 2017 Esteve 2012 NeubauerHarzhauser amp Kroh 2013) ought to be considered withinthe framework of PLE Likewise research exploringputative canalization of fossil organisms (eg PimientoTang Zamora Klug amp Saacutenchez‐Villagra 2018 Webster

Mor

phol

ogy

Time

Mor

phol

ogy

Time

(a) (b) EnvironmentalOptimum

Population

FIGURE 4 Illustration of two hypothetical responses of an evolutionary lineage to an environmental perturbation (a) Natural selectioncauses an adaptive shift of morphology towards a new environmental optimum in line with the predictions of phyletic gradualism (b) Theenvironmentally mediated release of cryptic genetic variation increases phenotypic variability facilitating rapid adaptive evolution towards anew optimum as per the process of plasticity‐led evolution Stabilizing selection subsequently leads to canalization around this newoptimum

JACKSON | 9 of 15

2015) could explicitly test for indications of PLE havingoccurred

15 | EVIDENCE OFDEVELOPMENTALLY BIASEDEVOLUTIONARY TRAJECTORIES

Adaptive radiations represent periods of rapid evolu-tion in which ancestral forms of a lineage diversify inenvironment morphology functionality and so forthWhen such patterns feature repeated evolutionarysequences such as in the case of the Anolis lizards ofthe Caribbean (Losos Jackman Larson de Queiroz ampRodriacuteguez‐Schettino 1998) they afford us the oppor-tunity to examine the fossil record for evidence ofanother component of PLE namely whether or notrepeated patterns of phenotypic evolution are some-times better explained by developmental bias thanconvergent selection At the heart of this issue lies therole of the environment in adaptive radiationsTraditionally the environment is seen as facilitatingsuch episodes by providing ecological opportunity todiversifying clades (Stroud amp Losos 2016) However asoutlined above the environment may also act as aninducer of novel variants by interacting with develop-mental processes Erwin (2017) describes these per-spectives as emphasizing ldquoenvironmental pullrdquo andldquodevelopmental pushrdquo respectively

West‐Eberhardrsquos (2003) flexible‐stem model sug-gests that plasticity mediated by the regulatorystructure of an ancestral populationrsquos developmentwill bias repeated independent diversificationsalong similar evolutionary trajectories via the pro-cesses of phenotypic and genetic accommodationThis represents an alternative or complementaryexplanation as to why independently derived lineages

of an ancestral population repeatedly evolve similarphenotypes such as the ldquoecomorphsrdquo of the Anolislizards mentioned above (Losos et al 1998 Williams1972) While this model has been explored in aneontological context (Gibert 2017 Parsons et al2016 Schneider amp Meyer 2017 Wund Baker ClancyGolub amp Foster 2008) it has not yet seen attention ina paleontological setting

Although the identification of adaptive radiations inthe fossil record is not a straightforward endeavor(Lieberman 2012) essentially any sequence of repeateddivergence from an ancestral stock population is suitablefor investigating the flexible‐stem informed model ofparallelism Similarly to the discussion on phenotypicand genetic accommodation above the focal point ofresearch in such systems should be on the variability ofmorphology rather than variation relative environmen-tal gradients If it is indeed phenotypic accommodationthat is the environmentally mediated release of CGVthat directs these repeated evolutionary trajectories weshould predict an increasing degree of phenotypicvariability to coincide with these divergences (Figure 5)This coupling of increasing variability and shiftingvariation is not conventionally anticipated in a patternof repeated parallel divergences that is convergentselection as explained above

The positive identification of such patterns in thefossil record would thus provide distinctive support forthe underlying processes of PLE both in terms of thesuggested environmentally mediated release of CGV aswell as the hypothesized bias exercised by the regulatorynature of development Research towards uncoveringsuch support should be directed towards reassessing thephenotypendashenvironment associations of studied fossiladaptive radiations (eg Abe amp Lieberman 2012 NeigeDera amp Dommergues 2013 Neubauer et al 2013) in thecontext of PLE

Mor

phol

ogy

Time

EnvironmentalOptimum

Ancestral Population

Diverging Population

FIGURE 5 Repeated divergences from an ancestral or stock population The bias native to the ancestral developmental network directsmorphological change along similar evolutionary trajectories as the successive independently diverging populations encounter the newenvironment An increase in morphological variability coupled with such parallelism of phenotypic mean change is predicted by the releaseof cryptic genetic variation in plasticity‐led evolution but not by the conventional process of convergent evolution resulting from similar andrepeated selective pressures

10 of 15 | JACKSON

16 | CONCLUDING REMARKS

Viewing development as constructive is a promisingconceptual framework for enriching our understandingof the processes and patterns of evolution It provides thescaffolding for properly addressing issues of centralimportance such as the role of developmental biases indirecting evolutionary change However the majority ofefforts towards this end are conducted on extantorganisms with minimal reference to paleontologicaldata This is unfortunate as the fossil record is afantastically rich source of information that is yet to befully explored in this intriguing and stimulating contextWhile microevolutionary concepts such as developmentalbias are challenging to approach in the fossil record Iargue that it is manifest in paleontological materials inthe form of macroevolutionary patterns concordant withthe predictions of plasticity‐led evolution Indeed suchpredictions recapitulate earlier theories of quantumevolution and punctuated equilibrium

These predictions are specifically an evolutionarymodel consisting of two major parts (a) stabilizingselection during periods of environmental stabilityleading to a reduction in phenotypic variation but anincrease in canalization CGV and hence evolutionarycapacitance and (b) episodes of environmentally inducedplasticity facilitating rapid evolution via the expression ofpreviously cryptic variation

Research aiming to uncover support for PLE in thefossil record should be directed towards the environ-mentndashphenotype association of fossil assemblages Spe-cifically towards patterns of correlation between envir-onmental stress and phenotypic variability as a proxy fordevelopmental instability These may be manifest alongsingle lineages or throughout adaptive radiations Thefossil record offers enormous potential to provide datathat when viewed through the lens of PLE providesbiologically meaningful information about evolution interms of both patterns and processes

ACKNOWLEDGMENTS

I would like to thank Armin Moczek for inviting me tocontribute to this special issue on developmental bias inevolution My thanks to two anonymous reviewers whoprovided thoughtful constructive and immensely helpfulcomments that greatly improved this paper I am gratefulto Graham Budd for our many discussions on this topicover the years I am thankful to the Uller group at LundUniversity for formative discussions particularly TobiasUller who also provided valuable comments on an earlydraft of this paper My thoughts presented here weregreatly informed by the SFI workshop ldquoDevelopmental

Bias and Evolutionrdquo and I direct my gratitude towards itsorganizers and participants This study was funded by theKnut and Alice Wallenberg Foundation grant to TobiasUller (KAW 20120155)

CONFLICT OF INTEREST

The author declares that there is no conflict of interest

ORCID

Illiam S C Jackson httporcidorg0000-0002-7948-2860

REFERENCES

Abe F R amp Lieberman B S (2012) Quantifying morphologicalchange during an evolutionary radiation of Devonian trilobitesPaleobiology 38 292ndash307 httpsdoiorg101666100471

Arthur W (2004) The effect of development on the directionof evolution Toward a twenty‐first century consensusEvolution amp Development 6(4) 282ndash288 httpsdoiorg101111j1525‐142X200404033x

Badyaev A V (2005) Stress‐induced variation in evolution Frombehavioural plasticity to genetic assimilation Proceedings of theRoyal Society B Biological Sciences 272 877ndash886 httpsdoiorg101098rspb20043045

Badyaev A V (2009) Evolutionary significance of phenotypicaccommodation in novel environments An empirical test of theBaldwin effect Philosophical Transactions of the Royal SocietyB Biological Sciences 364 1125ndash1141 httpsdoiorg101098rstb20080285

Baedke J (2018) Organism where art thou Old and new challengesfor organism‐centered biology Journal of the History of Biology52 293ndash324 httpsdoiorg101007s10739‐018‐9549‐4

Bookstein F L (1997) Morphometric tools for landmark dataGeometry and biology Cambridge Cambridge University Press

Brakefield P M (2011) Evo‐devo and accounting for Darwinsendless forms Philosophical Transactions of the Royal Society BBiological Sciences 366(1574) 2069ndash2075 httpsdoiorg101098rstb20110007

Bridle J R amp Vines T H (2007) Limits to evolution at rangemargins When and why does adaptation fail Trends in Ecologyamp Evolution 22(3) 140ndash147 httpsdoiorg101016jtree200611002

Bush A M Powell M G Arnold W S Bert T M amp Daley G M(2002) Time‐averaging evolution and morphologic variationPaleobiology 28 9ndash25 httpsdoiorg1016660094‐8373(2002)0283C0009TAEAMV3E20CO2

Carroll S B (2005) Evolution at two levels On genes and form PLOSBiology 3 e245 httpsdoiorg101371journalpbio0030245

Carroll S B (2008) Evo‐devo and an expanding evolutionarysynthesis A genetic theory of morphological evolution Cell134 25ndash36 httpsdoiorg101016jcell200806030

Charlesworth B Lande R amp Slatkin M (1982) A neo‐Darwiniancommentary on macroevolution Evolution 36 474ndash498 httpsdoiorg1023072408095

JACKSON | 11 of 15

11 | EVIDENCE OF PLASTICITY‐LEDEVOLUTION IN THE FOSSIL RECORD

When searching for evidence of PLE in the fossil recordit is insufficient to simply point to patterns of stasis ordiscontinuity As Gould and Eldredge (1993) point outldquoExamples of stasis alone [hellip] and simple abruptreplacement although conforming to expectations ofpunctuated equilibrium are not direct evidence for ourmechanism for stasis might just be a lull in anageneticgradualism [hellip] and replacement might represent rapidtransformation without branching or migration of adistant (phyletic or geographic) relative rather thanevolution in siturdquo Likewise discontinuities may indeedhave geological rather than biological explanationscaused by erosion of fossil‐bearing strata or a hiatus ina sedimentary deposition for example Furthermore theprocesses of phenotypic and genetic accommodation arechallenging to study in the fossil record As Pigliucciand Murren (2003) caution us ldquogiven the hypothesisthat [genetic] assimilation can occur within a fewgenerations it ironically may be too fast for us tocatchrdquo Indeed the rapidity of phenotypic transitionsreported by Hunt (2010) points to a similar problemThese challenges present a problem to the study of PLEin the fossil record

Beyond these challenges a further difficulty is thatthe central macroevolutionary prediction of PLE issuperficially identical to that of phyletic gradualismPLE predicts that novel environments lead to evolu-tionary change mediated by phenotypic and geneticaccommodation Phyletic gradualism predicts that novelenvironments lead to evolutionary change mediated bynatural selection on gradually accumulating mutationsThus the observation of phenotypic change accompa-nying environmental change is insufficient to supportPLE over phyletic gradualism without access to geneticmaterial and consequently the underlying processes weare left unable to determine which explanation is morelikely In general this is a difficulty faced wheninvestigating the fossil record in search for support forPLE macroevolutionary patterns have historically pro-ven an inadequate source of discrimination betweengenetic variation and plasticity (Hopkins 2014 Schoch2014 Webster 2019)

While the fact that the process of PLE is coherent withthe structure of the fossil record should not be ignored asinsignificant it remains the case that to provide definitivesupport for PLE as a mode of evolution we need to directour attention towards identifying evidence that distin-guishes it from phyletic gradualism Approaches towardsdifferentiating these processes rely on the study ofprocesses instead of patterns In other words they focus

on providing support for the underlying mechanisms ofPLE rather than the broad‐scale outcome

12 | ENVIRONMENTAL STRESSAND PHENOTYPIC VARIABILITY

Specifically the key feature distinguishing PLE fromphyletic gradualism is the environmentally mediateddevelopmental plasticity that facilitates rapid phenotypicchange While such change can be measured in thedifference between mean morphologies this does little tohelp separate hypotheses of PLE and phyletic gradualismseeing as both predict shifting phenotypic means as aresult of environmental change A stronger indicator ofreduced canalization and subsequent increased develop-mental plasticity is a measure of the morphologicalvariability induced by environmental stress (Hoffmann ampHercus 2000)

The relationship between environmental stress andmorphology is the subject of the Plus ccedila Change model(Sheldon 1997) In short it seeks to explain theobservation that morphological stasis appears the mostcommon response to fluctuations in environmentalconditions in the fossil record Drawing on a deep‐timeperspective it argues that long‐term generalists are thespecies most likely to survive such fluctuations andconsequently predicts gradualistic phenotypic changeunder stable or near‐stable environmental conditionsand phenotypic stasis during periods of strong environ-mental fluctuations

In its general sense this model seemingly suggests theopposite of PLE to manifest in the fossil record Yet itcontains the caveat that upon reaching a critical thresh-old of stress rapid phyletic evolution may be the result(Figure 3) The parallels between this prediction and theprocess of PLE outlined above are readily apparent thismodel suggests that in a strongly fluctuating environmentnatural selection favors those organisms that sufficientlycanalize their development to best track an underlyingphenotypic optimum yet when an extreme environmentis encountered developmental canalization is overcomeand previously cryptic variation is released

The prediction that environmental stressors candecrease developmental stability by overcoming cana-lization and releasing CGV implies that phenotypicvariability can be regarded as a proxy for develop-mental instability (Klingenberg 2019 Webster 2019Willmore Young amp Richtsmeier 2007) Thus thephenotypendashenvironment association of interest for thestudy of PLE in the fossil record is that betweenenvironmental conditions and phenotypic variabilityrather than phenotypic variation

JACKSON | 7 of 15

This distinction is key for separating the predictions ofPLE and phyletic gradualism Variation that is morpho-logical differences between assemblages associated withenvironmental change is insufficient to support PLE forthe reasons discussed above However an associationbetween environmental change and variability that isthe variance in form within assemblages informs usabout their developmental stability and thus permitsinvestigation into the unique prediction of PLE thatenvironmental stress may overcome canalization andinduce the release of CGV with consequent expression ofnovel variants upon which selection can act Theconsequence of this process is an increase in phenotypicvariability coupled with a shift in phenotypic meanassociated with environmental stress a pattern notpredicted by phyletic gradualism

Below I provide two examples of approaches towardsproviding evidence of PLE in the fossil record that usethis association Before discussing them however I firstpresent their necessary methodological framework

13 | FRAMEWORK OF ANALYSIS

The focal point for these investigations is the relation-ship between environmental factors and morphology offossil assemblages Morphology is essentially the onlybiological information available in the fossil record andit is therefore in general of paramount importance tocapture and study morphology in as high resolution aspossible and with careful consideration of potentialtaphonomic biases such as disarticulation of hardparts or later structural deformation (Webster 2019)This is particularly true however when studyingintraspecific patterns of variation such as thoseI will present below Many modern phylogeneticcladistic analyses rely on discretizing variation intocharacters which strongly limits their ability to studyintraspecific variation An alternative approach toanalyzing morphological variation is through the useof morphometric methods which quantify shapes intosets of measurements that can be mathematicallyanalyzed (MacLeod 2017) Such methods includelandmark analyses (Bookstein 1997) eigenshape ana-lysis (MacLeod 2002) and elliptical Fourier analysis(Kuhl amp Giardina 1982) These numerical approachesto describing variation in form translate morphologyinto continuous data and are consequently more suitedto the study of the subtle variation exhibited intraspe-cifically than the use of discrete characters and indeedare a prerequisite to examining fine‐scale patterns ofmorphological variability

While straightforward use of statistical variance isoften employed in studies on developmental instability(eg Imasheva Loeschcke Zhivotovsky amp Lazebny1997) novel methods for quantifying variance specificallystemming from developmental plasticity have seen recentdevelopment (Pertoldi amp Kristensen 2015 Pertoldi et al2014) Coupling shifts in variability to shifts in pheno-typic mean is the key to identifying developmentalplasticity in the fossil record as I will discuss belowThe putative confounding effect of time‐averaging onmorphological variance in fossil specimens has not beensupported by empirical studies Rather it seems thatpaleontological and neontological samples have compar-able levels of variability (Kidwell amp Flessa 1995 BushPowell Arnold Bert amp Daley 2002 Hunt 2004 Hunt2004 Hunt Bell amp Travis 2008 Webster 2019)

To approach the environmental aspect of PLE it isnecessary to identify indications of environmental con-ditions in fossil assemblages Patterns of intraspecificvariation in fossil species across a geographic andputative environmental range are by no means uncom-mon (eg Hopkins amp Webster 2009 Jackson amp Budd2017) Such patterns have been studied through the

FIGURE 3 Sheldonrsquos (1997) Plus ca Change model Stronglyfluctuating environments are predicted to increase selection forstasis until a threshold is reached and a punctuation sensuEldredge and Gould (1972) is generated This model can beinterpreted as predicting selection for increased canalization influctuating environments until a threshold is reached at whichpoint cryptic genetic variation is expressed and phenotypicaccommodation occurs

8 of 15 | JACKSON

assessment of associations between morphology andvarious environmental gradients (eg Cisne ChandleeRabe amp Cohen 1982 Scarponi amp Kowalewski 2004Webber amp Hunda 2007) Indeed a variety of geochemicalproxies are available for the investigation and approx-imation of depositional environments (Webster 2019)such as using trace metals as a proxy for paleoredox andpaleoproductivity (Tribovillard Algeo Lyons amp Riboul-leau 2006) for example However care must be takenwhen interpreting such proxies to minimize environ-mental variation generated by taphonomic processessuch as diagenetic alterations of the rock (Webster 2019)

The relationship between environmental stress anddevelopmental plasticity by proxy of morphologicalvariability is used in the following two approachestowards providing evidence of PLE as a mode ofevolution in the fossil record

14 | EVIDENCE OF PHENOTYPICAND GENETIC ACCOMMODATION

As mentioned above both phyletic gradualism and PLEpredict a shift in morphology following exposure tonovel environments albeit resulting from differentprocesses making it difficult to uncover traces defini-tively of PLE Gradual adaptive evolution wouldmanifest as an assemblage shifting towards a newphenotypic optimum in response to environmentalperturbation Likewise PLE would present a similarpattern yet with the environment both inducing as wellas selecting for novel variants

However PLE predicts a pattern of increasingvariability as the expression of CGV is induced byenvironmental stress which distinguishes it fromphyletic gradualism In this mode of evolution pheno-typic accommodation increases variability such that thephenotype subsequently genetically accommodated or

canalized lies outside its ancestral normal range (Figure4 cf Figure 1) While developmental instability isknown to occur during stress including on the rangemargins of populations (Bridle amp Vines 2007 RiesebergArcher amp Wayne 1999) such dynamics do not predictthe induction of novel variation outside the ancestralrange Consequently shifts in mean morphologycoupled with strongly increasing variability beyond anancestrally normal range can be used as an indicator ofPLE (Figure 4)

This means that the study of morphological varia-bility across environmental gradients offers an oppor-tunity to test specifically for evidence of the phenotypicand genetic accommodation inherent to PLE as de-scribed above

Naturally demonstrating this kind of intraspecificpattern requires a fossil organism or lineage with a highlyresolved stratigraphy sufficient abundance and knownpresence across temporal andor environmental gradients(Schoch 2014) Such research is not uncommon (egHopkins amp Webster 2009 Schoch 2014 Webber ampHunda 2007) yet morphological variability is rarely thefocal point and phenotypic accommodation rarer still

However the Cambrian trilobite‐like arthropod Ag-nostus pisiformis has been demonstrated to exhibitdifferentiated patterns of variation and variability acrossa gradient of dysoxic stress (Jackson amp Budd 2017)implying that environmental conditions may indeed beinducing as well as selecting for phenotypic variants

Further investigation in this framework is needed toproperly test the fossil record for evidence of PLEPatterns of phenotypic variation and chiefly variabilityuncovered by fine‐scale studies of fossil morphology (egDai Zhang Peng amp Yao 2017 Esteve 2012 NeubauerHarzhauser amp Kroh 2013) ought to be considered withinthe framework of PLE Likewise research exploringputative canalization of fossil organisms (eg PimientoTang Zamora Klug amp Saacutenchez‐Villagra 2018 Webster

Mor

phol

ogy

Time

Mor

phol

ogy

Time

(a) (b) EnvironmentalOptimum

Population

FIGURE 4 Illustration of two hypothetical responses of an evolutionary lineage to an environmental perturbation (a) Natural selectioncauses an adaptive shift of morphology towards a new environmental optimum in line with the predictions of phyletic gradualism (b) Theenvironmentally mediated release of cryptic genetic variation increases phenotypic variability facilitating rapid adaptive evolution towards anew optimum as per the process of plasticity‐led evolution Stabilizing selection subsequently leads to canalization around this newoptimum

JACKSON | 9 of 15

2015) could explicitly test for indications of PLE havingoccurred

15 | EVIDENCE OFDEVELOPMENTALLY BIASEDEVOLUTIONARY TRAJECTORIES

Adaptive radiations represent periods of rapid evolu-tion in which ancestral forms of a lineage diversify inenvironment morphology functionality and so forthWhen such patterns feature repeated evolutionarysequences such as in the case of the Anolis lizards ofthe Caribbean (Losos Jackman Larson de Queiroz ampRodriacuteguez‐Schettino 1998) they afford us the oppor-tunity to examine the fossil record for evidence ofanother component of PLE namely whether or notrepeated patterns of phenotypic evolution are some-times better explained by developmental bias thanconvergent selection At the heart of this issue lies therole of the environment in adaptive radiationsTraditionally the environment is seen as facilitatingsuch episodes by providing ecological opportunity todiversifying clades (Stroud amp Losos 2016) However asoutlined above the environment may also act as aninducer of novel variants by interacting with develop-mental processes Erwin (2017) describes these per-spectives as emphasizing ldquoenvironmental pullrdquo andldquodevelopmental pushrdquo respectively

West‐Eberhardrsquos (2003) flexible‐stem model sug-gests that plasticity mediated by the regulatorystructure of an ancestral populationrsquos developmentwill bias repeated independent diversificationsalong similar evolutionary trajectories via the pro-cesses of phenotypic and genetic accommodationThis represents an alternative or complementaryexplanation as to why independently derived lineages

of an ancestral population repeatedly evolve similarphenotypes such as the ldquoecomorphsrdquo of the Anolislizards mentioned above (Losos et al 1998 Williams1972) While this model has been explored in aneontological context (Gibert 2017 Parsons et al2016 Schneider amp Meyer 2017 Wund Baker ClancyGolub amp Foster 2008) it has not yet seen attention ina paleontological setting

Although the identification of adaptive radiations inthe fossil record is not a straightforward endeavor(Lieberman 2012) essentially any sequence of repeateddivergence from an ancestral stock population is suitablefor investigating the flexible‐stem informed model ofparallelism Similarly to the discussion on phenotypicand genetic accommodation above the focal point ofresearch in such systems should be on the variability ofmorphology rather than variation relative environmen-tal gradients If it is indeed phenotypic accommodationthat is the environmentally mediated release of CGVthat directs these repeated evolutionary trajectories weshould predict an increasing degree of phenotypicvariability to coincide with these divergences (Figure 5)This coupling of increasing variability and shiftingvariation is not conventionally anticipated in a patternof repeated parallel divergences that is convergentselection as explained above

The positive identification of such patterns in thefossil record would thus provide distinctive support forthe underlying processes of PLE both in terms of thesuggested environmentally mediated release of CGV aswell as the hypothesized bias exercised by the regulatorynature of development Research towards uncoveringsuch support should be directed towards reassessing thephenotypendashenvironment associations of studied fossiladaptive radiations (eg Abe amp Lieberman 2012 NeigeDera amp Dommergues 2013 Neubauer et al 2013) in thecontext of PLE

Mor

phol

ogy

Time

EnvironmentalOptimum

Ancestral Population

Diverging Population

FIGURE 5 Repeated divergences from an ancestral or stock population The bias native to the ancestral developmental network directsmorphological change along similar evolutionary trajectories as the successive independently diverging populations encounter the newenvironment An increase in morphological variability coupled with such parallelism of phenotypic mean change is predicted by the releaseof cryptic genetic variation in plasticity‐led evolution but not by the conventional process of convergent evolution resulting from similar andrepeated selective pressures

10 of 15 | JACKSON

16 | CONCLUDING REMARKS

Viewing development as constructive is a promisingconceptual framework for enriching our understandingof the processes and patterns of evolution It provides thescaffolding for properly addressing issues of centralimportance such as the role of developmental biases indirecting evolutionary change However the majority ofefforts towards this end are conducted on extantorganisms with minimal reference to paleontologicaldata This is unfortunate as the fossil record is afantastically rich source of information that is yet to befully explored in this intriguing and stimulating contextWhile microevolutionary concepts such as developmentalbias are challenging to approach in the fossil record Iargue that it is manifest in paleontological materials inthe form of macroevolutionary patterns concordant withthe predictions of plasticity‐led evolution Indeed suchpredictions recapitulate earlier theories of quantumevolution and punctuated equilibrium

These predictions are specifically an evolutionarymodel consisting of two major parts (a) stabilizingselection during periods of environmental stabilityleading to a reduction in phenotypic variation but anincrease in canalization CGV and hence evolutionarycapacitance and (b) episodes of environmentally inducedplasticity facilitating rapid evolution via the expression ofpreviously cryptic variation

Research aiming to uncover support for PLE in thefossil record should be directed towards the environ-mentndashphenotype association of fossil assemblages Spe-cifically towards patterns of correlation between envir-onmental stress and phenotypic variability as a proxy fordevelopmental instability These may be manifest alongsingle lineages or throughout adaptive radiations Thefossil record offers enormous potential to provide datathat when viewed through the lens of PLE providesbiologically meaningful information about evolution interms of both patterns and processes

ACKNOWLEDGMENTS

I would like to thank Armin Moczek for inviting me tocontribute to this special issue on developmental bias inevolution My thanks to two anonymous reviewers whoprovided thoughtful constructive and immensely helpfulcomments that greatly improved this paper I am gratefulto Graham Budd for our many discussions on this topicover the years I am thankful to the Uller group at LundUniversity for formative discussions particularly TobiasUller who also provided valuable comments on an earlydraft of this paper My thoughts presented here weregreatly informed by the SFI workshop ldquoDevelopmental

Bias and Evolutionrdquo and I direct my gratitude towards itsorganizers and participants This study was funded by theKnut and Alice Wallenberg Foundation grant to TobiasUller (KAW 20120155)

CONFLICT OF INTEREST

The author declares that there is no conflict of interest

ORCID

Illiam S C Jackson httporcidorg0000-0002-7948-2860

REFERENCES

Abe F R amp Lieberman B S (2012) Quantifying morphologicalchange during an evolutionary radiation of Devonian trilobitesPaleobiology 38 292ndash307 httpsdoiorg101666100471

Arthur W (2004) The effect of development on the directionof evolution Toward a twenty‐first century consensusEvolution amp Development 6(4) 282ndash288 httpsdoiorg101111j1525‐142X200404033x

Badyaev A V (2005) Stress‐induced variation in evolution Frombehavioural plasticity to genetic assimilation Proceedings of theRoyal Society B Biological Sciences 272 877ndash886 httpsdoiorg101098rspb20043045

Badyaev A V (2009) Evolutionary significance of phenotypicaccommodation in novel environments An empirical test of theBaldwin effect Philosophical Transactions of the Royal SocietyB Biological Sciences 364 1125ndash1141 httpsdoiorg101098rstb20080285

Baedke J (2018) Organism where art thou Old and new challengesfor organism‐centered biology Journal of the History of Biology52 293ndash324 httpsdoiorg101007s10739‐018‐9549‐4

Bookstein F L (1997) Morphometric tools for landmark dataGeometry and biology Cambridge Cambridge University Press

Brakefield P M (2011) Evo‐devo and accounting for Darwinsendless forms Philosophical Transactions of the Royal Society BBiological Sciences 366(1574) 2069ndash2075 httpsdoiorg101098rstb20110007

Bridle J R amp Vines T H (2007) Limits to evolution at rangemargins When and why does adaptation fail Trends in Ecologyamp Evolution 22(3) 140ndash147 httpsdoiorg101016jtree200611002

Bush A M Powell M G Arnold W S Bert T M amp Daley G M(2002) Time‐averaging evolution and morphologic variationPaleobiology 28 9ndash25 httpsdoiorg1016660094‐8373(2002)0283C0009TAEAMV3E20CO2

Carroll S B (2005) Evolution at two levels On genes and form PLOSBiology 3 e245 httpsdoiorg101371journalpbio0030245

Carroll S B (2008) Evo‐devo and an expanding evolutionarysynthesis A genetic theory of morphological evolution Cell134 25ndash36 httpsdoiorg101016jcell200806030

Charlesworth B Lande R amp Slatkin M (1982) A neo‐Darwiniancommentary on macroevolution Evolution 36 474ndash498 httpsdoiorg1023072408095

JACKSON | 11 of 15

This distinction is key for separating the predictions ofPLE and phyletic gradualism Variation that is morpho-logical differences between assemblages associated withenvironmental change is insufficient to support PLE forthe reasons discussed above However an associationbetween environmental change and variability that isthe variance in form within assemblages informs usabout their developmental stability and thus permitsinvestigation into the unique prediction of PLE thatenvironmental stress may overcome canalization andinduce the release of CGV with consequent expression ofnovel variants upon which selection can act Theconsequence of this process is an increase in phenotypicvariability coupled with a shift in phenotypic meanassociated with environmental stress a pattern notpredicted by phyletic gradualism

Below I provide two examples of approaches towardsproviding evidence of PLE in the fossil record that usethis association Before discussing them however I firstpresent their necessary methodological framework

13 | FRAMEWORK OF ANALYSIS

The focal point for these investigations is the relation-ship between environmental factors and morphology offossil assemblages Morphology is essentially the onlybiological information available in the fossil record andit is therefore in general of paramount importance tocapture and study morphology in as high resolution aspossible and with careful consideration of potentialtaphonomic biases such as disarticulation of hardparts or later structural deformation (Webster 2019)This is particularly true however when studyingintraspecific patterns of variation such as thoseI will present below Many modern phylogeneticcladistic analyses rely on discretizing variation intocharacters which strongly limits their ability to studyintraspecific variation An alternative approach toanalyzing morphological variation is through the useof morphometric methods which quantify shapes intosets of measurements that can be mathematicallyanalyzed (MacLeod 2017) Such methods includelandmark analyses (Bookstein 1997) eigenshape ana-lysis (MacLeod 2002) and elliptical Fourier analysis(Kuhl amp Giardina 1982) These numerical approachesto describing variation in form translate morphologyinto continuous data and are consequently more suitedto the study of the subtle variation exhibited intraspe-cifically than the use of discrete characters and indeedare a prerequisite to examining fine‐scale patterns ofmorphological variability

While straightforward use of statistical variance isoften employed in studies on developmental instability(eg Imasheva Loeschcke Zhivotovsky amp Lazebny1997) novel methods for quantifying variance specificallystemming from developmental plasticity have seen recentdevelopment (Pertoldi amp Kristensen 2015 Pertoldi et al2014) Coupling shifts in variability to shifts in pheno-typic mean is the key to identifying developmentalplasticity in the fossil record as I will discuss belowThe putative confounding effect of time‐averaging onmorphological variance in fossil specimens has not beensupported by empirical studies Rather it seems thatpaleontological and neontological samples have compar-able levels of variability (Kidwell amp Flessa 1995 BushPowell Arnold Bert amp Daley 2002 Hunt 2004 Hunt2004 Hunt Bell amp Travis 2008 Webster 2019)

To approach the environmental aspect of PLE it isnecessary to identify indications of environmental con-ditions in fossil assemblages Patterns of intraspecificvariation in fossil species across a geographic andputative environmental range are by no means uncom-mon (eg Hopkins amp Webster 2009 Jackson amp Budd2017) Such patterns have been studied through the

FIGURE 3 Sheldonrsquos (1997) Plus ca Change model Stronglyfluctuating environments are predicted to increase selection forstasis until a threshold is reached and a punctuation sensuEldredge and Gould (1972) is generated This model can beinterpreted as predicting selection for increased canalization influctuating environments until a threshold is reached at whichpoint cryptic genetic variation is expressed and phenotypicaccommodation occurs

8 of 15 | JACKSON

assessment of associations between morphology andvarious environmental gradients (eg Cisne ChandleeRabe amp Cohen 1982 Scarponi amp Kowalewski 2004Webber amp Hunda 2007) Indeed a variety of geochemicalproxies are available for the investigation and approx-imation of depositional environments (Webster 2019)such as using trace metals as a proxy for paleoredox andpaleoproductivity (Tribovillard Algeo Lyons amp Riboul-leau 2006) for example However care must be takenwhen interpreting such proxies to minimize environ-mental variation generated by taphonomic processessuch as diagenetic alterations of the rock (Webster 2019)

The relationship between environmental stress anddevelopmental plasticity by proxy of morphologicalvariability is used in the following two approachestowards providing evidence of PLE as a mode ofevolution in the fossil record

14 | EVIDENCE OF PHENOTYPICAND GENETIC ACCOMMODATION

As mentioned above both phyletic gradualism and PLEpredict a shift in morphology following exposure tonovel environments albeit resulting from differentprocesses making it difficult to uncover traces defini-tively of PLE Gradual adaptive evolution wouldmanifest as an assemblage shifting towards a newphenotypic optimum in response to environmentalperturbation Likewise PLE would present a similarpattern yet with the environment both inducing as wellas selecting for novel variants

However PLE predicts a pattern of increasingvariability as the expression of CGV is induced byenvironmental stress which distinguishes it fromphyletic gradualism In this mode of evolution pheno-typic accommodation increases variability such that thephenotype subsequently genetically accommodated or

canalized lies outside its ancestral normal range (Figure4 cf Figure 1) While developmental instability isknown to occur during stress including on the rangemargins of populations (Bridle amp Vines 2007 RiesebergArcher amp Wayne 1999) such dynamics do not predictthe induction of novel variation outside the ancestralrange Consequently shifts in mean morphologycoupled with strongly increasing variability beyond anancestrally normal range can be used as an indicator ofPLE (Figure 4)

This means that the study of morphological varia-bility across environmental gradients offers an oppor-tunity to test specifically for evidence of the phenotypicand genetic accommodation inherent to PLE as de-scribed above

Naturally demonstrating this kind of intraspecificpattern requires a fossil organism or lineage with a highlyresolved stratigraphy sufficient abundance and knownpresence across temporal andor environmental gradients(Schoch 2014) Such research is not uncommon (egHopkins amp Webster 2009 Schoch 2014 Webber ampHunda 2007) yet morphological variability is rarely thefocal point and phenotypic accommodation rarer still

However the Cambrian trilobite‐like arthropod Ag-nostus pisiformis has been demonstrated to exhibitdifferentiated patterns of variation and variability acrossa gradient of dysoxic stress (Jackson amp Budd 2017)implying that environmental conditions may indeed beinducing as well as selecting for phenotypic variants

Further investigation in this framework is needed toproperly test the fossil record for evidence of PLEPatterns of phenotypic variation and chiefly variabilityuncovered by fine‐scale studies of fossil morphology (egDai Zhang Peng amp Yao 2017 Esteve 2012 NeubauerHarzhauser amp Kroh 2013) ought to be considered withinthe framework of PLE Likewise research exploringputative canalization of fossil organisms (eg PimientoTang Zamora Klug amp Saacutenchez‐Villagra 2018 Webster

Mor

phol

ogy

Time

Mor

phol

ogy

Time

(a) (b) EnvironmentalOptimum

Population

FIGURE 4 Illustration of two hypothetical responses of an evolutionary lineage to an environmental perturbation (a) Natural selectioncauses an adaptive shift of morphology towards a new environmental optimum in line with the predictions of phyletic gradualism (b) Theenvironmentally mediated release of cryptic genetic variation increases phenotypic variability facilitating rapid adaptive evolution towards anew optimum as per the process of plasticity‐led evolution Stabilizing selection subsequently leads to canalization around this newoptimum

JACKSON | 9 of 15

2015) could explicitly test for indications of PLE havingoccurred

15 | EVIDENCE OFDEVELOPMENTALLY BIASEDEVOLUTIONARY TRAJECTORIES

Adaptive radiations represent periods of rapid evolu-tion in which ancestral forms of a lineage diversify inenvironment morphology functionality and so forthWhen such patterns feature repeated evolutionarysequences such as in the case of the Anolis lizards ofthe Caribbean (Losos Jackman Larson de Queiroz ampRodriacuteguez‐Schettino 1998) they afford us the oppor-tunity to examine the fossil record for evidence ofanother component of PLE namely whether or notrepeated patterns of phenotypic evolution are some-times better explained by developmental bias thanconvergent selection At the heart of this issue lies therole of the environment in adaptive radiationsTraditionally the environment is seen as facilitatingsuch episodes by providing ecological opportunity todiversifying clades (Stroud amp Losos 2016) However asoutlined above the environment may also act as aninducer of novel variants by interacting with develop-mental processes Erwin (2017) describes these per-spectives as emphasizing ldquoenvironmental pullrdquo andldquodevelopmental pushrdquo respectively

West‐Eberhardrsquos (2003) flexible‐stem model sug-gests that plasticity mediated by the regulatorystructure of an ancestral populationrsquos developmentwill bias repeated independent diversificationsalong similar evolutionary trajectories via the pro-cesses of phenotypic and genetic accommodationThis represents an alternative or complementaryexplanation as to why independently derived lineages

of an ancestral population repeatedly evolve similarphenotypes such as the ldquoecomorphsrdquo of the Anolislizards mentioned above (Losos et al 1998 Williams1972) While this model has been explored in aneontological context (Gibert 2017 Parsons et al2016 Schneider amp Meyer 2017 Wund Baker ClancyGolub amp Foster 2008) it has not yet seen attention ina paleontological setting

Although the identification of adaptive radiations inthe fossil record is not a straightforward endeavor(Lieberman 2012) essentially any sequence of repeateddivergence from an ancestral stock population is suitablefor investigating the flexible‐stem informed model ofparallelism Similarly to the discussion on phenotypicand genetic accommodation above the focal point ofresearch in such systems should be on the variability ofmorphology rather than variation relative environmen-tal gradients If it is indeed phenotypic accommodationthat is the environmentally mediated release of CGVthat directs these repeated evolutionary trajectories weshould predict an increasing degree of phenotypicvariability to coincide with these divergences (Figure 5)This coupling of increasing variability and shiftingvariation is not conventionally anticipated in a patternof repeated parallel divergences that is convergentselection as explained above

The positive identification of such patterns in thefossil record would thus provide distinctive support forthe underlying processes of PLE both in terms of thesuggested environmentally mediated release of CGV aswell as the hypothesized bias exercised by the regulatorynature of development Research towards uncoveringsuch support should be directed towards reassessing thephenotypendashenvironment associations of studied fossiladaptive radiations (eg Abe amp Lieberman 2012 NeigeDera amp Dommergues 2013 Neubauer et al 2013) in thecontext of PLE

Mor

phol

ogy

Time

EnvironmentalOptimum

Ancestral Population

Diverging Population

FIGURE 5 Repeated divergences from an ancestral or stock population The bias native to the ancestral developmental network directsmorphological change along similar evolutionary trajectories as the successive independently diverging populations encounter the newenvironment An increase in morphological variability coupled with such parallelism of phenotypic mean change is predicted by the releaseof cryptic genetic variation in plasticity‐led evolution but not by the conventional process of convergent evolution resulting from similar andrepeated selective pressures

10 of 15 | JACKSON

16 | CONCLUDING REMARKS

Viewing development as constructive is a promisingconceptual framework for enriching our understandingof the processes and patterns of evolution It provides thescaffolding for properly addressing issues of centralimportance such as the role of developmental biases indirecting evolutionary change However the majority ofefforts towards this end are conducted on extantorganisms with minimal reference to paleontologicaldata This is unfortunate as the fossil record is afantastically rich source of information that is yet to befully explored in this intriguing and stimulating contextWhile microevolutionary concepts such as developmentalbias are challenging to approach in the fossil record Iargue that it is manifest in paleontological materials inthe form of macroevolutionary patterns concordant withthe predictions of plasticity‐led evolution Indeed suchpredictions recapitulate earlier theories of quantumevolution and punctuated equilibrium

These predictions are specifically an evolutionarymodel consisting of two major parts (a) stabilizingselection during periods of environmental stabilityleading to a reduction in phenotypic variation but anincrease in canalization CGV and hence evolutionarycapacitance and (b) episodes of environmentally inducedplasticity facilitating rapid evolution via the expression ofpreviously cryptic variation

Research aiming to uncover support for PLE in thefossil record should be directed towards the environ-mentndashphenotype association of fossil assemblages Spe-cifically towards patterns of correlation between envir-onmental stress and phenotypic variability as a proxy fordevelopmental instability These may be manifest alongsingle lineages or throughout adaptive radiations Thefossil record offers enormous potential to provide datathat when viewed through the lens of PLE providesbiologically meaningful information about evolution interms of both patterns and processes

ACKNOWLEDGMENTS

I would like to thank Armin Moczek for inviting me tocontribute to this special issue on developmental bias inevolution My thanks to two anonymous reviewers whoprovided thoughtful constructive and immensely helpfulcomments that greatly improved this paper I am gratefulto Graham Budd for our many discussions on this topicover the years I am thankful to the Uller group at LundUniversity for formative discussions particularly TobiasUller who also provided valuable comments on an earlydraft of this paper My thoughts presented here weregreatly informed by the SFI workshop ldquoDevelopmental

Bias and Evolutionrdquo and I direct my gratitude towards itsorganizers and participants This study was funded by theKnut and Alice Wallenberg Foundation grant to TobiasUller (KAW 20120155)

CONFLICT OF INTEREST

The author declares that there is no conflict of interest

ORCID

Illiam S C Jackson httporcidorg0000-0002-7948-2860

REFERENCES

Abe F R amp Lieberman B S (2012) Quantifying morphologicalchange during an evolutionary radiation of Devonian trilobitesPaleobiology 38 292ndash307 httpsdoiorg101666100471

Arthur W (2004) The effect of development on the directionof evolution Toward a twenty‐first century consensusEvolution amp Development 6(4) 282ndash288 httpsdoiorg101111j1525‐142X200404033x

Badyaev A V (2005) Stress‐induced variation in evolution Frombehavioural plasticity to genetic assimilation Proceedings of theRoyal Society B Biological Sciences 272 877ndash886 httpsdoiorg101098rspb20043045

Badyaev A V (2009) Evolutionary significance of phenotypicaccommodation in novel environments An empirical test of theBaldwin effect Philosophical Transactions of the Royal SocietyB Biological Sciences 364 1125ndash1141 httpsdoiorg101098rstb20080285

Baedke J (2018) Organism where art thou Old and new challengesfor organism‐centered biology Journal of the History of Biology52 293ndash324 httpsdoiorg101007s10739‐018‐9549‐4

Bookstein F L (1997) Morphometric tools for landmark dataGeometry and biology Cambridge Cambridge University Press

Brakefield P M (2011) Evo‐devo and accounting for Darwinsendless forms Philosophical Transactions of the Royal Society BBiological Sciences 366(1574) 2069ndash2075 httpsdoiorg101098rstb20110007

Bridle J R amp Vines T H (2007) Limits to evolution at rangemargins When and why does adaptation fail Trends in Ecologyamp Evolution 22(3) 140ndash147 httpsdoiorg101016jtree200611002

Bush A M Powell M G Arnold W S Bert T M amp Daley G M(2002) Time‐averaging evolution and morphologic variationPaleobiology 28 9ndash25 httpsdoiorg1016660094‐8373(2002)0283C0009TAEAMV3E20CO2

Carroll S B (2005) Evolution at two levels On genes and form PLOSBiology 3 e245 httpsdoiorg101371journalpbio0030245

Carroll S B (2008) Evo‐devo and an expanding evolutionarysynthesis A genetic theory of morphological evolution Cell134 25ndash36 httpsdoiorg101016jcell200806030

Charlesworth B Lande R amp Slatkin M (1982) A neo‐Darwiniancommentary on macroevolution Evolution 36 474ndash498 httpsdoiorg1023072408095

JACKSON | 11 of 15

assessment of associations between morphology andvarious environmental gradients (eg Cisne ChandleeRabe amp Cohen 1982 Scarponi amp Kowalewski 2004Webber amp Hunda 2007) Indeed a variety of geochemicalproxies are available for the investigation and approx-imation of depositional environments (Webster 2019)such as using trace metals as a proxy for paleoredox andpaleoproductivity (Tribovillard Algeo Lyons amp Riboul-leau 2006) for example However care must be takenwhen interpreting such proxies to minimize environ-mental variation generated by taphonomic processessuch as diagenetic alterations of the rock (Webster 2019)

The relationship between environmental stress anddevelopmental plasticity by proxy of morphologicalvariability is used in the following two approachestowards providing evidence of PLE as a mode ofevolution in the fossil record

14 | EVIDENCE OF PHENOTYPICAND GENETIC ACCOMMODATION

As mentioned above both phyletic gradualism and PLEpredict a shift in morphology following exposure tonovel environments albeit resulting from differentprocesses making it difficult to uncover traces defini-tively of PLE Gradual adaptive evolution wouldmanifest as an assemblage shifting towards a newphenotypic optimum in response to environmentalperturbation Likewise PLE would present a similarpattern yet with the environment both inducing as wellas selecting for novel variants

However PLE predicts a pattern of increasingvariability as the expression of CGV is induced byenvironmental stress which distinguishes it fromphyletic gradualism In this mode of evolution pheno-typic accommodation increases variability such that thephenotype subsequently genetically accommodated or

canalized lies outside its ancestral normal range (Figure4 cf Figure 1) While developmental instability isknown to occur during stress including on the rangemargins of populations (Bridle amp Vines 2007 RiesebergArcher amp Wayne 1999) such dynamics do not predictthe induction of novel variation outside the ancestralrange Consequently shifts in mean morphologycoupled with strongly increasing variability beyond anancestrally normal range can be used as an indicator ofPLE (Figure 4)

This means that the study of morphological varia-bility across environmental gradients offers an oppor-tunity to test specifically for evidence of the phenotypicand genetic accommodation inherent to PLE as de-scribed above

Naturally demonstrating this kind of intraspecificpattern requires a fossil organism or lineage with a highlyresolved stratigraphy sufficient abundance and knownpresence across temporal andor environmental gradients(Schoch 2014) Such research is not uncommon (egHopkins amp Webster 2009 Schoch 2014 Webber ampHunda 2007) yet morphological variability is rarely thefocal point and phenotypic accommodation rarer still

However the Cambrian trilobite‐like arthropod Ag-nostus pisiformis has been demonstrated to exhibitdifferentiated patterns of variation and variability acrossa gradient of dysoxic stress (Jackson amp Budd 2017)implying that environmental conditions may indeed beinducing as well as selecting for phenotypic variants

Further investigation in this framework is needed toproperly test the fossil record for evidence of PLEPatterns of phenotypic variation and chiefly variabilityuncovered by fine‐scale studies of fossil morphology (egDai Zhang Peng amp Yao 2017 Esteve 2012 NeubauerHarzhauser amp Kroh 2013) ought to be considered withinthe framework of PLE Likewise research exploringputative canalization of fossil organisms (eg PimientoTang Zamora Klug amp Saacutenchez‐Villagra 2018 Webster

Mor

phol

ogy

Time

Mor

phol

ogy

Time

(a) (b) EnvironmentalOptimum

Population

FIGURE 4 Illustration of two hypothetical responses of an evolutionary lineage to an environmental perturbation (a) Natural selectioncauses an adaptive shift of morphology towards a new environmental optimum in line with the predictions of phyletic gradualism (b) Theenvironmentally mediated release of cryptic genetic variation increases phenotypic variability facilitating rapid adaptive evolution towards anew optimum as per the process of plasticity‐led evolution Stabilizing selection subsequently leads to canalization around this newoptimum

JACKSON | 9 of 15

2015) could explicitly test for indications of PLE havingoccurred

15 | EVIDENCE OFDEVELOPMENTALLY BIASEDEVOLUTIONARY TRAJECTORIES

Adaptive radiations represent periods of rapid evolu-tion in which ancestral forms of a lineage diversify inenvironment morphology functionality and so forthWhen such patterns feature repeated evolutionarysequences such as in the case of the Anolis lizards ofthe Caribbean (Losos Jackman Larson de Queiroz ampRodriacuteguez‐Schettino 1998) they afford us the oppor-tunity to examine the fossil record for evidence ofanother component of PLE namely whether or notrepeated patterns of phenotypic evolution are some-times better explained by developmental bias thanconvergent selection At the heart of this issue lies therole of the environment in adaptive radiationsTraditionally the environment is seen as facilitatingsuch episodes by providing ecological opportunity todiversifying clades (Stroud amp Losos 2016) However asoutlined above the environment may also act as aninducer of novel variants by interacting with develop-mental processes Erwin (2017) describes these per-spectives as emphasizing ldquoenvironmental pullrdquo andldquodevelopmental pushrdquo respectively

West‐Eberhardrsquos (2003) flexible‐stem model sug-gests that plasticity mediated by the regulatorystructure of an ancestral populationrsquos developmentwill bias repeated independent diversificationsalong similar evolutionary trajectories via the pro-cesses of phenotypic and genetic accommodationThis represents an alternative or complementaryexplanation as to why independently derived lineages

of an ancestral population repeatedly evolve similarphenotypes such as the ldquoecomorphsrdquo of the Anolislizards mentioned above (Losos et al 1998 Williams1972) While this model has been explored in aneontological context (Gibert 2017 Parsons et al2016 Schneider amp Meyer 2017 Wund Baker ClancyGolub amp Foster 2008) it has not yet seen attention ina paleontological setting

Although the identification of adaptive radiations inthe fossil record is not a straightforward endeavor(Lieberman 2012) essentially any sequence of repeateddivergence from an ancestral stock population is suitablefor investigating the flexible‐stem informed model ofparallelism Similarly to the discussion on phenotypicand genetic accommodation above the focal point ofresearch in such systems should be on the variability ofmorphology rather than variation relative environmen-tal gradients If it is indeed phenotypic accommodationthat is the environmentally mediated release of CGVthat directs these repeated evolutionary trajectories weshould predict an increasing degree of phenotypicvariability to coincide with these divergences (Figure 5)This coupling of increasing variability and shiftingvariation is not conventionally anticipated in a patternof repeated parallel divergences that is convergentselection as explained above

The positive identification of such patterns in thefossil record would thus provide distinctive support forthe underlying processes of PLE both in terms of thesuggested environmentally mediated release of CGV aswell as the hypothesized bias exercised by the regulatorynature of development Research towards uncoveringsuch support should be directed towards reassessing thephenotypendashenvironment associations of studied fossiladaptive radiations (eg Abe amp Lieberman 2012 NeigeDera amp Dommergues 2013 Neubauer et al 2013) in thecontext of PLE

Mor

phol

ogy

Time

EnvironmentalOptimum

Ancestral Population

Diverging Population

FIGURE 5 Repeated divergences from an ancestral or stock population The bias native to the ancestral developmental network directsmorphological change along similar evolutionary trajectories as the successive independently diverging populations encounter the newenvironment An increase in morphological variability coupled with such parallelism of phenotypic mean change is predicted by the releaseof cryptic genetic variation in plasticity‐led evolution but not by the conventional process of convergent evolution resulting from similar andrepeated selective pressures

10 of 15 | JACKSON

16 | CONCLUDING REMARKS

Viewing development as constructive is a promisingconceptual framework for enriching our understandingof the processes and patterns of evolution It provides thescaffolding for properly addressing issues of centralimportance such as the role of developmental biases indirecting evolutionary change However the majority ofefforts towards this end are conducted on extantorganisms with minimal reference to paleontologicaldata This is unfortunate as the fossil record is afantastically rich source of information that is yet to befully explored in this intriguing and stimulating contextWhile microevolutionary concepts such as developmentalbias are challenging to approach in the fossil record Iargue that it is manifest in paleontological materials inthe form of macroevolutionary patterns concordant withthe predictions of plasticity‐led evolution Indeed suchpredictions recapitulate earlier theories of quantumevolution and punctuated equilibrium

These predictions are specifically an evolutionarymodel consisting of two major parts (a) stabilizingselection during periods of environmental stabilityleading to a reduction in phenotypic variation but anincrease in canalization CGV and hence evolutionarycapacitance and (b) episodes of environmentally inducedplasticity facilitating rapid evolution via the expression ofpreviously cryptic variation

Research aiming to uncover support for PLE in thefossil record should be directed towards the environ-mentndashphenotype association of fossil assemblages Spe-cifically towards patterns of correlation between envir-onmental stress and phenotypic variability as a proxy fordevelopmental instability These may be manifest alongsingle lineages or throughout adaptive radiations Thefossil record offers enormous potential to provide datathat when viewed through the lens of PLE providesbiologically meaningful information about evolution interms of both patterns and processes

ACKNOWLEDGMENTS

I would like to thank Armin Moczek for inviting me tocontribute to this special issue on developmental bias inevolution My thanks to two anonymous reviewers whoprovided thoughtful constructive and immensely helpfulcomments that greatly improved this paper I am gratefulto Graham Budd for our many discussions on this topicover the years I am thankful to the Uller group at LundUniversity for formative discussions particularly TobiasUller who also provided valuable comments on an earlydraft of this paper My thoughts presented here weregreatly informed by the SFI workshop ldquoDevelopmental

Bias and Evolutionrdquo and I direct my gratitude towards itsorganizers and participants This study was funded by theKnut and Alice Wallenberg Foundation grant to TobiasUller (KAW 20120155)

CONFLICT OF INTEREST

The author declares that there is no conflict of interest

ORCID

Illiam S C Jackson httporcidorg0000-0002-7948-2860

REFERENCES

Abe F R amp Lieberman B S (2012) Quantifying morphologicalchange during an evolutionary radiation of Devonian trilobitesPaleobiology 38 292ndash307 httpsdoiorg101666100471

Arthur W (2004) The effect of development on the directionof evolution Toward a twenty‐first century consensusEvolution amp Development 6(4) 282ndash288 httpsdoiorg101111j1525‐142X200404033x

Badyaev A V (2005) Stress‐induced variation in evolution Frombehavioural plasticity to genetic assimilation Proceedings of theRoyal Society B Biological Sciences 272 877ndash886 httpsdoiorg101098rspb20043045

Badyaev A V (2009) Evolutionary significance of phenotypicaccommodation in novel environments An empirical test of theBaldwin effect Philosophical Transactions of the Royal SocietyB Biological Sciences 364 1125ndash1141 httpsdoiorg101098rstb20080285

Baedke J (2018) Organism where art thou Old and new challengesfor organism‐centered biology Journal of the History of Biology52 293ndash324 httpsdoiorg101007s10739‐018‐9549‐4

Bookstein F L (1997) Morphometric tools for landmark dataGeometry and biology Cambridge Cambridge University Press

Brakefield P M (2011) Evo‐devo and accounting for Darwinsendless forms Philosophical Transactions of the Royal Society BBiological Sciences 366(1574) 2069ndash2075 httpsdoiorg101098rstb20110007

Bridle J R amp Vines T H (2007) Limits to evolution at rangemargins When and why does adaptation fail Trends in Ecologyamp Evolution 22(3) 140ndash147 httpsdoiorg101016jtree200611002

Bush A M Powell M G Arnold W S Bert T M amp Daley G M(2002) Time‐averaging evolution and morphologic variationPaleobiology 28 9ndash25 httpsdoiorg1016660094‐8373(2002)0283C0009TAEAMV3E20CO2

Carroll S B (2005) Evolution at two levels On genes and form PLOSBiology 3 e245 httpsdoiorg101371journalpbio0030245

Carroll S B (2008) Evo‐devo and an expanding evolutionarysynthesis A genetic theory of morphological evolution Cell134 25ndash36 httpsdoiorg101016jcell200806030

Charlesworth B Lande R amp Slatkin M (1982) A neo‐Darwiniancommentary on macroevolution Evolution 36 474ndash498 httpsdoiorg1023072408095

JACKSON | 11 of 15

2015) could explicitly test for indications of PLE havingoccurred

15 | EVIDENCE OFDEVELOPMENTALLY BIASEDEVOLUTIONARY TRAJECTORIES

Adaptive radiations represent periods of rapid evolu-tion in which ancestral forms of a lineage diversify inenvironment morphology functionality and so forthWhen such patterns feature repeated evolutionarysequences such as in the case of the Anolis lizards ofthe Caribbean (Losos Jackman Larson de Queiroz ampRodriacuteguez‐Schettino 1998) they afford us the oppor-tunity to examine the fossil record for evidence ofanother component of PLE namely whether or notrepeated patterns of phenotypic evolution are some-times better explained by developmental bias thanconvergent selection At the heart of this issue lies therole of the environment in adaptive radiationsTraditionally the environment is seen as facilitatingsuch episodes by providing ecological opportunity todiversifying clades (Stroud amp Losos 2016) However asoutlined above the environment may also act as aninducer of novel variants by interacting with develop-mental processes Erwin (2017) describes these per-spectives as emphasizing ldquoenvironmental pullrdquo andldquodevelopmental pushrdquo respectively

West‐Eberhardrsquos (2003) flexible‐stem model sug-gests that plasticity mediated by the regulatorystructure of an ancestral populationrsquos developmentwill bias repeated independent diversificationsalong similar evolutionary trajectories via the pro-cesses of phenotypic and genetic accommodationThis represents an alternative or complementaryexplanation as to why independently derived lineages

of an ancestral population repeatedly evolve similarphenotypes such as the ldquoecomorphsrdquo of the Anolislizards mentioned above (Losos et al 1998 Williams1972) While this model has been explored in aneontological context (Gibert 2017 Parsons et al2016 Schneider amp Meyer 2017 Wund Baker ClancyGolub amp Foster 2008) it has not yet seen attention ina paleontological setting

Although the identification of adaptive radiations inthe fossil record is not a straightforward endeavor(Lieberman 2012) essentially any sequence of repeateddivergence from an ancestral stock population is suitablefor investigating the flexible‐stem informed model ofparallelism Similarly to the discussion on phenotypicand genetic accommodation above the focal point ofresearch in such systems should be on the variability ofmorphology rather than variation relative environmen-tal gradients If it is indeed phenotypic accommodationthat is the environmentally mediated release of CGVthat directs these repeated evolutionary trajectories weshould predict an increasing degree of phenotypicvariability to coincide with these divergences (Figure 5)This coupling of increasing variability and shiftingvariation is not conventionally anticipated in a patternof repeated parallel divergences that is convergentselection as explained above

The positive identification of such patterns in thefossil record would thus provide distinctive support forthe underlying processes of PLE both in terms of thesuggested environmentally mediated release of CGV aswell as the hypothesized bias exercised by the regulatorynature of development Research towards uncoveringsuch support should be directed towards reassessing thephenotypendashenvironment associations of studied fossiladaptive radiations (eg Abe amp Lieberman 2012 NeigeDera amp Dommergues 2013 Neubauer et al 2013) in thecontext of PLE

Mor

phol

ogy

Time

EnvironmentalOptimum

Ancestral Population

Diverging Population

FIGURE 5 Repeated divergences from an ancestral or stock population The bias native to the ancestral developmental network directsmorphological change along similar evolutionary trajectories as the successive independently diverging populations encounter the newenvironment An increase in morphological variability coupled with such parallelism of phenotypic mean change is predicted by the releaseof cryptic genetic variation in plasticity‐led evolution but not by the conventional process of convergent evolution resulting from similar andrepeated selective pressures

10 of 15 | JACKSON

16 | CONCLUDING REMARKS

Viewing development as constructive is a promisingconceptual framework for enriching our understandingof the processes and patterns of evolution It provides thescaffolding for properly addressing issues of centralimportance such as the role of developmental biases indirecting evolutionary change However the majority ofefforts towards this end are conducted on extantorganisms with minimal reference to paleontologicaldata This is unfortunate as the fossil record is afantastically rich source of information that is yet to befully explored in this intriguing and stimulating contextWhile microevolutionary concepts such as developmentalbias are challenging to approach in the fossil record Iargue that it is manifest in paleontological materials inthe form of macroevolutionary patterns concordant withthe predictions of plasticity‐led evolution Indeed suchpredictions recapitulate earlier theories of quantumevolution and punctuated equilibrium

These predictions are specifically an evolutionarymodel consisting of two major parts (a) stabilizingselection during periods of environmental stabilityleading to a reduction in phenotypic variation but anincrease in canalization CGV and hence evolutionarycapacitance and (b) episodes of environmentally inducedplasticity facilitating rapid evolution via the expression ofpreviously cryptic variation

Research aiming to uncover support for PLE in thefossil record should be directed towards the environ-mentndashphenotype association of fossil assemblages Spe-cifically towards patterns of correlation between envir-onmental stress and phenotypic variability as a proxy fordevelopmental instability These may be manifest alongsingle lineages or throughout adaptive radiations Thefossil record offers enormous potential to provide datathat when viewed through the lens of PLE providesbiologically meaningful information about evolution interms of both patterns and processes

ACKNOWLEDGMENTS

I would like to thank Armin Moczek for inviting me tocontribute to this special issue on developmental bias inevolution My thanks to two anonymous reviewers whoprovided thoughtful constructive and immensely helpfulcomments that greatly improved this paper I am gratefulto Graham Budd for our many discussions on this topicover the years I am thankful to the Uller group at LundUniversity for formative discussions particularly TobiasUller who also provided valuable comments on an earlydraft of this paper My thoughts presented here weregreatly informed by the SFI workshop ldquoDevelopmental

Bias and Evolutionrdquo and I direct my gratitude towards itsorganizers and participants This study was funded by theKnut and Alice Wallenberg Foundation grant to TobiasUller (KAW 20120155)

CONFLICT OF INTEREST

The author declares that there is no conflict of interest

ORCID

Illiam S C Jackson httporcidorg0000-0002-7948-2860

REFERENCES

Abe F R amp Lieberman B S (2012) Quantifying morphologicalchange during an evolutionary radiation of Devonian trilobitesPaleobiology 38 292ndash307 httpsdoiorg101666100471

Arthur W (2004) The effect of development on the directionof evolution Toward a twenty‐first century consensusEvolution amp Development 6(4) 282ndash288 httpsdoiorg101111j1525‐142X200404033x

Badyaev A V (2005) Stress‐induced variation in evolution Frombehavioural plasticity to genetic assimilation Proceedings of theRoyal Society B Biological Sciences 272 877ndash886 httpsdoiorg101098rspb20043045

Badyaev A V (2009) Evolutionary significance of phenotypicaccommodation in novel environments An empirical test of theBaldwin effect Philosophical Transactions of the Royal SocietyB Biological Sciences 364 1125ndash1141 httpsdoiorg101098rstb20080285

Baedke J (2018) Organism where art thou Old and new challengesfor organism‐centered biology Journal of the History of Biology52 293ndash324 httpsdoiorg101007s10739‐018‐9549‐4

Bookstein F L (1997) Morphometric tools for landmark dataGeometry and biology Cambridge Cambridge University Press

Brakefield P M (2011) Evo‐devo and accounting for Darwinsendless forms Philosophical Transactions of the Royal Society BBiological Sciences 366(1574) 2069ndash2075 httpsdoiorg101098rstb20110007

Bridle J R amp Vines T H (2007) Limits to evolution at rangemargins When and why does adaptation fail Trends in Ecologyamp Evolution 22(3) 140ndash147 httpsdoiorg101016jtree200611002

Bush A M Powell M G Arnold W S Bert T M amp Daley G M(2002) Time‐averaging evolution and morphologic variationPaleobiology 28 9ndash25 httpsdoiorg1016660094‐8373(2002)0283C0009TAEAMV3E20CO2

Carroll S B (2005) Evolution at two levels On genes and form PLOSBiology 3 e245 httpsdoiorg101371journalpbio0030245

Carroll S B (2008) Evo‐devo and an expanding evolutionarysynthesis A genetic theory of morphological evolution Cell134 25ndash36 httpsdoiorg101016jcell200806030

Charlesworth B Lande R amp Slatkin M (1982) A neo‐Darwiniancommentary on macroevolution Evolution 36 474ndash498 httpsdoiorg1023072408095

JACKSON | 11 of 15

16 | CONCLUDING REMARKS

Viewing development as constructive is a promisingconceptual framework for enriching our understandingof the processes and patterns of evolution It provides thescaffolding for properly addressing issues of centralimportance such as the role of developmental biases indirecting evolutionary change However the majority ofefforts towards this end are conducted on extantorganisms with minimal reference to paleontologicaldata This is unfortunate as the fossil record is afantastically rich source of information that is yet to befully explored in this intriguing and stimulating contextWhile microevolutionary concepts such as developmentalbias are challenging to approach in the fossil record Iargue that it is manifest in paleontological materials inthe form of macroevolutionary patterns concordant withthe predictions of plasticity‐led evolution Indeed suchpredictions recapitulate earlier theories of quantumevolution and punctuated equilibrium

These predictions are specifically an evolutionarymodel consisting of two major parts (a) stabilizingselection during periods of environmental stabilityleading to a reduction in phenotypic variation but anincrease in canalization CGV and hence evolutionarycapacitance and (b) episodes of environmentally inducedplasticity facilitating rapid evolution via the expression ofpreviously cryptic variation

Research aiming to uncover support for PLE in thefossil record should be directed towards the environ-mentndashphenotype association of fossil assemblages Spe-cifically towards patterns of correlation between envir-onmental stress and phenotypic variability as a proxy fordevelopmental instability These may be manifest alongsingle lineages or throughout adaptive radiations Thefossil record offers enormous potential to provide datathat when viewed through the lens of PLE providesbiologically meaningful information about evolution interms of both patterns and processes

ACKNOWLEDGMENTS

I would like to thank Armin Moczek for inviting me tocontribute to this special issue on developmental bias inevolution My thanks to two anonymous reviewers whoprovided thoughtful constructive and immensely helpfulcomments that greatly improved this paper I am gratefulto Graham Budd for our many discussions on this topicover the years I am thankful to the Uller group at LundUniversity for formative discussions particularly TobiasUller who also provided valuable comments on an earlydraft of this paper My thoughts presented here weregreatly informed by the SFI workshop ldquoDevelopmental

Bias and Evolutionrdquo and I direct my gratitude towards itsorganizers and participants This study was funded by theKnut and Alice Wallenberg Foundation grant to TobiasUller (KAW 20120155)

CONFLICT OF INTEREST

The author declares that there is no conflict of interest

ORCID

Illiam S C Jackson httporcidorg0000-0002-7948-2860

REFERENCES

Abe F R amp Lieberman B S (2012) Quantifying morphologicalchange during an evolutionary radiation of Devonian trilobitesPaleobiology 38 292ndash307 httpsdoiorg101666100471

Arthur W (2004) The effect of development on the directionof evolution Toward a twenty‐first century consensusEvolution amp Development 6(4) 282ndash288 httpsdoiorg101111j1525‐142X200404033x

Badyaev A V (2005) Stress‐induced variation in evolution Frombehavioural plasticity to genetic assimilation Proceedings of theRoyal Society B Biological Sciences 272 877ndash886 httpsdoiorg101098rspb20043045

Badyaev A V (2009) Evolutionary significance of phenotypicaccommodation in novel environments An empirical test of theBaldwin effect Philosophical Transactions of the Royal SocietyB Biological Sciences 364 1125ndash1141 httpsdoiorg101098rstb20080285

Baedke J (2018) Organism where art thou Old and new challengesfor organism‐centered biology Journal of the History of Biology52 293ndash324 httpsdoiorg101007s10739‐018‐9549‐4

Bookstein F L (1997) Morphometric tools for landmark dataGeometry and biology Cambridge Cambridge University Press

Brakefield P M (2011) Evo‐devo and accounting for Darwinsendless forms Philosophical Transactions of the Royal Society BBiological Sciences 366(1574) 2069ndash2075 httpsdoiorg101098rstb20110007

Bridle J R amp Vines T H (2007) Limits to evolution at rangemargins When and why does adaptation fail Trends in Ecologyamp Evolution 22(3) 140ndash147 httpsdoiorg101016jtree200611002

Bush A M Powell M G Arnold W S Bert T M amp Daley G M(2002) Time‐averaging evolution and morphologic variationPaleobiology 28 9ndash25 httpsdoiorg1016660094‐8373(2002)0283C0009TAEAMV3E20CO2

Carroll S B (2005) Evolution at two levels On genes and form PLOSBiology 3 e245 httpsdoiorg101371journalpbio0030245

Carroll S B (2008) Evo‐devo and an expanding evolutionarysynthesis A genetic theory of morphological evolution Cell134 25ndash36 httpsdoiorg101016jcell200806030

Charlesworth B Lande R amp Slatkin M (1982) A neo‐Darwiniancommentary on macroevolution Evolution 36 474ndash498 httpsdoiorg1023072408095

JACKSON | 11 of 15

Cisne J L Chandlee G O Rabe B D amp Cohen J A (1982)Clinal variation episodic evolution and possible parapatricspeciation The trilobite flexicalymene senaria along anOrdovician depth gradient Lethaia 15 325ndash341 httpsdoiorg101111j1502‐39311982tb01697x

Dai T Zhang X L Peng S C amp Yao X Y (2017) Intraspecificvariation of trunk segmentation in the oryctocephalid trilobiteduyunaspis duyunensis from the Cambrian (Stage 4 Series 2) ofSouth China Lethaia 50 527ndash539 httpsdoiorg101111let12208

Davidson E H (2006) The regulatory genome San DiegoAcademic Press

Dawkins R (1976) The selfish gene Oxford Oxford University PressDobzhansky T (1937) Genetics and the origin of species New York

Columbia University PressEhrenreich I M amp Pfennig D W (2016) Genetic assimilation A

review of its potential proximate causes and evolutionaryconsequences Annals of Botany 117 769ndash779 httpsdoiorg101093aobmcv130

Eldredge N amp Gould S J (1972) Punctuated equilibria Analternative to phyletic gradualism In T J M Schopf (Ed)Models in paleobiology (pp 82ndash115) San Francisco US Free-man Cooper amp Company

Erwin D H (2017) Developmental push or environmental pullThe causes of macroevolutionary dynamics History andPhilosophy of the Life Sciences 39 36 httpsdoiorg101007s40656‐017‐0163‐0

Erwin D H amp Davidson E H (2009) The evolution ofhierarchical gene regulatory networks Nature Reviews Genetics10 141ndash148 httpsdoiorg101038nrg2499

Esteve J (2012) Intraspecific variability in paradoxidid trilobitesfrom the Purujosa trilobite assemblage (middle Cambriannortheast Spain) Acta Palaeontologica Polonica 59 215ndash241httpsdoiorg104202app20120006

Flatt T (2005) The evolutionary genetics of canalization TheQuarterly Review of Biology 80 287ndash316 httpsdoiorg101086432265

Futuyma D J (2015) Can modern evolutionary theory explainmacroevolution In E E Serreli amp N Gontier (Eds) Macro-evolution Interdisciplinary evolution research (2 pp 29ndash85)Cham Springer

Gerhart J amp Kirschner M (2007) The theory of facilitated variationProceedings of the National Academy of Sciences 104(suppl 1)8582ndash8589 httpsdoiorg101073pnas0701035104

Gibson G amp Dworkin I (2004) Uncovering cryptic geneticvariation Nature Reviews Genetics 5 681ndash690 httpsdoiorg101038nrg1426

Gibson G amp Wagner G (2000) Canalization in evolutionarygenetics A stabilizing theory BioEssays 22 372ndash380 httpsdoiorg101002(SICI)1521‐1878(200004)224lt372AID‐BIES7gt30CO2‐J

Gibert J M (2017) The flexible stem hypothesis Evidence fromgenetic data Development Genes and Evolution 227 297ndash307httpsdoiorg101007s00427‐017‐0589‐0

Gingerich P D (1984) Punctuated equilibria‐where is theevidence Systematic Zoology 33 335ndash338 httpsdoiorg1023072413079

Goswami A Binder W J Meachen J amp OrsquoKeefe F R (2015)The fossil record of phenotypic integration and modularity A

deep‐time perspective on developmental and evolutionarydynamics Proceedings of the National Academy of Sciences112 4891ndash4896 httpsdoiorg101073pnas1403667112

Gould S J (1980) Is a new and general theory of evolutionemerging Paleobiology 6 119ndash130 httpsdoiorg101017S0094837300012549

Gould S (1982) Darwinism and the expansion of evolutionary theoryScience 216 380ndash387 httpsdoiorg101126science7041256

Gould S J (1986) The hardening of the Modern Synthesis In MGrene (Ed) Dimensions of Darwinism Themes and Counter-themes in twentieth century evolutionary theory (pp 71ndash93) NewYork US Cambridge University Press

Gould S J (2002) The structure of evolutionary theory CambridgeHarvard University Press

Gould S J amp Eldredge N (1977) Punctuated equilibria Thetempo and mode of evolution reconsidered Paleobiology 3115ndash151 httpsdoiorg101017S0094837300005224

Gould S amp Eldredge N (1993) Punctuated equilibrium comes ofage Nature 366 223ndash227 httpsdoiorg101038366223a0

Hoffmann A A amp Hercus M J (2000) Environmental stress asan evolutionary force BioScience 50 217ndash226 httpsdoiorg1016410006‐3568(2000)050[0217ESAAEF]23CO2

Hopkins M J amp Webster M (2009) Ontogeny and geographicvariation of a new species of the corynexochine trilobiteZacanthopsis (Dyeran Cambrian) Journal of Paleontology 83524ndash547 httpsdoiorg10166608‐102R1

Hopkins M J (2014) The environmental structure of trilobitemorphological disparity Paleobiology 40 352ndash373 httpsdoiorg10166613049

Hunt G (2004) Phenotypic variance inflation in fossil samples Anempirical assessment Paleobiology 30 487ndash506 httpsdoiorg1016660094‐8373(2004)0303C0487PVIIFS3E20CO2

Hunt G (2004) Phenotypic variation in fossil samples Modelingthe consequences of time‐averaging Paleobiology 30 426ndash443httpsdoiorg1016660094‐8373(2004)0303C0426PVIFSM3E20CO2

Hunt G (2007) The relative importance of directional changerandom walks and stasis in the evolution of fossil lineagesProceedings of the National Academy of Sciences 10418404ndash18408 httpsdoiorg101073pnas0704088104

Hunt G (2010) Evolution in fossil lineages Paleontology and theorigin of species The American Naturalist 176(suppl 1)S61ndashS76 httpsdoiorg101086657057

Hunt G amp Rabosky D L (2014) Phenotypic evolution in fossilspecies Pattern and process Annual Review of Earth andPlanetary Sciences 42 421ndash441 httpsdoiorg101073pnas0704088104

Hunt G Bell M A amp Travis M P (2008) Evolution toward anew adaptive optimum Phenotypic evolution in a fossilstickleback lineage Evolution 62 700ndash710 httpsdoiorg101111j1558‐5646200700310x

Imasheva A G Loeschcke V Zhivotovsky L A amp Lazebny OE (1997) Effects of extreme temperatures on phenotypicvariation and developmental stability in Drosophila melano-gaster and Drosophila buzzatii Biological Journal of theLinnean Society 61 117ndash126 httpsdoiorg101111j1095‐83121997tb01780x

Iwasaki W M Tsuda M E amp Kawata M (2013) Genetic andenvironmental factors affecting cryptic variations in gene

12 of 15 | JACKSON

regulatory networks BMC Evolutionary Biology 13 91 httpsdoiorg1011861471‐2148‐13‐91

Jablonski D (2017) Approaches to macroevolution 1 Generalconcepts and origin of variation Evolutionary Biology 44427ndash450 httpsdoiorg101007s11692‐017‐9420‐0

Jablonski D amp Shubin N H (2015) The future of the fossilrecord Paleontology in the 21st century Proceedings of theNational Academy of Sciences 112 4852ndash4858 httpsdoiorg101073pnas1505146112

Jackson I S C amp Budd G E (2017) Intraspecific morphologicalvariation of Agnostus pisiformis a Cambrian Series 3 trilobite‐like arthropod Lethaia 50 467ndash485 httpsdoiorg101111let12201

Kidwell S M amp Flessa K W (1995) The quality of the fossilrecord Populations species and communities Annual Reviewof Ecology and Systematics 26 269ndash299 httpsdoiorg101146annurevearth241433

Kirkpatrick M (1982) Quantum evolution and punctuatedequilibria in continuous genetic characters The AmericanNaturalist 119 833ndash848 httpsdoiorg101086283958

Klingenberg C P (2019) Phenotypic plasticity developmentalinstability and robustness The concepts and how they areconnected Frontiers in Ecology and Evolution 7 7 httpsdoiorg103389fevo201900056

Kuhl F P amp Giardina C R (1982) Elliptic Fourier features of aclosed contour Computer Graphics and Image Processing 18236ndash258 httpsdoiorg1010160146‐664X(82)90034‐X

Kuumlttner E Parsons K J Easton A A Skuacutelason S DanzmannR G amp Ferguson M M (2014) Hidden genetic variationevolves with ecological specialization The genetic basis ofphenotypic plasticity in Arctic charr ecomorphs Evolution ampDevelopment 16 247ndash257 httpsdoiorg101111ede12087

Laland K N Uller T Feldman M W Sterelny K Muumlller G BMoczek A hellip Odling‐Smee J (2015) The extended evolu-tionary synthesis Its structure assumptions and predictionsProceedings of the Royal Society B Biological Sciences 28220151019 httpsdoiorg101098rspb20151019

Lande R (2009) Adaptation to an extraordinary environment byevolution of phenotypic plasticity and genetic assimilationJournal of Evolutionary Biology 22 1435ndash1446 httpsdoiorg101111j1420‐9101200901754x

Levis N A amp Pfennig D W (2016) Evaluating plasticity‐firstevolution in nature Key criteria and empirical approachesTrends in Ecology amp Evolution 31(7) 563ndash574 httpsdoiorg101016jtree201603012

Levis N A amp Pfennig D W (2019) Plasticity‐led evolutionEvaluating the key prediction of frequency‐dependent adapta-tion Proceedings of the Royal Society B Biological Sciences 28620182754 httpsdoiorg101098rspb20182754

Lieberman B S (2012) Adaptive radiations in the context ofmacroevolutionary theory A paleontological perspective Evo-lutionary Biology 39 181ndash191 httpsdoiorg101007s11692‐012‐9165‐8

Lieberman B S amp Eldredge N (2014) What is punctuatedequilibrium What is macroevolution A response to Pennellet al Trends in Ecology amp Evolution 29 185ndash186 httpsdoiorg101016jtree201402005

Losos J B Jackman T R Larson A de Queiroz K ampRodrıguez‐Schettino L (1998) Contingency and determinism

in replicated adaptive radiations of island lizards Science 2792115ndash2118 httpsdoiorg101126science27953592115

MacLeod N (2002) Geometric morphometrics and geologicalshape‐classification systems Earth‐Science Reviews 59 27ndash47httpsdoiorg101016S0012‐8252(02)00068‐5

MacLeod N (2017) Morphometrics History development meth-ods and prospects Zoological Systematics 42 4ndash33 httpsdoiorg1011865zs201702

Masel J (2006) Cryptic genetic variation is enriched for potentialadaptations Genetics 172 1985ndash1991 httpsdoiorg101534genetics105051649

McGuigan K amp Sgrograve C M (2009) Evolutionary consequences ofcryptic genetic variation Trends in Ecology amp Evolution 24305ndash311 httpsdoiorg101016jtree200902001

Moczek A P (2008) On the origins of novelty in development andevolution BioEssays 30 432ndash447 httpsdoiorg101002bies20754

Moczek A P (2012) The nature of nurture and the future ofevodevo Toward a theory of developmental evolution Inte-grative and Comparative Biology 52(1) 108ndash119 httpsdoiorg101093icbics048

Moczek A P (2015) Re‐evaluating the environment in develop-mental evolution Frontiers in Ecology and Evolution 3 httpsdoiorg103389fevo201500007

Moczek A P (2015) Developmental plasticity and evolutionmdashquovadis Heredity 115 302ndash305 httpsdoiorg101038hdy201514

Moczek A P Sultan S Foster S Ledoacuten‐Rettig C Dworkin INijhout H F hellip Pfennig D W (2011) The role ofdevelopmental plasticity in evolutionary innovation Proceed-ings of the Royal Society B Biological Sciences 278 2705ndash2713httpsdoiorg101098rspb20110971

Muumlller G B (2007) Evondashdevo Extending the evolutionarysynthesis Nature Reviews Genetics 8(12) 943ndash949 httpsdoiorg101038nrg2219

Myers C E amp Saupe E E (2013) A macroevolutionary expansionof the modern synthesis and the importance of extrinsic abioticfactors Palaeontology 56 1179ndash1198 httpsdoiorg101111pala12053

Neubauer T A Harzhauser M amp Kroh A (2013) Phenotypicevolution in a fossil gastropod species lineage Evidence foradaptive radiation Palaeogeography Palaeoclimatology Pa-laeoecology 370 117ndash126 httpsdoiorg101016jpalaeo201211025

Neige P Dera G amp Dommergues J L (2013) Adaptive radiationin the fossil record A case study among Jurassic ammonoidsPalaeontology 56 1247ndash1261 httpsdoiorg101111pala12062

Nicholson D J (2014) The return of the organism as afundamental explanatory concept in biology Philosophy Com-pass 9 347ndash359 httpsdoiorg101111phc312128

Noble D (2015) Evolution beyond neo‐Darwinism A newconceptual framework Journal of Experimental Biology 2187ndash13 httpsdoiorg101242jeb106310

Noble D (2015) Conrad Waddington and the origin of epigeneticsJournal of Experimental Biology 218 816ndash818 httpsdoiorg101242jeb120071

Organ C L Cooper L N amp Hieronymus T L (2015)Macroevolutionary developmental biology Embryos fossils

JACKSON | 13 of 15

and phylogenies Developmental Dynamics 244 1184ndash1192httpsdoiorg101002dvdy24318

Paaby A B amp Rockman M V (2014) Cryptic genetic variationEvolutions hidden substrate Nature Reviews Genetics 15247ndash258 httpsdoiorg101038nrg3688

Parsons K J Concannon M Navon D Wang J Ea I GroveasK hellip Albertson R C (2016) Foraging environment determinesthe genetic architecture and evolutionary potential of trophicmorphology in cichlid fishes Molecular Ecology 25 6012ndash6023httpsdoiorg101111mec13801

Pennell M W Harmon L J amp Uyeda J C (2014) Is there roomfor punctuated equilibrium in macroevolution Trends inEcology amp Evolution 29 23ndash32 httpsdoiorg101016jtree201307004

Pertoldi C Bundgaard J Loeschcke V amp Barker J S F (2014)The phenotypic variance gradientndasha novel concept Ecology andEvolution 4 nandashna httpsdoiorg101002ece31298

Pertoldi C amp Kristensen T (2015) A new fluctuating asymmetryindex or the solution for the scaling effect Symmetry 7327ndash335 httpsdoiorg103390sym7020327

Pigliucci M (2010) Genotypendashphenotype mapping and the end ofthe lsquogenes as blueprintrsquo metaphor Philosophical Transactions ofthe Royal Society B Biological Sciences 365 557ndash566 httpsdoiorg101098rstb20090241

Pigliucci M amp Murren C J (2003) Perspective Geneticassimilation and a possible evolutionary paradox Can macro-evolution sometimes be so fast as to pass us by Evolution 571455ndash1464 httpsdoiorg101111j0014‐38202003tb00354x

Pigliucci M (2006) Phenotypic plasticity and evolution by geneticassimilation Journal of Experimental Biology 209 2362ndash2367httpsdoiorg101242jeb02070

Pimiento C Tang K L Zamora S Klug C amp Saacutenchez‐VillagraM R (2018) Assessing canalisation of intraspecific variation ona macroevolutionary scale The case of crinoid arms through thePhanerozoic PeerJ 6 e4899 httpsdoiorg107717peerj4899

Rieseberg L H Archer M A amp Wayne R K (1999)Transgressive segregation adaptation and speciation Heredity83 363ndash372 httpsdoiorg101038sjhdy6886170

Rutherford S L (2000) From genotype to phenotype Bufferingmechanisms and the storage of genetic information BioEssays22 1095ndash1105 httpsdoiorg1010021521‐1878(200012)2212lt1095AID‐BIES7gt30CO2‐A

Scarponi D amp Kowalewski M (2004) Stratigraphic paleoecologyBathymetric signatures and sequence overprint of molluskassociations from upper Quaternary sequences of the Po PlainItaly Geology 32 989ndash992 httpsdoiorg101130G208081

Schlichting C D (1989) Phenotypic integration and environmen-tal change BioScience 39 460ndash464 httpsdoiorg1023071311138

Schlichting C D (2008) Hidden reaction norms crypticgenetic variation and evolvability Annals of the New YorkAcademy of Sciences 1133 187ndash203 httpsdoiorg101196annals1438010

Schlichting C D amp Smith H (2002) Phenotypic plasticityLinking molecular mechanisms with evolutionary outcomesEvolutionary Ecology 16 189ndash211 httpsdoiorg101023A1019624425971

Schlichting C D amp Wund M A (2014) Phenotypic plasticity andepigenetic marking An assessment of evidence for genetic

accommodation Evolution 68 656ndash672 httpsdoiorg101111evo12348

Schmalhausen I I (1949) Factors of Evolution Chicago TheUniversity of Chicago Press

Schneider R F amp Meyer A (2017) How plasticity geneticassimilation and cryptic genetic variation may contribute toadaptive radiations Molecular Ecology 26 330ndash350 httpsdoiorg101111mec13880

Schoch R R (2014) Life cycles plasticity and palaeoecologyin temnospondyl amphibians Palaeontology 57(3) 517ndash529httpsdoiorg101111pala12100

Schwab D B Casasa S amp Moczek A P (2019) On thereciprocally causal and constructive nature of developmentalplasticity and robustness Frontiers in Genetics 9 httpsdoiorg103389fgene201800735

Schwander T amp Leimar O (2011) Genes as leaders and followersin evolution Trends in Ecology amp Evolution 26 143ndash151httpsdoiorg101016jtree201012010

Sheldon P R (1997) The Plus ccedila change model Explainingstasis and evolution in response to abiotic stress overgeological timescales In R Bijlsma amp V Loeschcke (Eds)Environmental Stress Adaptation and Evolution (pp 307ndash319)Basel Birkhaumluser

Simpson G G (1944) Tempo and mode in evolution New YorkColumbia University Press

Simpson G G (1953) The major features of evolution New YorkColumbia University Press

Stroud J T amp Losos J B (2016) Ecological opportunity andadaptive radiation Annual Review of Ecology Evolution andSystematics 47 507ndash532 httpsdoiorg101146annurev‐ecolsys‐121415‐032254

Tribovillard N Algeo T J Lyons T amp Riboulleau A (2006)Trace metals as paleoredox and paleoproductivity proxies Anupdate Chemical Geology 232 12ndash32 httpsdoiorg101016jchemgeo200602012

Uller T Moczek A P Watson R A Brakefield P M amp LalandK N (2018) Developmental bias and evolution A regulatorynetwork perspective Genetics 209 949ndash966 httpsdoiorg101534genetics118300995

van Bergen E Osbaldeston D Kodandaramaiah U BrattstroumlmO Aduse‐Poku K amp Brakefield P M (2017) Conservedpatterns of integrated developmental plasticity in a group ofpolyphenic tropical butterflies BMC Evolutionary Biology 17httpsdoiorg101186s12862‐017‐0907‐1

Voje K L (2016) Tempo does not correlate with mode in thefossil record Evolution 70 2678ndash2689 httpsdoiorg101111evo13090

Waddington C H (1942) Canalization of development and theinheritance of acquired characters Nature 150 563ndash565httpsdoiorg101038150563a0

Waddington C H (1961) Advances in genetics Advances inGenetics 10 257ndash293 httpsdoiorg101016S0065‐2660(08)60119‐4

Wagner A (2011) The origins of evolutionary innovations a theoryof transformative change in living systems Oxford OxfordUniversity Press

Wagner G P Booth G amp Bagheri‐Chaichian H (1997) Apopulation genetic theory of canalization Evolution 51329ndash347 httpsdoiorg101111j1558‐56461997tb02420x

14 of 15 | JACKSON

Wagner G P amp Zhang J (2011) The pleiotropic structure of thegenotypendashphenotype map The evolvability of complex organ-isms Nature Reviews Genetics 12 204ndash213 httpsdoiorg101038nrg2949

Watson R A amp Szathmaacutery E (2016) How can evolution learnTrends in Ecology amp Evolution 31 147ndash157 httpsdoiorg101016jtree201511009

Watson R A Wagner G P Pavlicev M Weinreich D M ampMills R (2014) The evolution of phenotypic correlations andldquodevelopmental memoryrdquo Evolution 68 1124ndash1138 httpsdoiorg101111evo12337

Webber A J amp Hunda B R (2007) Quantitatively comparingmorphological trends to environment in the fossil record(Cincinnatian Series Upper Ordovician) Evolution 611455ndash1465 httpsdoiorg101111j1558‐5646200700123x

Webster M (2015) Ontogeny and intraspecific variation of theearly Cambrian trilobite Olenellus gilberti with implications forolenelline phylogeny and macroevolutionary trends in pheno-typic canalization Journal of Systematic Palaeontology 13 1ndash74httpsdoiorg101080147720192013852903

Webster M (2019) Morphological homeostasis in the fossil recordSeminars in Cell amp Developmental Biology 88 91ndash104 httpsdoiorg101016jsemcdb201805016

West‐Eberhard M J (2003) Developmental plasticity and evolutionOxford Oxford University Press

West‐Eberhard M J (2005) Developmental plasticity and theorigin of species differences Proceedings of the NationalAcademy of Sciences 102(suppl 1) 6543ndash6549 httpsdoiorg101073pnas0501844102

West‐Eberhard M J (2005) Phenotypic accommodation Adaptiveinnovation due to developmental plasticity Journal of Experi-mental Zoology Part B Molecular and Developmental Evolution304 610ndash618 httpsdoiorg101002jezb21071

Williams E E (1972) The origin of faunas Evolution of lizardcongeners in a complex island fauna A trial analysis In

T Dobzhansky M K Hecht amp W C Steere (Eds)Evolutionary Biology New York NY Springer

Willmore K E Young N M amp Richtsmeier J T (2007)Phenotypic variability Its components measurement andunderlying developmental processes Evolutionary Biology 3499ndash120 httpsdoiorg101007s11692‐007‐9008‐1

Wright S (1932) The roles of mutation inbreeding crossbreedingand selection in evolution Proceedings of the VI InternationalCongress of Genetics 1 356ndash366

Wund M A (2012) Assessing the impacts of phenotypic plasticityon evolution Integrative and comparative biology 52 5ndash15httpsdoiorg101093icbics050

Wund M A Baker J A Clancy B Golub J L amp Foster SA (2008) A test of the ldquoflexible stemrdquo model of evolutionAncestral plasticity genetic accommodation and morpho-logical divergence in the threespine stickleback radiationThe American Naturalist 172 449ndash462 httpsdoiorg101086590966

Wund M A Valena S Wood S amp Baker J A (2012) Ancestralplasticity and allometry in threespine stickleback revealphenotypes associated with derived freshwater ecotypesBiological Journal of the Linnean Society 105 573ndash583httpsdoiorg105061dryadhb824gd4

Zheng J Payne J L amp Wagner A (2019) Cryptic geneticvariation accelerates evolution by opening access to diverseadaptive peaks Science 365 347ndash353 httpsdoiorg101126scienceaax1837

How to cite this article Jackson ISCDevelopmental bias in the fossil record Evolutionamp Development 2019e12312httpsdoiorg101111ede12312

JACKSON | 15 of 15