Is a New and General Theory of Evolution Emerging

7/29/2019 Is a New and General Theory of Evolution Emerging

1/12

Paleobiology, 61), 1980, pp 119130

s a new and general theory of evolution emerging?

Stephen Jay Gould

Abstract.-T modern syntesis, as an exclusive proposition, as broken down on bot of its fundamental claims: extrapolationism (gradual allelic substitution as a model for all evolutionary cange) andnearly exclusive reliance on selection leading to adaptation Evolution is a ierarcical process witcomplementary, but different, modes of cange at its tree major levels variation witin populations,

speciation, and patterns of macroevolution Speciation is not always an extension of gradual, adaptiveallelic substitution to greater effect, but may represent, as Goldscmidt argued, a different style of geneticcangrapid reorganization of te genome, peraps nonadaptive. Macroevolutionary trends do notarise from te gradual, adaptive transformation of populations, but usually from a igerorder selectionoperating upon groups of species, wile te individual species temselves generally do not cange following teir geologically instantaneous origin. I refer to tese two discontinuities in te evolutionaryierarcy as te Goldscmidt break (between cange in populations and speciation) and te Wrigt break(between speciation and trends as differential success among species).

A new and general evolutionary teory will embody tis notion of ierarcy and stress a variety oftemes eiter ignored or explicitly rejected by te modern syntesis punctuational cange at all levels,important nonadaptive cange at all levels, control of evolution not only by selection, but equally byconstraints of istory, development and arcitecturtus restoring to evolutionary teory a concept oforganism.

Stephen Jay Gould. Museum of Comparative Zoology, Haard University, Cambridge, Massachusetts

238

Accepted October 15, 1979

I. The Modern Synthesis

In one of the last skeptical books written before the Darwinian tide of the modern synthesisasserted its hegemony, Robson and Richardscharacterized the expanding orthodoxy thatthey deplored

The theory of Natural Selection . postulates that the evolutionary process is unitary, and that not only are groups formed by

the multiplication of single variants havingsurvival value, but also that such divergencesare amplied to produce adaptations (bothspecializations and organization) It has beencustomary to admit that certain ancillary processes are operative (isolation, correlation),but the importance of these, as active principles, is subordinate to selection (1936, pp.370371)

Darwinism, as a set of ideas, is sufciently

broad and variously dened to include a multitude of truths and sins. Darwin himself dis

1980 Te Paleontological Society. All rigts reserved.

avowed many interpretations made in his name(1880, for example) The version known as the"modern synthesis or "Neo-Darwinism (different from what the late 19th century calledNeo-Darwinismsee Romanes, 1900) is, Ithink, fairly characterized in its essentials byRobson and Richards Its foundation rests upontwo major premises (1) Point mutations (micromutations) are the ultimate source of variabili

ty Evolutionary change is a process of gradualallelic substitution within a population Eventsat broader scale, from the origin of new speciesto long-ranging evolutionary trends, representthe same process, extended in time and effectlarge numbers of allelic substitutions incorporated sequentially over long periods of time Inshort, gradualism, continuity and evolutionarychange by the transformation of populations(2) Genetic variation is raw material only Natural selection directs evolutionary change

Rates and directions of change are controlled byselection with little constraint exerted by raw

00948373/80/06010011$1. 00


2/12

10 STEPHEN JAY GOULD

material (slow rates are due to weak selection,not insufcient variation). All genetic change isadaptive (though some phenotypic effects, dueto pleiotropy, etc., may not be) In short, selection leading to adaptation.

All these statements, as Robson and Richardsalso note, are subject to recognized exceptionsand this imposes a great frustration upon anyone who would characterize the modern synthesis in order to criticize it. All the synthesistsrecognized exceptions and "ancillary processes,but they attempted both to prescribe a low relative frequency for them and to limit their application to domains of little evolutionary importance. Thus, genetic drift certainly occursbut only in populations so small and so near thebrink that their rapid extinction will almost cer

tainly ensue And phenotypes include manynon-adaptive features by allometry and pleotropy, but all are epiphenomena of primarilyadaptive genetic changes and none can haveany marked effect upon the organism (for, ifinadaptive, they will lead to negative selectionand elimination and, if adaptive, will enter themodel in their own right) Thus, a synthesistcould always deny a charge of rigidity by invoking these ofcial exceptions, even thougtheir circumscription, both in frequency and effect, actually guaranteed the hegemony of thetwo cardinal principles This frustrting situation had been noted by critics of an earlier Darwinian orthodoxy, by Romanes writing of Wallace, for example (1900, p 21):

[For Wallace,] the law of utility is, to all ntents and purposes, universal, with the resultthat natural selection is virtually the onlycause of organic evolution. I say 'to all intentsand purposes, or 'virtually, because Mr.Wallace does not expressly maintain the abstract impossibility of laws and causes otherthan those of utility and natural selection; indeed, at the end of his treatise, he uotes withapproval Darwins judgement, that 'naturalselection has been the most mportant, but notthe exclusive means of modication Nevertheless, as he nowhere recognizes any othelaw or cause of adaptive evolution, he practically concludes that, on inductive or empir

ical grounds, there is no such other law orcause to be entertained.

Lest anyone think that Robson and Richards,as doubters, had characterized the oppositionunfairly, or that their two principles representtoo simplistic or unsubtle a view of the synthetictheory, I cite the characterization of one of thearchitects of the theory himself (Mayr 1963, p586the rst statement of his chapter on

species and transspecic evolution)

The proponents of the synthetic theorymaintain that all evolution is due to the accumulation of small genetic changes, guidedby natural selection, and that trans specicevolution is nothing but an extrapolation andmagnication of the events that take placewithin populations and species

The early classics of the modern synthesisparticularly Dobzhanskys rst edition (1937)and Simpsons rst book (1944)were quite expansive, generous and pluralistic But the synthesis hrdened throughout the late 40s and50s, and later editions of the same classics(Dobzhansky 1951; Simpson 1953) are more rigid in their insistence upon micromutation, gradual transformation and adaptaton guided byselection (see Gould 1980 for an analysis ofchanges between Simpsons two books) When

Watson and Crick then determined the structure of DNA, and when the triplet code wascracked a few years later, everything seemed tofall even further into place Chromosomes arelong strings of triplets coding, in sequence, forthe proteins that build organisms Most pointmutations are simple base substitutions Aphysics and chemistry had been added, and itsquared well with the prevailing orthodox

I well remember how the synthetic theorybeguiled me with its unifying power when I wasa graduate student in the mid-1960s Since thenI have been watching it slowly unravel as a universal description of evolution The molecularassault came rst, followed quickly by renewedatention to unorthodox theories of speciationand by challenges at the level of macroevolutionitself I have been reluctant to admit itsincebeguiling is often foreverbut if Mayrs characterization of the ynthetic theory is accurate,then that theory, as a general proposition, is

effectively dead, despite its persstence as textbook orthodoxy


3/12

EMERGING EVOLUTIONARY THEORY 121

II. Reduction and Hierarchy

The modern synthetic theory embodies astrong faith in reductionism It advocates asmoot extrapolation across all levels andscalesfrom the base substitution to the originof higher taxa The most sophisticated of lead

ing introductory textbooks in biology still proclaims:

[Can] more extensive evolutionary change,macroevolution, be explained as an outcomeof these microevolutionary shifts. Did birdsreally arise from reptiles by an accumulationof gene substitutions of the kind illustrated bythe raspberry eye-color gene

The answer is that it is entirely plausible,and no one has come up with a better expla

nation . The fossil record suggests thatmacroevolution is indeed gradual, paced at arate that leads to the conclusion that it isbased upon hundreds or thousands of genesubstitutions no different in kind from theones examined in our case histories (Wilson etal 1973, pp 793794)

The general alternative to such reductionismis a concept of hierarchya world constructednot as a smooth and seamless continuum, per

mitting simple extrapolation from the lowestlevel to the highest, but as a series of ascendinglevels, each bound to the one below it in someways and independent in others. Discontinuitiesand seams characterize the transitions "emergent features not implicit in the operation ofprocesses at lower levels, may control events athigher levels. The basic processesmutation,selection, etcmay enter into explanations atall scales (and in that sense we may still hopefor a general theory of evolution), but they workin different ways on the characteristic materialof divers levels (see Bateson 1978 and Koestler1978, for all its other inadequacies, for gooddiscussions of hierarchy and its anti-reductionistic implications; Eldredge and Cracraft 1980)

The molecular level, which once seemedthrough its central dogma and triplet code toprovide an excellent "atomic basis for smoothextrapolation, now demands hierarchical interpretation itself he triplet code is only machine

language (I tank E. Yates for this appropriatemetaphor). The program resides at a higher

level of control and regulationand we knowvirtually nothing about it With its inserted sequences and jumping genes, the genome contains sets of scissors and pots of glue to snip andunite bits and pieces from various sources Thirty to seventy percent of the mammalian genomeconsists of repetitive sequences, some repeatedhundreds or thousands of times What are theyfor (if anything)? What role do they play in theregulation of development? Molecular biologistsare groping to understand this higher controlupon primary products of the triplet code Inthat understanding, we will probably obtain abasis for styles of evolutionary change radicallydifferent from the sequential allelic substitutions, each of minute effect, that the modernsynthesis so strongly advocated The uncover

ing of hierarchy on the molecular level willprobably exclude smooth continuity across other levels (We may nd, for example, that structural gene substitutions control most small-scale,adaptive variation within local populations,while disruption of regulation lies behind mostkey innovations in macroevolution )

The modern synthesis drew most of its directconclusions from studies of local populationsand their immediate adaptations It then extrapolated the postulated mechanism of theseadaptationsgradual, allelic substitutiontoencompass all larger-scale events The synthesisis now breaking down on both sides of this argument. Many evolutionists now doubt exclusive control by selection upon genetic changewithin local populations Moreover, even if local populations alter as the synthesis maintains,we now doubt that the same style of changecontrols events at the two major higher levelsspeciation and patterns of macroevolution

III A Note on Local Populations andNeutrality

At the level of populations, the synthesis hasbroken on the issue of amounts of genetic variation Selection, though it eliminates variationin both its classical modes (directional and, especially, stabilizing) can also act to preservevariation through such phenomena as overdominance, frequency dependence, and response to

small-scale uctuation of spatial and temporalenvironments. Nonetheless, the copiousness of


4/12


genetic variation, as revealed rst in the electrophoetic techniques that resolve only some ofit (Lewontin and ubby 1966; Lewontin 1974),cannot be encompassed by our models of selective control (of course, the models, rather thannature, may be wrong) This fact has forcedmany evolutionists, once stout synthesists themselves, to embrace the idea that alleles oftendrift to high frequency or xation, and thatmany common variants are therefore neutral or

just slightly deleterous. This admission lendssupport to a previous interpretation of the approximately even ticking of the molecular clock(Wilson 1977)that it reects the neutral statusof most changes in structural genes rather thana grand averaging of various types of selectionover time

None of this evidence, of course, negates therole of conventional selection and adaptation inmolding parts of the phenotype with obvousimportance for survival and reproduction. Still,it rather damps Mayrs enthusiastic claim for"all evolution guided by natural selectionThe question, as with so many issues in thecomplex sciences of natural history, becomesone of relative frequency Are the Darwiniansubstitutions merely a surface skin on a sea ofvariation invisible to selection, or are the neutral substitutions merely a thin bottom layerunderlying a Darwinian ocean above Or wherein between?

In short, the specter of stochasticity has intruded upon explanations of evolutionarychange. This represents a fundamental challenge to Darwinism, which holds, as its verybasis, that random factors enter only in the production of raw material, and that the deterministic process of selection produces change and

direction (see Nei 1975).

IV. The Level of Speciation andthe Goldschmidt Break

Ever since Darwin called his book The Originof Species, evolutionists have regarded the formation of reproductively isolated units by speciation as a fundamental process of large-scalechange. Yet speciation occurs at too high a levelto be observed directly in nature or producedby experiment in most cases Therefore, theories of speciation have been based on analogy,

extrapolation and inference Darwin himself focused on articial selection and geographic variation e regarded subspecies as incipentspecies and viewed their gradual, accumulatingdivergence as the primary mode of origin fornew taxa The modern synthesis continued thistradtion of extrapolation from local populationsand used the accepted model for adaptive geographic variationgradual allelic substitutondrected by natural selectionas a paradigm forthe origin of species Mayrs (1942, 1963) modelof allopatric speciation did challenge Darwinsimplied notion of sympatric continuity It emphasized the crucal role of isolation from geneow and did promote the importance of smallfounding populations and relatively rapid ratesof change Thus, the small peripheral isolate,

rather than the large local population in persistent contact with other conspecics, became theincipient species Nonetheless, despite ths welcome departure from the purest form of Darwinian gradualism, the allopatric theory heldrmly to the two major principles that permitsmooth extrapolation from the Biston betulariamodel of adaptive, allelic substitution: (i) Theaccumulating changes that lead to speciationare adaptive Reproductive isolation is a consequence of sufcient accumulation (ii) Although aided by founder effects and even (possibly) by drift, although dependent uponisolation from gene ow, although proceedingmore rapidly than local differentiation withinlarge populations, successful speciation is stilla cumulative and sequential process poweredby selection throgh large numbers of generations. It is, if you will, Darwinism a little faster

I have no doubt that many species originatein this way; but it now appers that many, per

haps most, do not. The new models stand atvariance with the synthetic proposition that speciation is an extension of microevolution withinlocal populations. Some of the new models callupon genetic variation of a different kind, andthey regard reproductive isolation as potentiallyprimary and non-adaptive rather than secondary and adaptive Insofar as these new modelsbe valid in theory and numerically important inapplication, speciation is not a simple "conversion to larger effect of processes occurring atthe lower level of adaptive modelling within 10-


5/12


cal populations It represents a discontinuity inour hierarchy of explanations, as the much maligned Richard Goldschmidt argued explicitly in1940

There are many ways to synthesize the swirling set of apparently disparate challenges that

have rocked the allopatric orthodoxy and established an alternative set of models for speciation The following reconstruction is neither historically sequential nor the only logical pathwayof linkage, but it does summarize the challengeson population structure, place of origin, genetic style, rate, and relation to adaptationin some reasonable order

Under the allopatric orthodoxy, speciesare viewed as integrated units which, if not actually panmictic, are at least sufciently ho

mogenized by gene ow to be treated as entitiesThis belief in effective homogenization withincentral populations underlies the allopatric theory with its emphasis on peripheral isolation asa precondition for speciation But many evolutionists now believe that gene ow is often tooweak to overcome selection and other intrinsicprocesses within local demes (Ehrlich and Raven 1969). Thus, the model of a large, homogenized central population preventing local dif

ferentiation and requiring allopatric "ight ofisolated demes for speciation may not be generally valid Perhaps most local demes have therequired independence for potential speciation

2. The primary terms of reference for theories of speciationallopatry and sympatrylose their meaning if we accept the rst statement Objections to sympatric speciation centered upon the homogenizing force of gene ow.But if demes may be independent in all geographic domains of a species, then sympatry los

es its meaning and allopatry its necessity Independent demes within the central range(sympatric by location) function, in their freedom from gene ow, like the peripheral isolatesof allopatric theory. In other words, the termsmake no sense outside a theory of populationstructure that contrasts central panmixia withmarginal isolation. They should be abandoned

3 In this context "sympatric speciation loses its status as an extremely improbable event

If demes are largely independent, new speciesmay originate anywhere within the geographic

range of an ancestral form Moreover, manyevolutionists now doubt that parapatric distributions (far more common then previouslythought) must represent cases of secondary contact White (1978, p 342) believes that many,if not most, are primary and that speciation can

also occur between populations continually incontact if gene ow can be overcome either bystrong selection or by the sheer rapidity of potential xation for major chromosomal variants(see White, 1978, p 17 on clinal speciation)

4 Most "sympatric models of speciation arebased upon rates and styles of genetic changeinconsistent with the reliance placed by themodern synthesis on slow, or at least sequentialchange

The most exciting entry among punctuational

models for speciation in ecological time is theemphasis, now coming from several quarters,on chromosomal alterations as isolating mechanisms (White 1978; Bush 1975; Carson 1975,1978; Wilson et al. 1975; Bush et al 1977)sometimes called the theory of chromosomalspeciation In certain population structures,particularly in very small and circumscribedgroups with high derees of inbreeding, majorchromosomal changes can rise to xation in less

than a handful of generations (mating of heterozygous F sibs to produce F homozygotesfor a start)

Allan Wilson, Guy Bush and their colleague(Wilson et al 1975; Bush et al 1977) nd astrong correlation between rates of karyotypicand anatomical change, but no relation betweeamounts of substitution in structural genes andany conventional assessment of phenotypicmodication, either in speed or extent Theysuggest that speciation may be more a matter

of gene regulation and rearrangement than ofchanges in structural genes that adapt local populations in minor ways to uctuating environments (the Biston betulara model)

Carson (1975, 1978) has also stressed the importance of small demes, chromosomal chageand extremely rapid speciation in his founderush theory with its emphasis on extreme bottlenecking during crashes of the ush-crashcycle (see Powell 1978 for experimental sup

port) Explicitly contrasting this view with extrapolationist models based on sequential u-


6/12


stitution of structural genes, he writes (1975, p.88)

Most theories of speciation are wedded togradualism, using the mode of origin of intraspecic adaptations as a model Iwould nevertheless like to propose that

speciational events may be set in motion andimportant genetic saltations towards speciesformation accomplished by a series of catastrophic, stochastic genetic events . . . initiated when an unusual forced reorganizationof the epistatic supergenes o the closed variability system occurs . . I propose that thiscycle of disorganization and reorganization beviewed as the essence of the speciation process

5. Another consequence of such essentialysaltational origin is even more disturbing to conventional views than the rapidity of the processitself, as Carson has forcefully stated. The control of evolution by selection leading to adaptation lies at the heart of the modern synthesisThus, reproductive isolation, the denition ofspeciation, is attained as a by-product of adaptationthat is, a population diverges by sequential adaptation and eventually becomessufciently different from its ancestor to foreclose interbreeding. (Selection for reproductiveisolation may also be direct when two imperfectly-separate forms come into contact.) But insaltational, chromosomal speciation, reproductive isolation comes rst and cannot be considered as an adaptation at all. It is a stochasticevent that establishes a species by the technicaldenition of reproductive isolation. To be sure,the later success of this species in competitionmay depend upon its subsequent acquisition of

adaptations; but the origin itself may be nonadaptive We can, i fact, reverse the conventional view and argue that speciation, by forming new entities stochastically, provides rawmaterial for selection

These challenges can be summarized in theclaim that a discontinuity in explanation existsbetween allelic substitutions in local populations (sequential, slow and adaptive) and theorigin of new species (often discontinuous andnon-adaptive) During the heyday of the modern synthesis, Richard Goldschmidt wa castigated for his defense of punctuational specia-

tion. I was told as a graduate student that thisgreat geneticist had gone astray because he hadbeen a lab man with no feel for nature, a personwho hadnt studied the adaptation of local populations and couldnt appreciate its potentialpower, by extrapolation, to form new speciesBut I discovered, in writing Ontogeny and Phy-logeny, that Goldschmidt had spent a good partof his career studying geographic variation,largely in the cloration of lepidopteran larvae(where he developed the concept of rate genesto explain minor changes in pattern) I thenturned to his major book (1940) and found thathis defense of saltational speciation is not basedon ignorance of geographic variation, but on anexplicit study of it; half the book is devoted tothis subject. Goldschmidt concludes that geo

graphic variation is ubiquitous, adaptive, andessential for the persistence of establishedspecies. But it is simply not the stuff of speciation it is a different process Speciation, Goldschmidt argues, occurs at different rates anduses different kinds of genetic variation. We donot now accept all his arguments about the nature of variation, but his explicit anti-extrapolationist statement is the epitome and foundation of emerging views on speciation discussedin this section There is a discontinuity in causeand explanation between adaptation in localpopulations and speciation; they represent twodistinct, though interacting, levels of evolution.We might refer to this discontinuity as the Gold-schmidt break, for he wrote

The characters of subspecies are of a gradient type, the species limit is characterizedby a gap, an unbridged difference in manycharacters This gap cannot be bridged by

theoretically continuing the subspecic gradient or cline beyond its actually existing limits. The subspecies do not merge into thespecies either actually or ideally . . . Microevolution by accumulation f micromutationswe may also say neo-Darwinian evolutionis a process which leads todiversication strictly within the species, usually, if not exclusively, for the sake of adaptation of the species to specic conditionswithin the area which it is able to occupy . . Subspecies are actually, therefore, neither incipient species nor models for the origin


7/12


8/12

16 STEPHEN JA GOULD

the result of their differential origin and survival Speciation interposes itself as an irreduciblelevel between change in local populations andtrends in geological time. Macroevolution is, asStanley argues (1975, p. 648), decoupled frommicroevolution.

Sewall Wright recognized the hierarchical implications of viewing species as irreducible inputs to macroevolution when he argued (1967,p. 11) that the relationship between change inlocal populations and evolutionary trends canonly be analogical. Just as mutation is randomwith respect to the direction of change withina population, so too might speciation be randomwith respect to the direction of a macroevolutionary trend A higher form of selection, acting directly upon species through differential

rates of extinction, may then be the analog ofnatural selection working within populationsthrough differential mortality of individuals.

Evolutionary trends therefore represent athird level superposed upon speciation andchange within demes. Intrademic events cannotencompass speciation because rates, geneticstyles, and relation to adaptation differ for thetwo processes. ikewise, since trends "usespecies as their raw material, they represent aprocess at a higher level than speciation itself.They reect a sorting out of speciation events.With apologies for the pun, the hierarchicalrupture between speciation and macroevolutionary trends might be called the Wrightbreak.*

As a nal point about the extrapolation ofmethods for the study of events within populations, the cladogenetic basis of macroevolution virtually precludes any direct applicationof the primary apparatus for microevolutionary

theory classical population genetics. (I believe

* I had te onor-not a word I use frequently, but inescapable in tis case-of spending a long evening wit DrWrigt last year. I discovered tat is quip about macro

evolution, just paraprased, was no trowaway statementbut an embodiment of is deep commitment to a iearcical view of evolutionary causation. (Te failure of manyevolutionists to tink ierarcically is responsible for temost frequent misinterpretation of Wrigt's views He never believed tat genetic drift-te Sewall Wrigt effect asit once was called-is an important agent of evolutionarychange He regards it as input to te directional process of

interdemic selection for evolution witin species Drift canpus a deme off an adaptive peak; selection can ten drawit to anoter peak

that essentially all macroevolution is cladogenesis and its concatenated effects. What we call"anagenesis, and often attempt to delineate asa separate phyletic process eading to "progress, is just accumulated cladogenesis lteredthrough the directing force of species selection(Stanley 1975)Wrights higher level analog ofnatural selection Carson (1978, p 95) makesthe point forcefully, again recognizing SewallWright as its long and chief defender

Investigation of cladistic events as opposedto phyletic (anagenetic) ones requires a different perspective from that normally assumed in classical population genetics Thestatistical and mathematical comfort of theardy-Weinberg equilibrium in large popu

lations has to be abandoned in favor of thevague realization that nearly everywhere innature we are faced with data suggesting thepartial or indeed complete sundering of genepools. If we are to deal realistically withcladogenesis we must seek to delineate eachgenetic and environmental factor which maypromote isolation. The most important devices are clearly those which operate at thevery lowest population level sib from sib,family from family, deme from deme. Formalpopulation genetics just cannot deal withsuch things, as Wright pointed out long ago.

Eldredge (1979) has traced many conceptualerrors and prejudicial blockages to our tendencyfor conceiving of evolution as the transformation ofcharacters within phyletic lineages, rather than as the origin of new taxa by cladogenesis(the transformational versus the taxic view inhis terms). I believe that, in ways deeper than

we realize, our preference for transformationalthinking represents a cultural tie to the controlling Western themes of progress and ranking byintrinsic meritan attitude that can be tracedin evolutionary thought to amarcks distinction between the march up lifes ladder promoted by the pouvoir de la vie and the tangential departures imposed by l'inuence descirconstances, with the rst process essentialand the second deective. Nonetheless, macroevolution is fundamentally about the origin oftaxa by splitting

. The saltational initiation of major transi-


9/12


tions The absence of fossil evidence for intermediary stages between major transitions in organic design, indeed our inability, even in ourimagination, to construct functional intermediates in many cases, has been a persistent andnagging problem for gradualistic accounts ofevolution. St. George Mivart (1871), Darwinsmost cogent critic, referred to it as the dilemmaof "the incipient stages of useful structuresofwhat possible benet to a reptile is two percentof a wing? The dilemma has two potential solutions. The rst, preferred by Darwinians because it preserves both gradualism and adaptation, is the principle of preadaptation theintermediary stages functioned in another waybut were, by good fortune in retrospect, preadapted to a new role they could play only after

greater elaboration Thus, if feathers rst functioned "for insulation and later "for the trapping of insect prey (Ostrom 199), a proto-wingmight be built without any reference o ight.

I do not doubt the supreme importance ofpreadaptation, but the other alternative, treatedwith caution, reluctance, disdain or even fearby the modern synthesis, now deserves a rehearing in the light of renewed interest in development perhaps, in many cases, the intermediates never existed. I do not refer to thesaltational origin of entire new designs, complete in all their complex and integrated featuresa fantasy that would be truly anti-Darwinian in denying any creativity to selectionand relegating it to the role of eliminating oldmodels. Instead, I envisage a potential saltational origin for the essential features of key adaptations. Why may we not imagne tha gillarch bones of an ancestral agnathan moved forward in one step to surround the mouth and

form proto-jaws? Such a change would scarcelyestablish the Bauplan of the gnathostomes. Somuch more must be altered i the reconstruction of agnathan designthe building of a true

shoulder girdl with Qony, paired appendages,to say the least. But the discontinuous origin ofa proto-jaw might set up new regimes of development and selection that would quickly leadto other, coordinated modications. Yet Darwin, conating gradualism with natural selection as he di so often, wrongly proclaimed thatany such discontinuity, even for organs (muchless taxa) ould estroy his theory

If it could be demonstrated that any complex organ existed, which could not possiblyhave been formed by numerous, successive,slight modications, my theory would absolutely break down (1859, p 189).

During the past 30 years, such proposals havegenerally been treated as a fantasy signifyingsurrenderan invocation of hopeful monstersrather than a square facing of a difcult issueBut our renewed interest in development, theonly discipline of biology that might unify molecular and evolutionary approaches into a coherent science, suggests that such ideas are neither fantastic, utterly contrary to geneticprinciples, nor untestable

Goldschmidt conated two proposals as caus

es for hopeful monsters"systemic mutationsinvolving the entire genome (a spinoff from hisfallacious belief that the entire genome acted asan integrated unit), and small mutations withlarge impact upon adult phenotypes becausethey work upon early stages of ontogeny andlead to cascading effects throughout embryolog We reject his rst proposal, but the second,eminently plausible, theme might unite a Darwinian insistence upon continuity of geneticchange with a macroevolutionary suspicion ofphenetic discontinuity It s, after all, a majorfocus in the study of heterochrony (effects, oftenprofound, of small changes in developmentalrate upon adult phenotypes); it is also impliedin the emphasis now being placed upon regulatory genes in the genesis of macroevolutionarychange (King and Wilson 195)for regulationis fundamentally about timing in the complexorchestration of development. Moreover, although we cannot readily build "hopeful mon

sters, the subject of major change through alteration of developmental rate can be treated,perhaps more than analogically, both by experiment and comparative biology. The study ofspontaneous anomalies of development (teratology) and experimental perturbations of embryoenic rates explores the tendencies and boundaries of developmental systems and allows us tospecify potential pathways of macroevolutionary change (see, for example, the stunning experiment of amp 1959, on recreation of reptilian patterns in birds, after 00 million yearsof their phenotypic absence, by experimental


10/12


manipulations tat amount to alterations in rateof development for te bula) At te very least,tese approaces work wit real informationand seem so muc more fruitful tan te construction of adaptive stories or te invention ofypotetical intermediates

3. Te importance of non-adaptation Teempasis on natural selection as te only directing force of any importance in evolution ledinevitably to an analysis of all attributes of organisms as adaptations Indeed, te tendencyas infected our language, for, witout tinkingabout wat it implies, we use "adaptation asour favored, descriptive term for designatingany recognizable bit of canged morpology inevolution I believe tat tis "adaptationist program as ad decidedly unfortunate effects in

biology (Gould and ewontin, 1979) It as ldto a reliance on speculative storytelling in preference to te analysis of form and its constraints and, if wrong, in any case, it is virtually impossible to dislodge because te failureof one story leads to invention of anoter ratertan abandonment of te enterprise

Yet, as I argued earlier, te egemony of adaptation as been broken at te two lower levelsof our evolutionary ierarcy variation witin

populations, and speciation Most populationsmay contain too muc variation for selection tomaintain; moreover, if te neutralists are evenpart rigt, muc allelic substitution occurswitout controlling inuence from selection,and wit no direct relationsip to adaptationIf species often form as a result of major cromosomal alterations, ten teir originte establisment of reproductive isolationmay require no reference to adaptation Similarly, attis tird level of macroevolution, bot argu

ments previously cited against te conventionalextrapolationist view require tat we abandonstrict adaptationism

i) If trends are produced by te unidirectional transformation of populations (ortoselection), ten tey can scarcely receive oter tana conventional adaptive explanation After all,if adaptation lies beind single allelic substitutions in te Biston betularia model for cangein local populations, wat else but even strong

er, more persistent selection and adaptive orientation can render a trend tat persists formillions of years? But if trends represent a

iger-level process of differential og andmortality among species, ten a suite of potentially non-adaptive explanations must be considered Trends, for example, may occur because some kinds of species tend to speciatemore often tan oters Tis tendency may re

side in te caracter of environments or in attributes of beavior and population structurebearing no relationsip to morpologies tatspread troug lineages as a result of igerspeciation rates among some of teir membersOr trends may arise from te greater longevityof certain kinds of species Again, tis greaterpersistence may ave little to do wit te morpologies tat come to prevail as a result I suspect tat many morpological trends in paleontologya bugbear of te profession because we

ave been unable to explain tem in ordinaryadaptive termsare non-adaptive sequelae ofdifferential species success based upon environments and population structures

ii) If transitions represent te continuous andgradual transformation of populations, tentey must be regulated by adaptation trougout (even toug adaptive orientation may alteraccording to te principle of preadaptation)But if discontinuity arises troug sifts in de

velopment, ten directions of potential cangemay be limited and strongly constrained by teinerited program and developmental mecaics of an organism Adaptation may determineweter or not a opeful monster survives, butprimary constraint upon its genesis and direction resides wit inerited ontogeny, not witselective modelling

VI uo Vadis?

My crystal ball is clouded bot by te dust of

tese growing controversies and by te mists ofignorance emanating from molecular biology,were even te basis of regulation in eukaryotesremains srouded in mystery I tink I can seewat is breaking down in evolutionary teoryte strict construction of te modern syntesiswit its belief in pervasive adaptation, gradualism, and extrapolation by smoot continuityfrom causes of cange in local populations tomajor trends and transitions in te istory of

life I do not know wat will take its place asa unied teory, but I would venture to predictsome temes and outlines


11/12


The new theory will be rooted in a hiera

cal vew of nature. It will not emboy the

epressing notion that levels are fundamentally

istint and neessaily opposed to eah other ther dentaton of auses (as the older pa

leontologists held n mantaining that maro

voluton ould not, n prinple, be referred to

e same auses that regulate miroevoltion

g, Osborn ). It will possess a ommondy o auses and onstrants, but w reog

ze that they work in harateristially dffert ways pon the materal of differet levels

trademi hange, speiation, and patterns of

aroevolution.As ts seond major departure from urrent

rthodoxy, the new theory w restore to biol

y a onept of organism In an exeedngly

crious and unonsous bt of rony, strt se

ltonsm (whih was not, please remember,

Darwin's own view) debased what had been a

ainstay of biology-the organsm as an inte

rated entity exerting onstraint over its history

St George Mivart expressed the subtle pot

e in borrowing a metphor from Galton Iall al it Galton's polyhedron Mivart wrtes

(187, pp 8-9):

Ths oneption of suh iternal and latent

apabilities s somewhat like that of Mr Galton aordng to whh the organ worldonssts of entities, eah of whih s, as it

were, a spheroid wth many faets on its sur

fae, upon one of whih it reposes in stable

equilibrium When by the aumulated aton of ndent fores this equlibrium s dis

turbed, the spheroid s supposed to turn over

untl t settles on an adjaent faet one more

in stable equlbrium The internal tendeny

of an oransm to ertan onsderable and

denite hange would orrespond to the fa

ets on the surfae of the spherod

Under strt seletionism, the organism is a

here It exerts lttle onstrant upon the harte of ts potential hange; t an roll along all

aths Geneti varaton s opious, small n ts

nrements, and avalable n all diretonsthe

ssene of tte term "ndo as used to guarntee that varaton seres as raw materal onlyand that seleton ontrols the dreton of evo

lton

med no retun to the antquated and antiar

wnan view that mysterious internal fators

provide diretion inherently and that seleon

only elimnates the unt (orthogeness variousorms of \talsm and nalsm nstead the a

ets are onstraints exerted by th de\elopmentalintegraton of organisms themselves. Change

annot our in all iretions, or wth y in

rement the organism is not a metapora

sphere When the polyhedron tumbles, seleto

may usually be the propellng for But f ad

jaent faets are few in number and wde n

spang, then we annot dentify seletion as the

only, or even the primary ontrol upon evoluton For seletion is hanneled by the for of

the polyhedron t pushes and these onstrant

may exert a more powerful inuene upon evo

lutionary diretons than the external push t

self Ths is the legitmate sense of a muh ma

gned aim that internal fators are important

in evolution They hannel and onstran ar

winian fores they do not stand n oppostion

to them Most of the other hanges in evolu

tonary viewpoint that I have advoatedthroughout ths paper fall out of Galton's et

aphor: puntuatonal hange at all levels the

ip from faet to faet, sine homeostat sstems hange by abrupt shfting to ne equlb

ria); essental nonadaptaton, even n ajoparts of the phenotype (hange in an integated

organsm often has effets that reverbeate

throughout the system); hanneling of diretion

by onstraints of history and developmental ar

hiteture Organisms are not billiard ballsstruk n determnst fashon by the ue of nat

ural seleton, and rollng to optimal postions

on lfe's table They nuene ther own detn

in interestng, omplex, and omprehensble

ways We ust put th onept o oranis

bak nto evolutionary bology

Literature Cited

YALA F] 197 ca c, a " Pp. 1-20I: aa F ] o E\' A caSuderlad. Mass

BATESO", G 198 Mid d ; P D "el' YkBVSH G L 1975 des f ama spca A \ c

Syst 6 :33-3BUS, G L, S CA A C \ILSO, AD ] L PO

197 Rapid peciati ad chrmsma evlio i aP a cad Sci 7432-39

C H L 15 T i f d d


12/12


ARSON, H. L . 1978 Chromosomes and species formation Ev

lution 32:95-97

ARWIN, C 1859 On the Orgin of Species 490 pp John Murray;

London

ARWIN, C 1880 Sr Wylle Thomson and natural selection

Nature 23

OBZHANSKY, H. 197 Genetcs and the Orgin of Speces 64

pp Columba Uni Press; New York

OBZHANSKY, H. 1951 Genetcs and the Orign of Species (rd

ed) 64 pp Columbia Un Press New York

HRLICH, P R AND P H AVEN. 1969 Differentation of pop

ulatons Science 18-1

LDREDGE, N 1979 Alternate approaches to eolutionary the

ory Bull Carnege Mus Nat Hst pp 7-19

LDREDGE, N AND J CF 1980 Phylogenetc Patterns and

the Eolutonary Process Columbia U Press; New York

LDREDGE, N AND S J OULD 197 Punctuated equilbra An

alternate to phyletc gradualsm Pp 8-115 In Schopf, T J

M, ed Models n Paleobiology reeman, Cooper and Co; San

rancsco, Calforna

OLDSCHMIDT, R 1940 The Material Bass of Eoluton 46 pp

Yale Uni Press; New Haen, Conn

OULD, S J 1980 G G Smpson, paleontology and the modern

synthess In Mayr, E, ed Conference on the makng of the

Modern Synthess Harard Uni Press; Cambridge, Mass

OULD, S J AND N LDREDGE. 1977 Punctuated equlibra the

tempo and mode of eoluton reconsdered Paleobology 3115-

151

OULD, S J AND R C EWONTIN. 1979 The spandrels of San

Marco and the Panglossan paradgm A crtque of the adapta

tonist program Proc R Soc London 2581-598

AMP, A 1959 Contribution a l'ude du deloppement et dela regulaton des dcences et des excdents dans la patte de

l'embryone poulet Arch Anat Microsc Morphol Exp 45-

478

ING, M C AND A C ILSON. 1975 Eoluton at two leels n

humans and chimpanzees Science 107-116

OESTLER, A 1978 Janus a Summng Up Random House; New

York

EWONTIN, R C 1974 The Genetc Bass of Eoluton ary

Change 46 pp Columba Uni Press; New York

EWONTIN, R C. AND J L UBBY. 1966 A molecular approach

to the study of genic heterozygosty in natural populations II.

Amount of ariaton and degree of heteozygosity in natural pop

ulatons of Drosophil pseudoobscur. Genetcs 595-609AYR, E 194 Systematics and the Orgin of Species 4 pp

Columbia Uni Press; New York.

AYR, E 196 Anmal Speces and Eoluton 797 pp Belknap

Press of Harard Uni. Press; Cambrdge, Mass

IVART, T G 1871 On the Genesis of Species 96 pp

MacMllan; London

N, M. 1975 Molecular Population Genetics and Eolution

American Elseier; New York

SBORN, H 19 Orthogenesis as obsered from paleontolog

ical eidence begnnng n the yea 1889 Am Nat 14-14

STROM, J H 1979 Brd ght How did it begin Am Sci 46-

56

OWELL, J R 1978 The founderush speciaton theory an ex

permental approach Eoluton 32465-474

OBSON, G C AND O W ICHARDS. 196 The Variaton of

Anmals n Nature Longmans, Green, and Co; London

OMANE, G J 1900 Darwn and after Darwin, ol Post

Darwinian questons Heredty and Utlty 44 pp. Longmans,

Green, and Co; London.

IMPSON, G G 1944 Tempo and Mode n Eoluton 7 pp.

Columba Un Press; New York

IMPSON, G G 195 The Major eatures of Eolution 44 pp

Columba Un Press; New York

TANLEY, S M 1975 A theory of eoluton aboe the species

leel Proc Nat Acad Sc 2646-650

HITE, M J D 1978 Modes of Speciaton 455 pp WH. ree

man; San rancisco, California

ILSON, A C, G L USH, S M ASE, AND M C ING.

1975 Socal structuring of mammalian populations and rate of

chromosomal eolution Proc Nat Acad Sci 25061-5065

ILSON, A C, S S ARLSON, AND T J HITE 1977 Bio

chemcal eolution Annu Re Bochem 57-69

ILSON, E O ET AL 197 Lfe on Earth Snauer Associates;

Sunderland, Mass

RIGHT, S 1967 Comments on the prelmnary workng papers

of Eden and Waddington In Moorehead, P S and M M Kap

lan, eds Mathematcal Challenges to the NeoDarwnan Theory

of Eolution Wistar Inst Symp 117-10

Documents

Is a New and General Theory of Evolution Emerging