Fifty years of progress in research on species and speciation

10 Pages·1993·5.3 MB·English

by  Ernst Mayr

PROCEEDINGS
OF THE
CALIFORNIA ACADEMY OF SCIENCES
Vol. 48, No. 6, pp. 131-140 December 21, 1993
FIFTY YEARS OF PROGRESS IN RESEARCH ON
SPECIES AND SPECIATIONanne
Biological Labo
*oods Hole OceiograScStion
By Ub«7.
Ernst Mayr
MuseumofComparativeZoology,HarvardUniversity,
Cambridge,Massachusetts02138
Woods Hole, MA 02543
AdaptedfromalecturedeliveredattheGoldenJubileeCelebrationofthepublicationofSystematicsandthe
OriginofSpeciesattheCaliforniaAcademyofSciencesonOctober16, 1992.
ReceivedDecember31, 1992.Accepted February 11, 1993.
Historiansofsciencehavetaughtushowmuch the nature of the genetic material, resulting in
onecanlearnfromstudyingthehistoryofafield the theory ofparticulate inheritance. However,
ofscience. This is excellently illustrated by the they drew from this the wrongconclusion as far
history ofevolutionary biology as a whole, and asevolutionisconcerned,claimingthatnewspe-
byourgrowingunderstandingofspeciesandspe- cieswereproducedbynewmutationsina single
ciation, in particular. saltation, completely rejecting Darwin's theory
After 1859, two ofDarwin's theorieswere ac- ofgradualism. Their opponents were the biom-
cepted almost at once. First, evolution as such, etricians, suchas PearsonandWeldon,whocor-
and secondly, the branching theory ofcommon rectly insisted on the gradualness of evolution
descent.Naturalselectionwaswithalmostequal but incorrectly claimed that inheritance was
unanimityrejected,beingacceptedonlybyasmall equally gradual, that is, blending. As far as ge-
groupofnaturalists. Thiswas nottoo surprising netics is concerned, the Mendelians were right;
since at that time no one understood variation as far as evolution is concerned, the biometri-
anditsorigin. Finally, anactive, almostviolent, cians were right. There was no genuine popula-
controversy developed over two other major tion thinking in either camp and the biometri-
Darwinian theories, that ofspeciation and that ciansandotheropponentsofMendelismadopted
ofevolutionarygradualism. As a matteroffact, Lamarckian inheritance in order to account for
Darwin himselfwas vacillating with respect to the gradualness ofevolution.
these two theories. In the ensuingyears, thegap between the two
Let us now proceed to the year 1900 and the camps narrowed appreciably as a result ofthe
rediscovery ofMendel's laws. At that time, two new findings ofgenetics and systematics. Even-
camps became established in evolutionary bi- tually the biometricians disappeared from the
ology. One consisted ofthe Mendelians repre- scene and were replaced by a group of evolu-
sentedbyBateson,DeVries,andJohannsen.They tionists I shall call the naturalists. At the same
werestricttypologistswhosawdiscontinuityev- time the Mendelians were replaced by the pop-
erywhere in nature and applied this correctly to ulation geneticists.
U3i]
132 PROCEEDINGSOFTHECALIFORNIAACADEMYOFSCIENCES,Vol. 48,No. 6
The major difference between these two new thecausesofevolutionarychangeandspeciation
constellations ofevolutionists was their sphere is well-documented by the widely read work of
ofinterest. To document this it is necessary to Robson and Richards (1936).
define"evolution."Thegeneticistsadoptedare- At that time there appeared on the scene a
ductionist definition that, as we now see it, was young beetle systematist who had grown up in
quite misleading. They defined evolution as the the thinking of the Russian tradition with its
change ofgene frequencies in populations. This emphasis on organic diversity but who also had
definition emphasized the wrong level ofactiv- had the advantage ofnearly 10 years ofwork in
ity. Evolution is the story ofadaptation, ofthe (and stimulation by) an American laboratory in
develomentofnewkinds ofanimals andplants, populationgenetics. I am, ofcourse, referringto
ofthe origin ofmodes ofreproduction, and of TheodosiusDobzhansky,whointegratedthetwo
all aspects of the history of organisms. Evolu- great traditions I have just described and who
tionary biology deals not merely with genes but produced what could almost be called the Bible
with two major sets ofproblems: (1) the acqui- of the evolutionary synthesis, his magnificent
sition and maintenance ofadaptedness, and (2) GeneticsandtheOriginofSpecies(1937).Inspite
the origin and nature oforganic diversity. ofsomeomissionsandevenafewoutrighterrors,
Thegeneticistsdealtonlywith theproblem of thisworkcontainedthegistofthenewparadigm
adaptedness. Their approach was reductionist, oftheevolutionarysynthesis.Withinafewyears
concernedwiththegeneticchangeswithinapop- the synthesis was completed in zoology, as doc-
ulation. It dealt only with the time dimension, umentedbythepublicationsofJ.Huxley(1942),
with what one might call "vertical evolution." E. Mayr (1942), G. G. Simpson (1944), and B.
The naturalists concentrated on the other as- Rensch (1947). In 1950 G. Ledyard Stebbins
pectofevolution,thenatureoforganicdiversity. brought in botany, showing in his monumental
Their interest was in populations, species, and Variation andEvolution in Plantsthat the prin-
macroevolution,withparticularemphasisonthe ciples developed in the synthesis were equally
process ofspeciation and thegeographical com- applicable to plants, contrary to the claims of
ponents ofevolution, what one might call the some other botanists. This is also true even for
"horizontaldimension."Thisdifferenceinbasic the numerous special phenomena and processes
concerns was, in my opinion, more important encounteredin plants(MayrandProvine 1980).
than their differences in the genetic interpreta- Thesynthesisofthe 1930s-40swastheendof
tion. old arguments, in particular the final refutation
There were also conspicuous national differ- ofthe various non-Darwinian theories ofevo-
ences. Evolutionary genetics flourished in the lution, but itwasalso thebeginningofa newset
Anglophonecountriesasindicatedbythenames ofcontroversies. My limitation oftime permits
T. H. Morgan, H. J. Muller, S. Wright, R. A. me only to deal with those concerning species
Fisher, and J. B. S. Haldane, while the study of and speciation. Even under that limitation one
organic diversity in an evolutionary manner has to cover scores ofbooks and thousands of
flourishedinRussia,Scandinavia,andGermany, smaller publications. This forces me to present
but was poorly represented in English-speaking myfindingsasthefinalconclusionsoflongdrawn-
countries, in the United States byF. B. Sumner, outargumentsandperhapsseeminglyinarather
L. R. Dice, D. S. Jordan, andJ. Grinnell, and in dogmatic manner.
England by E. Poulton and E. B. Ford. I am often asked, what in particularhad been
We thus had two ratherdifferenttraditions in my own contributions to the Evolutionary Syn-
the 1920s and early 1930s: an Anglophone ge- thesis?Theycanberecordedunderthetwohead-
netic tradition studying the vertical component ings: species and speciation.
ofevolution,thatis,adaptivegeneticchange,and
an essentially continental European tradition in
systematics studyingthe horizontal component,
Species
that is the geographical changes ofpopulations
leadingtospeciationandmacroevolution.Aslate Asfarasspeciesareconcerned,Idemonstrated
astheearly 1930s(upto 1935 and 1936),several theweakness, ifnotinvalidity, ofthepreviously
authors declared that the gap between the two mostpopularspeciescriteria,particularlyasstat-
camps was unbridgeable. The confusion about ed in the morphological and genetical species
MAYR: SPECIESANDSPECIATION 133
definitions. Instead, I promoted acceptance of definition: "an evolutionary species is a lineage
the biological species concept with its emphasis (an ancestral-descendant sequence of popula-
on populations and on reproductive isolation: tions) evolving separately from others and with
"A species is a group of interbreeding natural its own unitary role and tendencies." The re-
populations that is reproductively isolated from placement ofthe clear-cut criterion ofthe bio-
other such groups." I was not the first to adopt logical species (reproductive isolation) by such
the biological species concept but there is little undefinedvaguetermsas"evolutionarytenden-
doubtthatitwasthesupportIgaveitinmy 1942 cies" and"evolvingseparately" doesnotpermit
book that led to its rapid subsequent adoption. discrimination between good species and iso-
In particular, I pointed out the weakness ofthe lates. Itisnotapplicabletopolytypicspeciesthat
morphologicaldefinitionbecauseit providedno contain geographical isolates. Nor does it even
criterion by which to determine the status of permitthedelimitationofan"evolutionaryspe-
highly distinct intraspecific variants. Another cies" within a phyletic lineage. (For a more de-
weaknessofthemorphologicaldefinitionwasits taileddiscussion see Mayr 1988a:323.) Simpson
inability to cope with a phenomenon for which wasnot raisedasanaturalistand, in spite ofhis
I introduced the term "sibling species," that is, biometric work on samples offossils, his mate-
morphologically virtually identical populations rial did not allow a study ofgeographic specia-
that were nevertheless reproductively isolated. I tion.
expandedmytreatmentofsiblingspeciesin 1948 Anothergroup ofopponents ofthe biological
and 1963, and their extreme frequency is now speciesconceptconsistedofcertainmuseumand
generallyacknowledged.Previouslymostofthem, herbariumtaxonomists.Theyhadtoassignspec-
ifrecognized at all, had been listed as biological imens to species, particularly such from widely
races. After 25 years ofargument, I finally per- distant geographic locations, and were puzzled
suaded even Tracy Sonneborn to recognize the what criteria to use in order to infer whether or
so-calledvarietiesofParameciumassiblingspe- not these isolates were reproductively isolated.
cies. I believe there are some 14 such sibling As a result, they returned to a typological/mor-
species in the Paramecium aurelia group alone phological speciesconcept. Butthisat oncebur-
(Sonneborn 1975).Recentmolecularstudieshave dened them again with the two formidable de-
shown that most sibling species are as different ficiencies ofthe morphological species concept,
from each other on a genetic-molecularbasis as the treatment ofsibling speciesand ofpolymor-
are morphologically distinct species. phism. Both problems require a biological spe-
Theadoptionofthebiologicalspeciesconcept cies concept for their resolution. Whatever de-
was perhaps even more important for the field cisionone makes, itcan onlybeinferredbutnot
worker. Ecologistsandstudentsofbehaviorusu- proven. The procedure ofsuch inference is de-
ally work in a local situation, and there the rec- scribed in Mayr and Ashlock 1991:100-105.
ognition of species as non-interbreeding, coex- Here I must make a short aside. Those who
isting populations is ofthe utmost importance. studyspeciesmostintenselyarenaturalistswork-
I also showed that in most species that are not ingatagivenlocality. Theystudytheinteraction
host-specific, there is considerable geographic ofspecieseitheras members ofan ecosystem or
variation requiring the adoption of polytypic in connection with the study ofbehavioral in-
species. Suchgeographicvariation maybeclinal teractions among different species. The ranking
when populations are continuous, or discontin- of geograpically distant populations is usually
uous when populations, particularly peripheral irrelevant for their objectives. For instance, for
populations,arespatiallyisolatedfromeachoth- someone who studies the song sparrows ofthe
er. Throughout I continued to emphasize that San FranciscoBayregionfromthepointofview
species are not types but populations. of ecology, adaptation, or behavior, it is quite
The biological species concept, based as it is irrelevant whether one calls the song sparrows
on population thinking, was not palatable to ofthe Aleutian Islands conspecific or a full spe-
workersinseveralfields.Thepaleontologists,for cies. HoweveronerankstheAleutiansongspar-
instance, who study species in the time dimen- rowwill have no effect whatsoever on the study
sion, looked fora species concept that would be ofthe song sparrows ofthe San Francisco Bay
particularlysuitableforthedelimitationoffossil region.
species.Here,Simpson(1961:153)proposedthis In a study ofthe over 600 species ofNorth
134 PROCEEDINGSOFTHECALIFORNIAACADEMYOFSCIENCES, Vol.48, No. 6
American birds undertaken jointly with L. L. Inothergroupswithmanylocalizeddiagnosable
Short (1970), we found that with one single ex- populations the inflation might be even greater.
ception (Pipilo) all ofthem were fully consistent Newpaperssupportingorproposingunortho-
with the biological species concept. In a recent dox species concepts continue to be published
analysisofthe vascularplantsofConcordtown- quitefrequently,sometimesentirevolumesstress
shipinMassachusetts(Mayr 1992a),Ifoundthat suchdefinitions(e.g., Otteand Endler 1989). As
thenumberofcaseswherethebiological species far as I am concerned, none ofthese proposals
concept led to an ambiguous decision was less strikes me as particularly convincing. Coyne
than 10%.Evenhere,thedifficultiesweremostly (1988), in a review in Nature of one of these
thoseofinsufficientscientificanalysisratherthan volumes, refers to the commotion produced by
aconsequenceofusingthebiologicalspeciescon- the new proposals and then continues, "When
cept. Professor Stebbins has recently informed the dust has settled only Mayr is still on his feet
me that he found the number ofspecies causing with hisoriginalconceptremainingthe simplest
difficultiesintwolocalflorastobearound4-6%. andmostusefulviewofspecies." ForawhileH.
I am quoting these figures merely in order to Paterson's(1981)"recognitionconcept"wasquite
makethe pointthattheclaim thebiological spe- popularbutneitherPatersonhimselfnoranyone
cies concept is inapplicable to plants is not sub- else seems to support it any longer after it was
stantiatedwhenanactualanalysisofalocalflora shown that, first, itdid not reallydifferfrom the
is undertaken. biologicalspeciesconcept,andsecondly,thatPa-
In theirendeavorto apply cladistic principles terson'sattackofthelatterconceptwasbasedon
even to intraspecific populations, that is to the a number ofmisconceptions (Mayr 1988b, and
verylowestbranchingpoints,somecladistshave Coyne et al. 1988).
recentlyproposedaso-called "phylogeneticspe- The major reason why it is impossible at this
cies"concept. Indeedtherearenowatleastthree time to say the last word on species concepts is
versionsofthis concept inexistence (Nixon and that the biological species concept is based on
Wheeler 1990). This concept was first suggested thestudyofdiploid,sexuallyreproducingorgan-
by Rosen (1979:277) who proposed to consider isms, forming standard biological populations.
the lowest population or population aggregate Yet our knowledge ofthe population structure
showing a new character (apomorphy) as a sep- and mode of reproduction of many groups of
arate species: "a 'species' is merely apopulation lower invertebrates, lower plants, protists, and
orgroup ofpopulations defined by one or more particularly prokaryotes is altogether insuffi-
apomorphousfeatures,itisalsothesmallestnat- cient. Many ofthem clearly do not fit the stan-
uralaggregationofindividualswitha specifiable dardbiologicalspeciesconcept. Theplacewhere
geographic integrity that can be defined by any thebiologicalspeciesconceptrunsintoparticular
current set of analytical techniques." His defi- difficulties is in its application to asexually (uni-
nition would have required raising the popula- parentally)reproducingorganisms. Thesedonot
tionofjustabouteverytributaryofeveryCentral formpopulationsinthesenseofthepopulations
American riverto species rank, for nearly all of ofdiploid sexually reproducing organisms, and
them have some special color gene or other pe- therefore a species definition based on popula-
culiar characteristic. The most widely accepted tions is inapplicable. It will be up to the spe-
definition of the phylogenetic species is: the cialists of such organisms to develop a species
smallest cluster oforganisms which is diagnos- definition that is particularly suitable for them.
ably distinct from other such clusters. The real Thequestionisoftenaskedwhatinfluencemo-
purposeofthephylogeneticspeciesistoserveas lecularbiology has had on ourunderstandingof
"thesmallestunitsuitableforcladisticanalysis" speciesandspeciation.Iconcludethatithasmade
(Nixon and Wheeler 1990:212). I am not aware three major contributions.
ofanybiologicalsignificanceofthisunit.Toadopt First, it led to the discovery or confirmation
this reductionist approach would lead to a mas- ofmanysiblingspecies.Biologicalracesandoth-
siveincreaseinthenumberofrecognizedspecies ersuspectedsiblingspecieswereoftenshownby
inallgroupswithgeographicalvariationandiso- molecularmethodstobejustasdifferentasmor-
lation. I estimate that birds would have over phologically highly distinct species. Further-
20,000phylogenetic"species,"asdefinedabove, more, in the case ofmorphologically highly sta-
insteadoftheabout9,500orsonowrecognized. blegroupsofspecies,relationshipscansometimes
MAYR: SPECIESANDSPECIATION 135
be worked out very reliably by molecularmeth- zhanskyfavoredtheWallacetheory, Mullerand
ods, and more than that, the branching point I the Darwin theory. All the recent studies of
between various lineages can often be assigned secondaryhybridzonesindicatethatDarwinwas,
to definite points in the geological time scale. on the whole, right. Recent developments indi-
Second,whenthereisdoubtconcerningapar- catethatbehavioralisolatingmechanismsinan-
ticularvariety, whetherit is merely an intraspe- imals may well be due to a change offunction
cific variant or a good species, molecular meth- ofpropertiesacquiredasaresultofsexual selec-
odscanusuallygiveaclear-cutanswer.Nomatter tionduringthepreviousisolation(Mayr 1988b).
howdifferenttheymayappear, intraspecificvar- In 1942 I distinguished four types ofspecia-
iants usually differfrom othermembers oftheir tion: (1) geographical (allopatric), (2) semi-geo-
population by only very few genes. graphical (now called parapatric), (3) sympatric,
Third, in asexually reproducing organisms, and(4)instantaneous.Ididnotrejectotherforms
molecularmethodshavecastagreatdealoflight ofspeciation outright, but I insisted that, as far
ontheamountofdifferenceamongvariousclones as higher organisms are concerned, allopatric
and on various cryptic methods of gene ex- speciation was the mostcommon mode. I think
change. itislegitimatetostatethatthisevaluationisstill
It must be stressed that, opposing claims not- valid today. It is ofinterest to note that by far
withstanding, molecular methods have not, in thelongestchapterofmy 1942bookwasdevoted
anyway,weakenedthebiologicalspeciesconcept towhatI calledthe "biology ofspeciation," that
nor have they affected the standard interpreta- is, all the ecological and behavioral factors in-
tion ofspeciation. volved. SofarasIknowtherewasno suchtreat-
ment in the earlier literature.
Perhaps my major contribution was that I
Speciation
solvedtheoldconundrumofhowonecouldrec-
The recognition that species are populations, oncilethesharpdemarcationofspeciesinalocal
not types, was particularly important in the ex- fauna and flora with the Darwinian concept of
planation ofspeciation. I reported in 1942 that gradual evolution. I demonstrated that in the
in birds, mammals, butterflies, and snails, that local situation species are indeed sharply sepa-
is, inalltaxonomicallywell-studiedgroups, spe- rated by gaps, but that ifone looks at a species
ciation invariably turned out to have been geo- taxon in its total geographical representation
graphical.Thismeansthatapopulationthathad throughitsentirerange, onefindsthatmostspe-
been isolated by geographical or vegetational cies consist ofa large aggregation oflocal pop-
barriers had acquired genetical isolating mech- ulations. Some ofthese, particularly those iso-
anismsduringthisgeographicalisolationandthat lated at the periphery ofthe species range, are
thissubsequentlypermittedittocoexistwiththe actually incipient species, that is, in transition
parental specieswithout interbreeding. Here my from the status oflocal population to that ofan
studies of island faunas were particularly im- independent new species. This refuted the old
portantbecause itenabled me to show thatgeo- claim ofDarwin's opponents that the sharp de-
graphic speciation is a continuous process: pop- limitation oflocal species, emphasized by nat-
ulations on the most recently colonized islands uralistsfromLinnaeuson,wasincompatiblewith
are still almost identical with the source popu- Darwin'sconceptofgradualevolution. Thepuz-
lationwhile the longeran islandpopulation had zle is solved by expanding the non-dimensional
been isolated, the more different it was, until species ofthe local naturalist to the geographi-
finally after a sufficient time interval, complete callyvariablespeciesinamulti-dimensionalap-
species status had been reached. proach. Furthermorethisshowedthatitwasun-
How the isolating mechanisms were acquired necessary, indeed incorrect, to postulate
continued to be controversial. There were two speciation by saltation because geographic spe-
opposingtheories,thatofAlfredRusselWallace, ciation is a gradualistic process.
whothoughttheywereacquiredbyselectionwhen To explain how new species originate had
the previously isolated populations came into clearlybeenthe mostimportantobjectiveofmy
secondarycontact, andthatofDarwin, whosaid SystematicsandtheOrigin ofSpecies. Asiswell
that selection could never complete the process known, Darwin had supported geographic spe-
of speciation under those circumstances. Dob- ciationuptothe 1840sbuthadbecomeuncertain
136 PROCEEDINGSOFTHECALIFORNIAACADEMYOFSCIENCES,Vol. 48,No. 6
when encountering certain situations in plants distant from the core area ofthe species, tended
and eventually allowed for massive sympatric to be the ones most different. Indeed, some of
speciation (Mayr 1992b). And that was still the them were so different that they had been de-
majority opinion among biologists when I gave scribed as different genera (Mayr 1954).
the Jesup lectures in 1941. And this indicates the second reason for the
Thenumberofconfusionsexistingatthattime importance ofperipatric speciation. It provides
was considerable. Many authors, for instance, a bridge from speciation to macroevolution. I
made nodistinctionbetweenphyleticchange(as publishedthese ideasin 1954, andEldredgeand
observedby paleontologists)and an actual mul- Gouldbased theirtheory ofpunctuated equilib-
tiplication of species. Other authors, such as ria on it in 1972. This theory, in its most sim-
Goldschmidt(1940), inthetraditionoftheorig- plifiedversion, statesthat manyseeminggapsin
inal Mendelians, still thought ofspeciation as a the fossil record are due to the fact that such
phenomenoninvolvingasingleindividualgiving highly localized events, as genetic reorganiza-
rise to a new species. tions in founderpopulations, are not likelyever
Soon after 1942 it became clear to me that to be found in the fossil record, and secondly
geographicspeciationwasnotthesimpleunitary thataftersuchspeciationiscompletedandanew
phenomenon I had first thought. I discovered well-balanced genotype has been formed, evo-
that there are indeed two kinds of geographic lution willgreatly slowdown and there may fol-
speciation. In the classical type, now also called low a period ofstasis lasting millions ofyears.
the dumbbell model or dichopatric speciation, In the long controversy about punctuated equi-
the range ofa moreorlesswidespreadspeciesis libria, itwould seem to me that more factswere
divided by a barrier, resulting in two isolated published supporting the theory (in the simpli-
parts ofthe species. These in due time may be- fied version here presented by me) than facts
comesufficientlydifferenttoacttowardeachoth- opposing it. However, most likely we have plu-
er as good species. The other mode, later called ralismhereasinmostevolutionaryphenomena,
by me peripatric speciation (Mayr 1982), takes and in a few cases a phyletic lineage may dras-
placewhensomedispersingindividualsofaspe- tically change in the course oftime, also under-
ciesestablishan isolatedfounderpopulationbe- goingchangesofmacroevolutionarysignificance.
yond the species periphery and this population When we go back to the classical writings of
becomes in due time sufficiently different and Fisher and Haldane ofthe early 1930s we find
acquires the necessary isolating mechanisms to that they, by only considering additive gene ef-
rank as a separate species. fects,assumedthatevolutionwouldproceedmost
Such speciation through the founder effect is rapidly in large, populous species. The more re-
ofdoublesignificance.First,thegeneticvariation centevidencerefutesthisbelief. Letmeaddthat
among the few founder individuals, sometimes the isolation of an incipient species has to be
only a single fertilized female, is only a small pretty nearly complete. Situations such as those
sample ofthe total genetic variability ofthe pa- described by Sewall Wright oftemporarily iso-
rentalspecies. Inbreedingduringthefounderpe- lated demes may contribute to the variability of
riodwillleadtofurtherlossofgeneticvariability. aspecies,butnottotheformationofnewspecies.
This bottleneck effect may lead to a shift in ep- Evolutionary changes corresponding to peri-
istaticinteractionsandtoaconsiderablerestruc- patric speciation take place also in refugia. As
turing ofthe genotype. Such a genetic reorgani- wasfirstpointedoutbyStresemann(1919),later
zation takes place only in maybe 1 out of50 or bymyself(1942), andmorerecentlyparticularly
1 out of500 ofthe founder populations. But if by Haffer (1969), contracting relict populations
ittakesplaceitprovidesan opportunityforma- resultingfromclimaticchangesmayalsobecome
jorshiftsinadaptation, particularlysincethese- incipient species and with a greatly accelerated
lectingfactors in the founder population are apt rateofdivergenceduringtheirperiodofreduced
to be quite different from those in the parental population size.
population. I arrived at this model ofperipatric
speciation noton thebasisoftheoreticalconsid- Unorthodox Modes of Speciation
erations but rather because I observed that in
South Sea islands' birds, nearly always the most Even though geographic speciation was ac-
isolated populations, or the populations most cepted by most evolutionists as by far the most
MAYR:SPECIESANDSPECIATION 137
frequent mode ofspeciation, a minority ofau- years. It is clearly associated with the Wallace-
thors promoted the occurrence of alternate Dobzhansky theory of the sympatric origin of
modes. One can assign these various proposals isolating mechanisms. The genetic analysis of
to three classes. numerous cases of secondary hybrid belts has
Instantaneous speciation occurs by the pro- shown that such belts are indeed secondary and
duction ofan individual that is reproductively do not lead to a gradual strengthening of the
isolatedfrom theparental species. Actually, one isolating mechanisms. The only seeming excep-
typeofinstantaneousspeciation,allopolyploidy, tionsarecasesofrapidlyspreadingspeciesover-
produced by the doubling ofthe chromosomes running part of the range of a closely related
ofa sterile species hybrid, is quite common in species, a situation in which the first colonists
plants and not controversial. A different type of have a greatly reduced opportunity for mating
such instantaneous speciation is not infrequent withconspecificindividuals, with theresultthat
in certain groups of animals, that is, a shift of occasionalhybridizationoccurs.Suchhybridiza-
specieshybridstoparthenogeneticreproduction. tion ends when the colonization is completed.
Such cases have been particularly well analyzed As farasthereinforcementofisolatingmech-
in lizards and fishes (White 1974). anisms is concerned, one must make a distinc-
White (1978)proposedone furthertypeofin- tionbetweenspeciesandpopulations.Wheretwo
stantaneous chromosomal speciation, called by closely related species overlap, indeed one finds
him stasipatric speciation. In this model, a new sometimesacharacterdivergenceintheisolating
chromosomeisproducedthroughachromosom- mechanisms. On the other hand, it is now well
al mutation which, although somewhat inferior established that gene flow and recombination
asheterozygote,canconqueranewsuperiorniche preventsuchanoccurrenceintheareaofmeeting
when homozygous. And then it can live side by ofconspecific populations.
side with the parental species owingto selection Finally, there is the possibility of sympatric
againstthedeleteriousheterozygotes. Whitewas speciation, that is, the development ofisolating
led to this model by the observation that allo- mechanisms within the cruising range of indi-
patric species in morabine grasshoppers invari- vidualsoftheparentalspecies.Asiswellknown,
ablydifferfromeachotherinsomechromosomal Darwin, forvariousreasons, finallyacceptedthe
rearrangement that is inferior in the heterozy- widespread occurrence of sympatric speciation
gotes.However,asKey(1981),Mayr(1982),and (Mayr 1992b). For80yearsitwasconsideredthe
others have since shown, White's model is not prevailing mode of speciation by most of his
supportedbyanyfacts.Ifitwerevalid,oneshould followers, particularly the entomologists. Even
find many cases ofreproductively isolated new though some perceptive evolutionists, like Ed-
chromosome types inside the range of the pa- ward Poulton and particularly Karl Jordan, had
rental species but one never finds this. It is far refuted numerous putative cases of such sym-
simplerto postulate, andthisis supportedbyall patric speciation, it was still a popular, if not
known facts, that the chromosomal rearrange- majority,beliefin 1942.WilliamThorpe(1930),
ment takes place in a small founderpopulation. forinstance,publishedseveralpapersdescribing
The mistake made by White was to ask, is spe- biological races believed by him to be incipient
ciation chromosomal or geographical? In fact, sympatric species. My own researchesled me to
however, all chromosomal speciation is simul- claimin 1942, 1947,and 1963thatallthesecases
taneously also geographical, and there is no dif- couldbeaswellorbetterexplainedbygeographic
ficulty in developing a model which combines speciation and furthermore, that most so-called
both geographical and chromosomal speciation. biologicalraceswerenothingbutgoodbiological
A second model not involving complete geo- species even though morphologically very sim-
graphic isolation is so-called parapatric specia- ilar, that is, they were sibling species. I never
tion. Here it is postulated that isolating mecha- denied the possibility of sympatric speciation,
nisms between two continuously distributed but I denied that this form of speciation was
populations arise along an ecological escarp- indeed substantiated by the proposed cases. In
ment. Thismodel ofspeciation, particularly fa- recentyears,GuyBushhasadvancedagreatdeal
vored by Murray and Clarke (1980) and by En- ofevidencebelievedbyhimtosupportsympatric
dler(1977), has not been substaniated in any of speciation (Bush and Zwolfer 1984). He states,
the very numerous test cases analyzed in recent for instance, that there are more than 100,000
138 PROCEEDINGSOFTHECALIFORNIAACADEMYOFSCIENCES,Vol. 48, No. 6
species ofinsects where mating and egg laying in a flourishingcolony ofone species, inquilines
takesplaceon the hostplant. He believesa shift ofanother species are found (Buschinger 1990).
toa new host plant might quickly result in sym- Thelatterare so-to-speakparasiticbecause they
patric speciation. He minimizes, however, the have no workercasteand benefitfrom the labor
numerousdifficultiesas, forinstance, the occur- of the workers of the host species. In the few
rence ofback colonization ofthe parental host casesthathavebeenwelldescribed,theinquiline
and the development of a two-host species, species seems to be most closely related to the
pointedout by various authors. Such difficulties host speciesand there is no mode ofgeographic
disappear when one assumes that the host shift speciation that could produce such a result.
takes place in a founder population. Bush's ev- Inspiteofalltheseadvancesandclarifications,
idence has been severely criticized by Futuyma there are still areas ofgreat uncertainty. For in-
and Mayer (1980), by Jaenike (1981), and by stance, how do pelagic species speciate? Each of
Paterson (1981). In the best studied case, the such species is usually associated with a partic-
original host of the fruit fly species Rhagoletis ular water mass, but how can such a species be
pomonella is hawthorne {Crataegus). After the divided into two? There is a possible scenario.
fly had colonized apple orchards around 1850, Onecanassumethatsomeperipheralportionof
the apple population developed certain differ- suchawatermassbecomesisolatedbyadifferent
encesinlifecycleandmorphologyfromtheorig- water mass intervening between it and the re-
inal hawthorne population. However, vast mainderofthespeciespopulationandthateven-
monocultures, like apple orchards, provide an tually this isolated population acquires species
unusualsituation.Ispendmysummersinsouth- status. However, except for one or two cases of
ernNewHampshire,wheretherearesquaremiles fossil species, there is so far very little evidence
ofapple orchards, and yet in all my botanizing indicating the validity ofthis scenario.
in that area I have not yet found a single wild Asinthecaseofspecies,ourconclusionsabout
Crataegustree. This population ofRhagoletisis speciation are based on the situation in diploid
therefore virtually isolated on apple trees. sexual species. There is still great uncertainty
Attemptswere also made in the laboratory to about speciation in asexual organisms, particu-
produce sympatrically reproductively isolated larly in view ofthe fact that it is so uncertain
species. One or two ofthese experiments, par- whataspeciesisintheseorganisms.Hence,what
ticularlyonebyThoday,haveindeedbeenseem- is speciation? However, we have the fact ofthe
ingly successful (Thoday and Gibson 1962). bdelloidrotifers,ataxonwithsome200-oddspe-
However, there are literally scores ofothersuch cies, all ofthem asexual with no males ever en-
endeavors that led to negative results. countered. This is best explained by the as-
Disruptive selection has been suggested as an sumption that there is a continuing production
effectivemechanismtoproducedifferentspecies. ofasexualclones,andthatsubsequentextinction
If we had within a single population different ofthe less-viable clones leads to gaps between
morphsspecializingindifferentfoodsources,this what ultimately will be called species.
might in due time lead to their isolation as dif-
ferentspecies, itwasbelieved. Suchcasesoftro- Genetics of Speciation
phic polymorphism have indeed been found in
various species of fishes, particularly cichlids In view of the intensive work on speciation
(Meyer 1990) and in a species ofbirds (Smith during the last 100 years, it is shocking to have
1990). However,innoneofthecasesisthereany to admit how little we actually know about the
indication ofa development ofisolating mech- geneticsofspeciation. Thishasanumberofrea-
anisms between the trophic types. Indeed, as a sons. The first is that many geneticists did not
Darwinian,oneshouldexpectthatalineagehav- fully understand the differencebetween phyletic
ingboth trophictypeswould havegreaterinclu- change and multiplication of species. They
sive fitness than one giving up one of the two thought that by extrapolating from the changes
sources offood. This would be true in any vari- withinagenepooltheywouldbeabletoexplain
ableenvironment,andthat ofcoursemeansitis speciation. However, it is now quite clear that
true for all environments. thesolutionwillcomefromacomparisonofdif-
Up to now the best evidence for sympatric ferent populations. Everythingwe know, partic-
speciationisprovidedbyants, wheresometimes ularlythroughthestudyofperipatricspeciation,
MAYR:SPECIESANDSPECIATION 139
indicates that certain genetic turnovers may go ticularlywith respectto the lowerinvertebrates,
on in a founderpopulation that are not encoun- lower plants, fungi, prokaryotes, and also such
teredinanormalspeciespopulation.Thesecond ecological specialistsas pelagicanimals. Finally,
reason forthe slow progress is the heterogeneity the area where there is perhaps still the greatest
ofthe genotype. Owing to the ease ofthe elec- uncertainty is the genetic basis ofthe isolating
trophoresis method, most authors have studied mechanisms. But here again the researches are
the variation ofenzyme genes. However, there not likely to lead to any refutation ofcurrently
is no evidence that this class ofgenes is partic- accepted views. To repeat, the concepts ofspe-
ularly involved in the origin ofisolating mech- ciesandspeciationasdevelopedduringtheevo-
anisms. On the contrary, many of the enzyme lutionary synthesis are likely to endure.
allelesareapparentlyin Kimura'sclassofquasi-
neutral genes. It would certainly be misleading Literature Cited
to base one's explanation of speciation on this
class ofgenes. There is much to suggest that the Buesvcohliuntgieorn,ofAs.oci1a9ll9y0.parSasyimtpiactarnitcs—spHeecrieattiiconhyapnodthreasdeisatainvde
genes for behavioral isolating mechanisms be- their factual background. Z. Zool. Syst. Evolutionsforsch.
long to different classes ofgenes from the genes 28:241-260.
for sterility factors. But very little concrete evi- Bush,G.L.andH.Zwolfer. 1984. Sympatrischeundpara-
dence on these classes ofgenes is available. patrische Artbildung. Z. Zool. Syst. Evolutionsforsch. 22:
211-233.
There is little hope for a valid explanation of Coyne, J. A. 1988. What do we know about speciation?
thegeneticsofspeciationuntilgenesareclassified Nature331:485^86.
into different categories, some ofwhich do and Coyne,J.A., H.A. Orr,andD.J.Futuyma. 1988. Dowe
some ofwhich do not play a role in speciation. needanewspeciesconcept?Syst.Zool. 37(2):190-200.
Furthermore,itisquitelikelythatdifferentkinds DoCbozlhuamnbsikay,UnTi.ver1s9i3t7y.PrGesesn,etNiecswaYnodrkt.he origin ofspecies.
ofgenes are involved in the speciation ofdiffer- Eldredge,N.andS.J.Gould. 1972. Punctuatedequilibria:
entkindsoforganisms. TheworkontheAfrican analternativetophyleticgradualism.Pp.305-332inModels
cichlid fishes indicates that behaviorgenes may inpaleobiology,T.J. M.Schopf,ed.Freeman,Cooper,San
be very important in this taxon, becoming iso- Francisco.
lating mechanisms by way of sexual selection, Encdllieners,.JP.riAn.cet1o9n77U.nivGeerosgirtyapPhriecss,vaPrriiantcieotn,on,spNeceiwatiJoenrseayn.d
and may permitthe evolution ofreproductively Futuyma, D. J. and G. C. Mayer. 1980. Non-allopatric
isolated populations within incredibly short pe- speciationinanimals. Syst.Zool. 29:254-271.
riods, periods ofonly a couple ofhundredyears Goldschmidt, R. B. 1940. Thematerialbasisofevolution.
(Mayr 1988a). On the other hand, the conspe- YaleUniversityPress,NewHaven,Connecticut.
cificityofcertaineasternNorthAmericanplants Haefnfceer,16J.5:113916-9.137S.peciationinAmazonianforestbirds.Sci-
with their representatives in eastern Asia indi- Huxley,J. 1942. Evolution: the modern synthesis. George
cates that in this case an isolation of5, 8, or 10 AllenandUnwin, London.
million years was not sufficient for the origin of Jaenike, J. 1981. Criteria for ascertaining the existence of
efficientisolatingmechanisms.Whereverwehave Kehyo,stK.race1s9.81A.m.SpNeacti.es,11p7a:r8a3p0a-t8r3y4.and the morabine grass-
drastically different rates ofspeciation, one can hoppers.Syst.Zool. 30:425-458.
be reasonably sure that different kinds ofisolat- Mayr, E. 1942. Systematicsand the origin ofspecies. Co-
ingmechanismsandtheirgeneticsareinvolved. lumbiaUniversityPress,NewYork.
. 1947. Ecological factors in speciation. Evolution
1:263-288.
Summary 1948. The bearingofnewsystematicson genetical
problems.Thenatureofspecies.Adv.Genet. 2:205-237.
In conclusion, I think it is legitimate to say 1954. Changeofgeneticenvironmentandevolution.
that the basic picture ofspecies and speciation Pp. 157-180 in Evolution as a process, J. Huxley, A. C.
developedbytaxonomistsduringtheEvolution- Hardy, E. B. Ford,eds. AllenandUnwin, London.
ary Synthesisand presentedin 1942 in mybook versi.ty1P9r6e3s.s,CAanmibmraildgsepe,ciMeasssaancdhuesveotltust.ion. HarvardUni-
Systematics and the Origin ofSpecies did not . 1982. Processofspeciationinanimals. Pp. 1-19 in
have to be changed subsequently. Most of the Mechanisms ofspeciation, C. Barigozzi, ed. Alan R. Liss,
aetvteancktshaoguagihnstthietrehahvaevebebeenenthcolraoruigfihclaytiroenfsutaendd, UNneivwe.rYso1ir9t8ky8.aP.resTso,wCaarmdbraindegwe,phMialsossaocphhuyseotftbsi.ology.Harvard
thedevelopmentofamorepluralisticapproach. . 1988b. Thewhyandhowofspecies.Biol.&Philos.
However, there are still vast uncertainties, par- 3:431-441.
140 PROCEEDINGSOFTHECALIFORNIAACADEMYOFSCIENCES,Vol. 48, No. 6
. 1992a. Alocalfloraandthebiologicalspeciescon- Rosen,D. 1979. Fishesfromtheuplandintermontanebasins
cept.Am.J. Bot. 79(2): 222-238. ofGuatemala. Bull.Am. Mus.Nat. Hist. 162:269-375.
. 1992b. Darwin's Principle ofDivergence. J. Hist. Simpson, G. G. 1944. Tempo and mode in evolution. Co-
Biol. 25:343-359. lumbiaUniversityPress,NewYork.
Mayr, E. and W. Provtne. (eds) 1980. The evolutionary . 1961. Principles of animal taxonomy. Columbia
synthesis. Harvard University Press, Cambridge, Massa- UniversityPress,NewYork.
chusetts. Smith,T.B. 1990. Patternsofmorphologicalandgeographic
Mayr, E. and L. L. Short. 1970. Species taxa ofNorth variation in trophic billmorphs ofthe African finch Pyr-
Americanbirds:acontributiontocomparativesystematics. enestes. Biol.J. Linn. Soc.41:381-414.
NuttallOrn.Club,Publ.No. 9,Cambridge, Massachusetts. Sonneborn,T.M. 1975. TheParameciumAureliacomplex
Meyer.A. 1990. Ecologicalandevolutionaryconsequences offourteensiblingspecies.Trans.Am. Microsc.Soc.94(2):
ofthetrophicpolymorphisminCichlasomacitrinellum(Pi- 155-178.G
sces,Cichlidae).Biol.J. Linn.Soc. 39:279-299. Setbbins, L. 1950. Variation and evolution in plants.
Murray, J. and B. Clarke. 1980. The genus Partula on ColumbiaUniversityPress,NewYork.
Moorea.Speciationinprogress. Proc. R.Soc.Lond.Ser. B. Stresemann, E. 1919. Uber die europaischen Baumlaufer.
Biol.Sci. 211:83-117. Verh.Ornithol.Ges. Bayern. 14:39-74.
Nixon, K. C. andQ. D. Wheeler. 1990. Anamplification Thoday,J.M.andJ. B.Gibson. 1962. Isolationbydisrup-
ofthephylogeneticspeciesconcept. Cladistics6:211-223. tiveselection.Nature 193:1164-1166.
Otte, D. and J. E. Endler. eds. 1989. Speciation and its Thorpe, W. H. 1930. Biological races in insects andallied
consequences.Sinauer, Sunderland, Massachusetts. groups. Biol. Rev. Camb. Philos. Soc. 5:177-212.
Paterson, E. H. 1981. The continuing search for the un- White,M.J.D. 1974. Animalcytologyandevolution.Cam-
known and unknowable: a critique ofcontemporary ideas bridgeUniversityPress,Cambridge, Massachusetts.
onspeciation.S. Afr.J. Sci. 77:113-119. . 1978. Modes ofspeciation. W. H. Freeman, San
Rensch.B. 1947. DieTransspezifischeEvolution.Ferdinand Francisco.
Enke,Stuttgart.
Robson,G. C. andO. W. Richards. 1936. Thevariations
ofanimalsinnature. Longmans,Green, London.
CALIFORNIA ACADEMY OF SCIENCES, 1993
Golden Gate Park
San Francisco, California 94118