TheOriginofSpecies‐Chapter4:NaturalSelectionbyCharlesDarwin

NaturalSelection‐itspowercomparedwithman'sselection‐itspoweroncharactersoftriflingimportance‐itspoweratallagesandonbothsexes‐SexualSelection‐Onthegeneralityofintercrossesbetweenindividualsofthesamespecies‐CircumstancesfavourableandunfavourabletoNaturalSelection,namely,intercrossing,isolation,numberofindividuals‐Slowaction‐ExtinctioncausedbyNaturalSelection‐DivergenceofCharacter,relatedtothediversityofinhabitantsofanysmallarea,andtonaturalisation‐ActionofNaturalSelection,throughDivergenceofCharacterandExtinction,onthedescendantsfromacommonparent‐ExplainstheGroupingofallorganicbeings

Howwillthestruggleforexistence,discussedtoobrieflyinthelastchapter,actinregardtovariation?Cantheprincipleofselection,whichwehaveseenissopotentinthehandsofman,applyinnature?Ithinkweshallseethatitcanactmosteffectually.Letitbeborneinmindinwhatanendlessnumberofstrangepeculiaritiesourdomesticproductions,and,inalesserdegree,thoseundernature,vary;andhowstrongthehereditarytendencyis.Underdomestication,itmaybetrulysaidthatthe,wholeorganisationbecomesinsomedegreeplastic.Letitbeborneinmindhowinfinitelycomplexandclose‐fittingarethemutualrelationsofallorganicbeingstoeachotherandtotheirphysicalconditionsoflife.Canit,then,bethoughtimprobable,seeingthatvariationsusefultomanhaveundoubtedlyoccurred,thatothervariationsusefulinsomewaytoeachbeinginthegreatandcomplexbattleoflife,shouldsometimesoccurinthecourseofthousandsofgenerations?Ifsuchdooccur,canwedoubt(rememberingthatmanymoreindividualsarebornthancanpossiblysurvive)thatindividualshavinganyadvantage,howeverslight,overothers,wouldhavethebestchanceofsurvivingandofprocreatingtheirkind?Ontheotherhand,wemayfeelsurethatanyvariationintheleastdegreeinjuriouswouldberigidlydestroyed.Thispreservationoffavourablevariationsandtherejectionofinjuriousvariations,IcallNaturalSelection.Variationsneitherusefulnorinjuriouswouldnotbeaffectedbynaturalselection,andwouldbeleftafluctuatingelement,asperhapsweseeinthespeciescalledpolymorphic.

Weshallbestunderstandtheprobablecourseofnaturalselectionbytakingthecaseofacountryundergoingsomephysicalchange,forinstance,ofclimate.Theproportionalnumbersofitsinhabitantswouldalmostimmediatelyundergoachange,andsomespeciesmightbecomeextinct.Wemayconclude,fromwhatwehaveseenoftheintimateandcomplexmannerinwhichtheinhabitantsofeachcountryareboundtogether,thatanychangeinthenumericalproportionsofsomeoftheinhabitants,independentlyofthechangeofclimateitself,wouldmostseriouslyaffectmanyoftheothers.Ifthecountrywereopenonitsborders,newformswouldcertainlyimmigrate,andthisalsowouldseriouslydisturbtherelationsofsomeoftheformerinhabitants.Letitberememberedhowpowerfultheinfluenceofasingleintroducedtreeormammalhasbeenshowntobe.Butinthecaseofanisland,orofacountrypartlysurroundedbybarriers,intowhichnewandbetteradaptedformscouldnotfreelyenter,weshouldthenhaveplacesintheeconomyofnaturewhichwouldassuredlybebetterfilledup,ifsomeoftheoriginalinhabitantswereinsomemannermodified;for,hadtheareabeenopentoimmigration,thesesameplaceswouldhavebeenseizedonbyintruders.Insuchcase,everyslightmodification,whichinthecourseofageschancedtoarise,andwhichinanywayfavouredtheindividualsofanyofthespecies,bybetteradaptingthemtotheiralteredconditions,wouldtendtobepreserved;andnaturalselectionwouldthushavefreescopefortheworkofimprovement.

Wehavereasontobelieve,asstatedinthefirstchapter,thatachangeintheconditionsoflife,byspeciallyactingonthereproductivesystem,causesorincreasesvariability;andintheforegoingcasetheconditionsoflifearesupposedtohaveundergoneachange,andthiswouldmanifestlybefavourabletonaturalselection,bygivingabetterchanceofprofitablevariationsoccurring;andunlessprofitablevariationsdooccur,naturalselectioncandonothing.Notthat,asIbelieve,anyextremeamountofvariabilityisnecessary;asmancancertainlyproducegreatresultsbyaddingupinanygivendirectionmereindividualdifferences,socouldNature,butfarmoreeasily,fromhavingincomparablylongertimeatherdisposal.NordoIbelievethatanygreatphysicalchange,asofclimate,oranyunusualdegreeofisolationtocheckimmigration,isactuallynecessarytoproducenewandunoccupiedplacesfornaturalselectiontofillupbymodifyingandimprovingsomeofthevaryinginhabitants.Forasalltheinhabitantsofeachcountryarestrugglingtogetherwithnicelybalancedforces,extremelyslightmodificationsinthestructureorhabitsofoneinhabitantwouldoftengiveitanadvantageoverothers;andstillfurthermodificationsofthesamekindwouldoftenstillfurtherincreasetheadvantage.Nocountrycanbenamedinwhichallthenativeinhabitantsarenowsoperfectlyadaptedtoeachotherandtothephysicalconditionsunderwhichtheylive,thatnoneofthemcouldanyhowbeimproved;forinallcountries,thenativeshavebeensofarconqueredbynaturalisedproductions,thattheyhaveallowedforeignerstotakefirmpossessionoftheland.Andasforeignershavethuseverywherebeatensomeofthenatives,wemaysafelyconcludethatthenativesmighthavebeenmodifiedwithadvantage,soastohavebetterresistedsuchintruders.

Asmancanproduceandcertainlyhasproducedagreatresultbyhismethodicalandunconsciousmeansofselection,whatmaynotnatureeffect?Mancanactonlyonexternalandvisiblecharacters:naturecaresnothingforappearances,exceptinsofarastheymaybeusefultoanybeing.Shecanactoneveryinternalorgan,oneveryshadeofconstitutionaldifference,onthewholemachineryoflife.Manselectsonlyforhisowngood;Natureonlyforthatofthebeingwhichshetends.Everyselectedcharacterisfullyexercisedbyher;andthebeingisplacedunderwell‐suitedconditionsoflife.Mankeepsthenativesofmanyclimatesinthesamecountry;heseldomexerciseseachselectedcharacterinsomepeculiarandfittingmanner;hefeedsalongandashortbeakedpigeononthesamefood;hedoesnotexercisealong‐backedorlong‐leggedquadrupedinanypeculiarmanner;heexposessheepwithlongandshortwooltothesameclimate.Hedoesnotallowthemostvigorousmalestostruggleforthefemales.Hedoesnotrigidlydestroyallinferioranimals,butprotectsduringeachvaryingseason,asfarasliesinhispower,allhisproductions.Heoftenbeginshisselectionbysomehalf‐monstrousform;oratleastbysomemodificationprominentenoughtocatchhiseye,ortobeplainlyusefultohim.Undernature,theslightestdifferenceofstructureorconstitutionmaywellturnthenicely‐balancedscaleinthestruggleforlife,andsobepreserved.Howfleetingarethewishesandeffortsofman!howshorthistime!andconsequentlyhowpoorwillhisproductsbe,comparedwiththoseaccumulatedbynatureduringwholegeologicalperiods.Canwewonder,then,thatnature'sproductionsshouldbefar'truer'incharacterthanman'sproductions;thattheyshouldbeinfinitelybetteradaptedtothemostcomplexconditionsoflife,andshouldplainlybearthestampoffarhigherworkmanship?

Itmaybesaidthatnaturalselectionisdailyandhourlyscrutinising,throughouttheworld,everyvariation,eventheslightest;rejectingthatwhichisbad,preservingandaddingupallthatisgood;silentlyandinsensiblyworking,wheneverandwhereveropportunityoffers,attheimprovementofeachorganicbeinginrelationtoitsorganicandinorganicconditionsoflife.Weseenothingoftheseslowchangesinprogress,untilthehandoftimehasmarkedthelonglapsesofages,andthensoimperfectis

ourviewintolongpastgeologicalages,thatweonlyseethattheformsoflifearenowdifferentfromwhattheyformerlywere.

Althoughnaturalselectioncanactonlythroughandforthegoodofeachbeing,yetcharactersandstructures,whichweareapttoconsiderasofverytriflingimportance,maythusbeactedon.Whenweseeleaf‐eatinginsectsgreen,andbark‐feedersmottled‐grey;thealpineptarmiganwhiteinwinter,thered‐grousethecolourofheather,andtheblack‐grousethatofpeatyearth,wemustbelievethatthesetintsareofservicetothesebirdsandinsectsinpreservingthemfromdanger.Grouse,ifnotdestroyedatsomeperiodoftheirlives,wouldincreaseincountlessnumbers;theyareknowntosufferlargelyfrombirdsofprey;andhawksareguidedbyeyesighttotheirprey,somuchso,thatonpartsoftheContinentpersonsarewarnednottokeepwhitepigeons,asbeingthemostliabletodestruction.HenceIcanseenoreasontodoubtthatnaturalselectionmightbemosteffectiveingivingthepropercolourtoeachkindofgrouse,andinkeepingthatcolour,whenonceacquired,trueandconstant.Noroughtwetothinkthattheoccasionaldestructionofananimalofanyparticularcolourwouldproducelittleeffect:weshouldrememberhowessentialitisinaflockofwhitesheeptodestroyeverylambwiththefaintesttraceofblack.Inplantsthedownonthefruitandthecolourofthefleshareconsideredbybotanistsascharactersofthemosttriflingimportance:yetwehearfromanexcellenthorticulturist,Downing,thatintheUnitedStatessmooth‐skinnedfruitssufferfarmorefromabeetle,acurculio,thanthosewithdown;thatpurpleplumssufferfarmorefromacertaindiseasethanyellowplums;whereasanotherdiseaseattacksyellow‐fleshedpeachesfarmorethanthosewithothercolouredflesh.If,withalltheaidsofart,theseslightdifferencesmakeagreatdifferenceincultivatingtheseveralvarieties,assuredly,inastateofnature,wherethetreeswouldhavetostrugglewithothertreesandwithahostofenemies,suchdifferenceswouldeffectuallysettlewhichvariety,whetherasmoothordowny,ayelloworpurplefleshedfruit,shouldsucceed.

Inlookingatmanysmallpointsofdifferencebetweenspecies,which,asfarasourignorancepermitsustojudge,seemtobequiteunimportant,wemustnotforgetthatclimate,food,&c.,probablyproducesomeslightanddirecteffect.Itis,however,farmorenecessarytobearinmindthattherearemanyunknownlawsofcorrelationofgrowth,which,whenonepartoftheorganisationismodifiedthroughvariation,andthemodificationsareaccumulatedbynaturalselectionforthegoodofthebeing,willcauseothermodifications,oftenofthemostunexpectednature.

Asweseethatthosevariationswhichunderdomesticationappearatanyparticularperiodoflife,tendtoreappearintheoffspringatthesameperiod;forinstance,intheseedsofthemanyvarietiesofourculinaryandagriculturalplants;inthecaterpillarandcocoonstagesofthevarietiesofthesilkworm;intheeggsofpoultry,andinthecolourofthedownoftheirchickens;inthehornsofoursheepandcattlewhennearlyadult;soinastateofnature,naturalselectionwillbeenabledtoactonandmodifyorganicbeingsatanyage,bytheaccumulationofprofitablevariationsatthatage,andbytheirinheritanceatacorrespondingage.Ifitprofitaplanttohaveitsseedsmoreandmorewidelydisseminatedbythewind,Icanseenogreaterdifficultyinthisbeingeffectedthroughnaturalselection,thaninthecotton‐planterincreasingandimprovingbyselectionthedowninthepodsonhiscotton‐trees.Naturalselectionmaymodifyandadaptthelarvaofaninsecttoascoreofcontingencies,whollydifferentfromthosewhichconcernthematureinsect.Thesemodificationswillnodoubtaffect,throughthelawsofcorrelation,thestructureoftheadult;andprobablyinthecaseofthoseinsectswhichliveonlyforafewhours,andwhichneverfeed,alargepartoftheirstructureismerelythecorrelatedresultofsuccessivechangesin

thestructureoftheirlarvae.So,conversely,modificationsintheadultwillprobablyoftenaffectthestructureofthelarva;butinallcasesnaturalselectionwillensurethatmodificationsconsequentonothermodificationsatadifferentperiodoflife,shallnotbeintheleastdegreeinjurious:foriftheybecameso,theywouldcausetheextinctionofthespecies.

Naturalselectionwillmodifythestructureoftheyounginrelationtotheparent,andoftheparentinrelationtotheyoung.Insocialanimalsitwilladaptthestructureofeachindividualforthebenefitofthecommunity;ifeachinconsequenceprofitsbytheselectedchange.Whatnaturalselectioncannotdo,istomodifythestructureofonespecies,withoutgivingitanyadvantage,forthegoodofanotherspecies;andthoughstatementstothiseffectmaybefoundinworksofnaturalhistory,Icannotfindonecasewhichwillbearinvestigation.Astructureusedonlyonceinananimal'swholelife,ifofhighimportancetoit,mightbemodifiedtoanyextentbynaturalselection;forinstance,thegreatjawspossessedbycertaininsects,andusedexclusivelyforopeningthecocoonorthehardtiptothebeakofnestlingbirds,usedforbreakingtheegg.Ithasbeenasserted,thatofthebestshort‐beakedtumbler‐pigeonsmoreperishintheeggthanareabletogetoutofit;sothatfanciersassistintheactofhatching.Now,ifnaturehadtomakethebeakofafull‐grownpigeonveryshortforthebird'sownadvantage,theprocessofmodificationwouldbeveryslow,andtherewouldbesimultaneouslythemostrigorousselectionoftheyoungbirdswithintheegg,whichhadthemostpowerfulandhardestbeaks,forallwithweakbeakswouldinevitablyperish:or,moredelicateandmoreeasilybrokenshellsmightbeselected,thethicknessoftheshellbeingknowntovarylikeeveryotherstructure.

SexualSelection

Inasmuchaspeculiaritiesoftenappearunderdomesticationinonesexandbecomehereditarilyattachedtothatsex,thesamefactprobablyoccursundernature,andifso,naturalselectionwillbeabletomodifyonesexinitsfunctionalrelationstotheothersex,orinrelationtowhollydifferenthabitsoflifeinthetwosexes,asissometimesthecasewithinsects.AndthisleadsmetosayafewwordsonwhatIcallSexualSelection.Thisdepends,notonastruggleforexistence,butonastrugglebetweenthemalesforpossessionofthefemales;theresultisnotdeathtotheunsuccessfulcompetitor,butfewornooffspring.Sexualselectionis,therefore,lessrigorousthannaturalselection.Generally,themostvigorousmales,thosewhicharebestfittedfortheirplacesinnature,willleavemostprogeny.Butinmanycases,victorywilldependnotongeneralvigour,butonhavingspecialweapons,confinedtothemalesex.Ahornlessstagorspurlesscockwouldhaveapoorchanceofleavingoffspring.Sexualselectionbyalwaysallowingthevictortobreedmightsurelygiveindomitablecourage,lengthtothespur,andstrengthtothewingtostrikeinthespurredleg,aswellasthebrutalcock‐fighter,whoknowswellthathecanimprovehisbreedbycarefulselectionofthebestcocks.Howlowinthescaleofnaturethislawofbattledescends,Iknownot;malealligatorshavebeendescribedasfighting,bellowing,andwhirlinground,likeIndiansinawar‐dance,forthepossessionofthefemales;malesalmonshavebeenseenfightingalldaylong;malestag‐beetlesoftenbearwoundsfromthehugemandiblesofothermales.Thewaris,perhaps,severestbetweenthemalesofpolygamousanimals,andtheseseemoftenestprovidedwithspecialweapons.Themalesofcarnivorousanimalsarealreadywellarmed;thoughtothemandtoothers,specialmeansofdefencemaybegiventhroughmeansofsexualselection,asthemanetothelion,theshoulder‐padtotheboar,andthehookedjawtothemalesalmon;fortheshieldmaybeasimportantforvictory,astheswordorspear.

Amongstbirds,thecontestisoftenofamorepeacefulcharacter.Allthosewhohaveattendedtothesubject,believethatthereistheseverestrivalrybetweenthemalesofmanyspeciestoattractbysingingthefemales.Therock‐thrushofGuiana,birdsofparadise,andsomeothers,congregate;andsuccessivemalesdisplaytheirgorgeousplumageandperformstrangeanticsbeforethefemales,whichstandingbyasspectators,atlastchoosethemostattractivepartner.Thosewhohavecloselyattendedtobirdsinconfinementwellknowthattheyoftentakeindividualpreferencesanddislikes:thusSirR.Heronhasdescribedhowonepiedpeacockwaseminentlyattractivetoallhishenbirds.Itmayappearchildishtoattributeanyeffecttosuchapparentlyweakmeans:Icannothereenteronthedetailsnecessarytosupportthisview;butifmancaninashorttimegiveelegantcarriageandbeautytohisbantams,accordingtohisstandardofbeauty,Icanseenogoodreasontodoubtthatfemalebirds,byselecting,duringthousandsofgenerations,themostmelodiousorbeautifulmales,accordingtotheirstandardofbeauty,mightproduceamarkedeffect.Istronglysuspectthatsomewell‐knownlawswithrespecttotheplumageofmaleandfemalebirds,incomparisonwiththeplumageoftheyoung,canbeexplainedontheviewofplumagehavingbeenchieflymodifiedbysexualselection,actingwhenthebirdshavecometothebreedingageorduringthebreedingseason;themodificationsthusproducedbeinginheritedatcorrespondingagesorseasons,eitherbythemalesalone,orbythemalesandfemales;butIhavenotspaceheretoenteronthissubject.

Thusitis,asIbelieve,thatwhenthemalesandfemalesofanyanimalhavethesamegeneralhabitsoflife,butdifferinstructure,colour,orornament,suchdifferenceshavebeenmainlycausedbysexualselection;thatis,individualmaleshavehad,insuccessivegenerations,someslightadvantageoverothermales,intheirweapons,meansofdefence,orcharms;andhavetransmittedtheseadvantagestotheirmaleoffspring.Yet,Iwouldnotwishtoattributeallsuchsexualdifferencestothisagency:forweseepeculiaritiesarisingandbecomingattachedtothemalesexinourdomesticanimals(asthewattleinmalecarriers,horn‐likeprotuberancesinthecocksofcertainfowls,&c.),whichwecannotbelievetobeeitherusefultothemalesinbattle,orattractivetothefemales.Weseeanalogouscasesundernature,forinstance,thetuftofhaironthebreastoftheturkey‐cock,whichcanhardlybeeitherusefulorornamentaltothisbird;indeed,hadthetuftappearedunderdomestication,itwouldhavebeencalledamonstrosity.

IllustrationsoftheactionofNaturalSelection

Inordertomakeitclearhow,asIbelieve,naturalselectionacts,Imustbegpermissiontogiveoneortwoimaginaryillustrations.Letustakethecaseofawolf,whichpreysonvariousanimals,securingsomebycraft,somebystrength,andsomebyfleetness;andletussupposethatthefleetestprey,adeerforinstance,hadfromanychangeinthecountryincreasedinnumbers,orthatotherpreyhaddecreasedinnumbers,duringthatseasonoftheyearwhenthewolfishardestpressedforfood.Icanundersuchcircumstancesseenoreasontodoubtthattheswiftestandslimmestwolveswouldhavethebestchanceofsurviving,andsobepreservedorselected,providedalwaysthattheyretainedstrengthtomastertheirpreyatthisoratsomeotherperiodoftheyear,whentheymightbecompelledtopreyonotheranimals.Icanseenomorereasontodoubtthis,thanthatmancanimprovethefleetnessofhisgreyhoundsbycarefulandmethodicalselection,orbythatunconsciousselectionwhichresultsfromeachmantryingtokeepthebestdogswithoutanythoughtofmodifyingthebreed.

Evenwithoutanychangeintheproportionalnumbersoftheanimalsonwhichourwolfpreyed,acubmightbebornwithaninnatetendencytopursuecertainkindsofprey.Norcanthisbethoughtveryimprobable;forweoftenobservegreatdifferencesinthenaturaltendenciesofourdomesticanimals;onecat,forinstance,takingtocatchrats,anothermice;onecat,accordingtoMr.St.John,bringinghomewingedgame,anotherharesorrabbits,andanotherhuntingonmarshygroundandalmostnightlycatchingwoodcocksorsnipes.Thetendencytocatchratsratherthanmiceisknowntobeinherited.Now,ifanyslightinnatechangeofhabitorofstructurebenefitedanindividualwolf,itwouldhavethebestchanceofsurvivingandofleavingoffspring.Someofitsyoungwouldprobablyinheritthesamehabitsorstructure,andbytherepetitionofthisprocess,anewvarietymightbeformedwhichwouldeithersupplantorcoexistwiththeparent‐formofwolf.Or,again,thewolvesinhabitingamountainousdistrict,andthosefrequentingthelowlands,wouldnaturallybeforcedtohuntdifferentprey;andfromthecontinuedpreservationoftheindividualsbestfittedforthetwosites,twovarietiesmightslowlybeformed.Thesevarietieswouldcrossandblendwheretheymet;buttothissubjectofintercrossingweshallsoonhavetoreturn.Imayadd,that,accordingtoMr.Pierce,therearetwovarietiesofthewolfinhabitingtheCatskillMountainsintheUnitedStates,onewithalightgreyhound‐likeform,whichpursuesdeer,andtheothermorebulky,withshorterlegs,whichmorefrequentlyattackstheshepherd'sflocks.

Letusnowtakeamorecomplexcase.Certainplantsexcreteasweetjuice,apparentlyforthesakeofeliminatingsomethinginjuriousfromtheirsap:thisiseffectedbyglandsatthebaseofthestipulesinsomeLeguminosae,andatthebackoftheleafofthecommonlaurel.Thisjuice,thoughsmallinquantity,isgreedilysoughtbyinsects.Letusnowsupposealittlesweetjuiceornectartobeexcretedbytheinnerbasesofthepetalsofaflower.Inthiscaseinsectsinseekingthenectarwouldgetdustedwithpollen,andwouldcertainlyoftentransportthepollenfromoneflowertothestigmaofanotherflower.Theflowersoftwodistinctindividualsofthesamespecieswouldthusgetcrossed;andtheactofcrossing,wehavegoodreasontobelieve(aswillhereafterbemorefullyalludedto),wouldproduceveryvigorousseedlings,whichconsequentlywouldhavethebestchanceofflourishingandsurviving.Someoftheseseedlingswouldprobablyinheritthenectar‐excretingpower.Thoseinindividualflowerswhichhadthelargestglandsornectaries,andwhichexcretedmostnectar,wouldbeoftenestvisitedbyinsects,andwouldbeoftenestcrossed;andsointhelong‐runwouldgaintheupperhand.Thoseflowers,also,whichhadtheirstamensandpistilsplaced,inrelationtothesizeandhabitsoftheparticularinsectswhichvisitedthem,soastofavourinanydegreethetransportaloftheirpollenfromflowertoflower,wouldlikewisebefavouredorselected.Wemighthavetakenthecaseofinsectsvisitingflowersforthesakeofcollectingpolleninsteadofnectar;andaspollenisformedforthesoleobjectoffertilisation,itsdestructionappearsasimplelosstotheplant;yetifalittlepollenwerecarried,atfirstoccasionallyandthenhabitually,bythepollen‐devouringinsectsfromflowertoflower,andacrossthuseffected,althoughnine‐tenthsofthepollenweredestroyed,itmightstillbeagreatgaintotheplant;andthoseindividualswhichproducedmoreandmorepollen,andhadlargerandlargeranthers,wouldbeselected.

Whenourplant,bythisprocessofthecontinuedpreservationornaturalselectionofmoreandmoreattractiveflowers,hadbeenrenderedhighlyattractivetoinsects,theywould,unintentionallyontheirpart,regularlycarrypollenfromflowertoflower;andthattheycanmosteffectuallydothis,Icouldeasilyshowbymanystrikinginstances.Iwillgiveonlyonenotasaverystrikingcase,butaslikewiseillustratingonestepintheseparationofthesexesofplants,presentlytobealludedto.Someholly‐trees

bearonlymaleflowers,whichhavefourstamensproducingratherasmallquantityofpollen,andarudimentarypistil;otherholly‐treesbearonlyfemaleflowers;thesehaveafull‐sizedpistil,andfourstamenswithshrivelledanthers,inwhichnotagrainofpollencanbedetected.Havingfoundafemaletreeexactlysixtyyardsfromamaletree,Iputthestigmasoftwentyflowers,takenfromdifferentbranches,underthemicroscope,andonall,withoutexception,therewerepollen‐grains,andonsomeaprofusionofpollen.Asthewindhadsetforseveraldaysfromthefemaletothemaletree,thepollencouldnotthushavebeencarried.Theweatherhadbeencoldandboisterous,andthereforenotfavourabletobees,neverthelesseveryfemaleflowerwhichIexaminedhadbeeneffectuallyfertilisedbythebees,accidentallydustedwithpollen,havingflownfromtreetotreeinsearchofnectar.Buttoreturntoourimaginarycase:assoonastheplanthadbeenrenderedsohighlyattractivetoinsectsthatpollenwasregularlycarriedfromflowertoflower,anotherprocessmightcommence.Nonaturalistdoubtstheadvantageofwhathasbeencalledthe'physiologicaldivisionoflabour;'hencewemaybelievethatitwouldbeadvantageoustoaplanttoproducestamensaloneinoneflowerorononewholeplant,andpistilsaloneinanotherfloweroronanotherplant.Inplantsundercultureandplacedundernewconditionsoflife,sometimesthemaleorgansandsometimesthefemaleorgansbecomemoreorlessimpotent;nowifwesupposethistooccurineversoslightadegreeundernature,thenaspollenisalreadycarriedregularlyfromflowertoflower,andasamorecompleteseparationofthesexesofourplantwouldbeadvantageousontheprincipleofthedivisionoflabour,individualswiththistendencymoreandmoreincreased,wouldbecontinuallyfavouredorselected,untilatlastacompleteseparationofthesexeswouldbeeffected.

Letusnowturntothenectar‐feedinginsectsinourimaginarycase:wemaysupposetheplantofwhichwehavebeenslowlyincreasingthenectarbycontinuedselection,tobeacommonplant;andthatcertaininsectsdependedinmainpartonitsnectarforfood.Icouldgivemanyfacts,showinghowanxiousbeesaretosavetime;forinstance,theirhabitofcuttingholesandsuckingthenectaratthebasesofcertainflowers,whichtheycan,withaverylittlemoretrouble,enterbythemouth.Bearingsuchfactsinmind,Icanseenoreasontodoubtthatanaccidentaldeviationinthesizeandformofthebody,orinthecurvatureandlengthoftheproboscis,&c.,fartooslighttobeappreciatedbyus,mightprofitabeeorotherinsect,sothatanindividualsocharacterisedwouldbeabletoobtainitsfoodmorequickly,andsohaveabetterchanceoflivingandleavingdescendants.Itsdescendantswouldprobablyinheritatendencytoasimilarslightdeviationofstructure.Thetubesofthecorollasofthecommonredandincarnateclovers(Trifoliumpratenseandincarnatum)donotonahastyglanceappeartodifferinlength;yetthehive‐beecaneasilysuckthenectaroutoftheincarnateclover,butnotoutofthecommonredclover,whichisvisitedbyhumble‐beesalone;sothatwholefieldsoftheredcloverofferinvainanabundantsupplyofpreciousnectartothehive‐bee.Thusitmightbeagreatadvantagetothehive‐beetohaveaslightlylongerordifferentlyconstructedproboscis.Ontheotherhand,Ihavefoundbyexperimentthatthefertilityofclovergreatlydependsonbeesvisitingandmovingpartsofthecorolla,soastopushthepollenontothestigmaticsurface.Hence,again,ifhumble‐beesweretobecomerareinanycountry,itmightbeagreatadvantagetotheredclovertohaveashorterormoredeeplydividedtubetoitscorolla,sothatthehive‐beecouldvisititsflowers.ThusIcanunderstandhowaflowerandabeemightslowlybecome,eithersimultaneouslyoroneaftertheother,modifiedandadaptedinthemostperfectmannertoeachother,bythecontinuedpreservationofindividualspresentingmutualandslightlyfavourabledeviationsofstructure.

Iamwellawarethatthisdoctrineofnaturalselection,exemplifiedintheaboveimaginaryinstances,isopentothesameobjectionswhichwereatfirsturgedagainstSirCharlesLyell'snobleviewson'the

modernchangesoftheearth,asillustrativeofgeology;'butwenowveryseldomheartheaction,forinstance,ofthecoast‐waves,calledatriflingandinsignificantcause,whenappliedtotheexcavationofgiganticvalleysortotheformationofthelongestlinesofinlandcliffs.Naturalselectioncanactonlybythepreservationandaccumulationofinfinitesimallysmallinheritedmodifications,eachprofitabletothepreservedbeing;andasmoderngeologyhasalmostbanishedsuchviewsastheexcavationofagreatvalleybyasinglediluvialwave,sowillnaturalselection,ifitbeatrueprinciple,banishthebeliefofthecontinuedcreationofneworganicbeings,orofanygreatandsuddenmodificationintheirstructure.

OntheIntercrossingofIndividuals

Imusthereintroduceashortdigression.Inthecaseofanimalsandplantswithseparatedsexes,itisofcourseobviousthattwoindividualsmustalwaysuniteforeachbirth;butinthecaseofhermaphroditesthisisfarfromobvious.NeverthelessIamstronglyinclinedtobelievethatwithallhermaphroditestwoindividuals,eitheroccasionallyorhabitually,concurforthereproductionoftheirkind.Thisview,Imayadd,wasfirstsuggestedbyAndrewKnight.Weshallpresentlyseeitsimportance;butImustheretreatthesubjectwithextremebrevity,thoughIhavethematerialspreparedforanamplediscussion.Allvertebrateanimals,allinsects,andsomeotherlargegroupsofanimals,pairforeachbirth.Modernresearchhasmuchdiminishedthenumberofsupposedhermaphrodites,andofrealhermaphroditesalargenumberpair;thatis,twoindividualsregularlyuniteforreproduction,whichisallthatconcernsus.Butstilltherearemanyhermaphroditeanimalswhichcertainlydonothabituallypair,andavastmajorityofplantsarehermaphrodites.Whatreason,itmaybeasked,isthereforsupposinginthesecasesthattwoindividualseverconcurinreproduction?Asitisimpossibleheretoenterondetails,Imusttrusttosomegeneralconsiderationsalone.

Inthefirstplace,Ihavecollectedsolargeabodyoffacts,showing,inaccordancewiththealmostuniversalbeliefofbreeders,thatwithanimalsandplantsacrossbetweendifferentvarieties,orbetweenindividualsofthesamevarietybutofanotherstrain,givesvigourandfertilitytotheoffspring;andontheotherhand,thatcloseinterbreedingdiminishesvigourandfertility;thatthesefactsaloneinclinemetobelievethatitisagenerallawofnature(utterlyignorantthoughwebeofthemeaningofthelaw)thatnoorganicbeingself‐fertilisesitselfforaneternityofgenerations;butthatacrosswithanotherindividualisoccasionallyperhapsatverylongintervals‐‐indispensable.

Onthebeliefthatthisisalawofnature,wecan,Ithink,understandseverallargeclassesoffacts,suchasthefollowing,whichonanyotherviewareinexplicable.Everyhybridizerknowshowunfavourableexposuretowetistothefertilisationofaflower,yetwhatamultitudeofflowershavetheiranthersandstigmasfullyexposedtotheweather!butifanoccasionalcrossbeindispensable,thefullestfreedomfortheentranceofpollenfromanotherindividualwillexplainthisstateofexposure,moreespeciallyastheplant'sownanthersandpistilgenerallystandsoclosetogetherthatself‐fertilisationseemsalmostinevitable.Manyflowers,ontheotherhand,havetheirorgansoffructificationcloselyenclosed,asinthegreatpapilionaceousorpea‐family;butinseveral,perhapsinall,suchflowers,thereisaverycuriousadaptationbetweenthestructureoftheflowerandthemannerinwhichbeessuckthenectar;for,indoingthis,theyeitherpushtheflower'sownpollenonthestigma,orbringpollenfromanotherflower.

Sonecessaryarethevisitsofbeestopapilionaceousflowers,thatIhavefound,byexperimentspublishedelsewhere,thattheirfertilityisgreatlydiminishedifthesevisitsbeprevented.Now,itisscarcelypossiblethatbeesshouldflyfromflowertoflower,andnotcarrypollenfromonetotheother,tothegreatgood,asIbelieve,oftheplant.Beeswillactlikeacamel‐hairpencil,anditisquitesufficientjusttotouchtheanthersofoneflowerandthenthestigmaofanotherwiththesamebrushtoensurefertilisation;butitmustnotbesupposedthatbeeswouldthusproduceamultitudeofhybridsbetweendistinctspecies;forifyoubringonthesamebrushaplant'sownpollenandpollenfromanotherspecies,theformerwillhavesuchaprepotenteffect,thatitwillinvariablyandcompletelydestroy,ashasbeenshownbyGärtner,anyinfluencefromtheforeignpollen.

Whenthestamensofaflowersuddenlyspringtowardsthepistil,orslowlymoveoneaftertheothertowardsit,thecontrivanceseemsadaptedsolelytoensureself‐fertilisation;andnodoubtitisusefulforthisend:but,theagencyofinsectsisoftenrequiredtocausethestamenstospringforward,asKölreuterhasshowntobethecasewiththebarberry;andcuriouslyinthisverygenus,whichseemstohaveaspecialcontrivanceforself‐fertilisation,itiswellknownthatifveryclosely‐alliedformsorvarietiesareplantedneareachother,itishardlypossibletoraisepureseedlings,solargelydotheynaturallycross.Inmanyothercases,farfromtherebeinganyaidsforself‐fertilisation,therearespecialcontrivances,asIcouldshowfromthewritingsofC.C.Sprengelandfrommyownobservations,whicheffectuallypreventthestigmareceivingpollenfromitsownflower:forinstance,inLobeliafulgens,thereisareallybeautifulandelaboratecontrivancebywhicheveryoneoftheinfinitelynumerouspollen‐granulesaresweptoutoftheconjoinedanthersofeachflower,beforethestigmaofthatindividualflowerisreadytoreceivethem;andasthisflowerisnevervisited,atleastinmygarden,byinsects,itneversetsaseed,thoughbyplacingpollenfromonefloweronthestigmaofanother,Iraisedplentyofseedlings;andwhilstanotherspeciesofLobeliagrowingcloseby,whichisvisitedbybees,seedsfreely.Inverymanyothercases,thoughtherebenospecialmechanicalcontrivancetopreventthestigmaofaflowerreceivingitsownpollen,yet,asC.C.Sprengelhasshown,andasIcanconfirm,eithertheanthersburstbeforethestigmaisreadyforfertilisation,orthestigmaisreadybeforethepollenofthatflowerisready,sothattheseplantshaveinfactseparatedsexes,andmusthabituallybecrossed.Howstrangearethesefacts!Howstrangethatthepollenandstigmaticsurfaceofthesameflower,thoughplacedsoclosetogether,asiffortheverypurposeofself‐fertilisation,shouldinsomanycasesbemutuallyuselesstoeachother!Howsimplyarethesefactsexplainedontheviewofanoccasionalcrosswithadistinctindividualbeingadvantageousorindispensable!

Ifseveralvarietiesofthecabbage,radish,onion,andofsomeotherplants,beallowedtoseedneareachother,alargemajority,asIhavefound,oftheseedlingsthusraisedwillturnoutmongrels:forinstance,Iraised233seedlingcabbagesfromsomeplantsofdifferentvarietiesgrowingneareachother,andoftheseonly78weretruetotheirkind,andsomeevenofthesewerenotperfectlytrue.Yetthepistilofeachcabbage‐flowerissurroundednotonlybyitsownsixstamens,butbythoseofthemanyotherflowersonthesameplant.How,then,comesitthatsuchavastnumberoftheseedlingsaremongrelised?Isuspectthatitmustarisefromthepollenofadistinctvarietyhavingaprepotenteffectoveraflower'sownpollen;andthatthisispartofthegenerallawofgoodbeingderivedfromtheintercrossingofdistinctindividualsofthesamespecies.Whendistinctspeciesarecrossedthecaseisdirectlythereverse,foraplant'sownpollenisalwaysprepotentoverforeignpollen;buttothissubjectweshallreturninafuturechapter.

Inthecaseofagigantictreecoveredwithinnumerableflowers,itmaybeobjectedthatpollencouldseldombecarriedfromtreetotree,andatmostonlyfromflowertofloweronthesametree,andthatflowersonthesametreecanbeconsideredasdistinctindividualsonlyinalimitedsense.Ibelievethisobjectiontobevalid,butthatnaturehaslargelyprovidedagainstitbygivingtotreesastrongtendencytobearflowerswithseparatedsexes.Whenthesexesareseparated,althoughthemaleandfemaleflowersmaybeproducedonthesametree,wecanseethatpollenmustberegularlycarriedfromflowertoflower;andthiswillgiveabetterchanceofpollenbeingoccasionallycarriedfromtreetotree.ThattreesbelongingtoallOrdershavetheirsexesmoreoftenseparatedthanotherplants,Ifindtobethecaseinthiscountry;andatmyrequestDrHookertabulatedthetreesofNewZealand,andDrAsaGraythoseoftheUnitedStates,andtheresultwasasIanticipated.Ontheotherhand,DrHookerhasrecentlyinformedmethathefindsthattheruledoesnotholdinAustralia;andIhavemadethesefewremarksonthesexesoftreessimplytocallattentiontothesubject.

Turningforaverybriefspacetoanimals:onthelandtherearesomehermaphrodites,asland‐molluscaandearth‐worms;buttheseallpair.AsyetIhavenotfoundasinglecaseofaterrestrialanimalwhichfertilisesitself.Wecanunderstandthisremarkablefact,whichofferssostrongacontrastwithterrestrialplants,ontheviewofanoccasionalcrossbeingindispensable,byconsideringthemediuminwhichterrestrialanimalslive,andthenatureofthefertilisingelement;forweknowofnomeans,analogoustotheactionofinsectsandofthewindinthecaseofplants,bywhichanoccasionalcrosscouldbeeffectedwithterrestrialanimalswithouttheconcurrenceoftwoindividuals.Ofaquaticanimals,therearemanyself‐fertilisinghermaphrodites;butherecurrentsinthewaterofferanobviousmeansforanoccasionalcross.And,asinthecaseofflowers,Ihaveasyetfailed,afterconsultationwithoneofthehighestauthorities,namely,ProfessorHuxley,todiscoverasinglecaseofanhermaphroditeanimalwiththeorgansofreproductionsoperfectlyenclosedwithinthebody,thataccessfromwithoutandtheoccasionalinfluenceofadistinctindividualcanbeshowntobephysicallyimpossible.Cirripedeslongappearedtometopresentacaseofverygreatdifficultyunderthispointofview;butIhavebeenenabled,byafortunatechance,elsewheretoprovethattwoindividuals,thoughbothareself‐fertilisinghermaphrodites,dosometimescross.

Itmusthavestruckmostnaturalistsasastrangeanomalythat,inthecaseofbothanimalsandplants,speciesofthesamefamilyandevenofthesamegenus,thoughagreeingcloselywitheachotherinalmosttheirwholeorganisation,yetarenotrarely,someofthemhermaphrodites,andsomeofthemunisexual.Butif,infact,allhermaphroditesdooccasionallyintercrosswithotherindividuals,thedifferencebetweenhermaphroditesandunisexualspecies,asfarasfunctionisconcerned,becomesverysmall.

FromtheseseveralconsiderationsandfromthemanyspecialfactswhichIhavecollected,butwhichIamnothereabletogive,Iamstronglyinclinedtosuspectthat,bothinthevegetableandanimalkingdoms,anoccasionalintercrosswithadistinctindividualisalawofnature.Iamwellawarethatthereare,onthisview,manycasesofdifficulty,someofwhichIamtryingtoinvestigate.Finallythen,wemayconcludethatinmanyorganicbeings,acrossbetweentwoindividualsisanobviousnecessityforeachbirth;inmanyothersitoccursperhapsonlyatlongintervals;butinnone,asIsuspect,canself‐fertilisationgoonforperpetuity.

CircumstancesfavourabletoNaturalSelection

Thisisanextremelyintricatesubject.Alargeamountofinheritableanddiversifiedvariabilityisfavourable,butIbelievemereindividualdifferencessufficeforthework.Alargenumberofindividuals,bygivingabetterchancefortheappearancewithinanygivenperiodofprofitablevariations,willcompensateforalesseramountofvariabilityineachindividual,andis,Ibelieve,anextremelyimportantelementofsuccess.Thoughnaturegrantsvastperiodsoftimefortheworkofnaturalselection,shedoesnotgrantanindefiniteperiod;forasallorganicbeingsarestriving,itmaybesaid,toseizeoneachplaceintheeconomyofnature,ifanyonespeciesdoesnotbecomemodifiedandimprovedinacorrespondingdegreewithitscompetitors,itwillsoonbeexterminated.

Inman'smethodicalselection,abreederselectsforsomedefiniteobject,andfreeintercrossingwillwhollystophiswork.Butwhenmanymen,withoutintendingtoalterthebreed,haveanearlycommonstandardofperfection,andalltrytogetandbreedfromthebestanimals,muchimprovementandmodificationsurelybutslowlyfollowfromthisunconsciousprocessofselection,notwithstandingalargeamountofcrossingwithinferioranimals.Thusitwillbeinnature;forwithinaconfinedarea,withsomeplaceinitspolitynotsoperfectlyoccupiedasmightbe,naturalselectionwillalwaystendtopreservealltheindividualsvaryingintherightdirection,thoughindifferentdegrees,soasbettertofilluptheunoccupiedplace.Butiftheareabelarge,itsseveraldistrictswillalmostcertainlypresentdifferentconditionsoflife;andthenifnaturalselectionbemodifyingandimprovingaspeciesintheseveraldistricts,therewillbeintercrossingwiththeotherindividualsofthesamespeciesontheconfinesofeach.Andinthiscasetheeffectsofintercrossingcanhardlybecounterbalancedbynaturalselectionalwaystendingtomodifyalltheindividualsineachdistrictinexactlythesamemannertotheconditionsofeach;forinacontinuousarea,theconditionswillgenerallygraduateawayinsensiblyfromonedistricttoanother.Theintercrossingwillmostaffectthoseanimalswhichuniteforeachbirth,whichwandermuch,andwhichdonotbreedataveryquickrate.Henceinanimalsofthisnature,forinstanceinbirds,varietieswillgenerallybeconfinedtoseparatedcountries;andthisIbelievetobethecase.Inhermaphroditeorganismswhichcrossonlyoccasionally,andlikewiseinanimalswhichuniteforeachbirth,butwhichwanderlittleandwhichcanincreaseataveryrapidrate,anewandimprovedvarietymightbequicklyformedonanyonespot,andmighttheremaintainitselfinabody,sothatwhateverintercrossingtookplacewouldbechieflybetweentheindividualsofthesamenewvariety.Alocalvarietywhenoncethusformedmightsubsequentlyslowlyspreadtootherdistricts.Ontheaboveprinciple,nurserymenalwaysprefergettingseedfromalargebodyofplantsofthesamevariety,asthechanceofintercrossingwithothervarietiesisthuslessened.

Eveninthecaseofslow‐breedinganimals,whichuniteforeachbirth,wemustnotoverratetheeffectsofintercrossesinretardingnaturalselection;forIcanbringaconsiderablecatalogueoffacts,showingthatwithinthesamearea,varietiesofthesameanimalcanlongremaindistinct,fromhauntingdifferentstations,frombreedingatslightlydifferentseasons,orfromvarietiesofthesamekindpreferringtopairtogether.

Intercrossingplaysaveryimportantpartinnatureinkeepingtheindividualsofthesamespecies,orofthesamevariety,trueanduniformincharacter.Itwillobviouslythusactfarmoreefficientlywiththose

animalswhichuniteforeachbirth;butIhavealreadyattemptedtoshowthatwehavereasontobelievethatoccasionalintercrossestakeplacewithallanimalsandwithallplants.Evenifthesetakeplaceonlyatlongintervals,Iamconvincedthattheyoungthusproducedwillgainsomuchinvigourandfertilityovertheoffspringfromlong‐continuedself‐fertilisation,thattheywillhaveabetterchanceofsurvivingandpropagatingtheirkind;andthus,inthelongrun,theinfluenceofintercrosses,evenatrareintervals,willbegreat.Ifthereexistorganicbeingswhichneverintercross,uniformityofcharactercanberetainedamongstthem,aslongastheirconditionsofliferemainthesame,onlythroughtheprincipleofinheritance,andthroughnaturalselectiondestroyinganywhichdepartfromthepropertype;butiftheirconditionsoflifechangeandtheyundergomodification,uniformityofcharactercanbegiventotheirmodifiedoffspring,solelybynaturalselectionpreservingthesamefavourablevariations.

Isolation,also,isanimportantelementintheprocessofnaturalselection.Inaconfinedorisolatedarea,ifnotverylarge,theorganicandinorganicconditionsoflifewillgenerallybeinagreatdegreeuniform;sothatnaturalselectionwilltendtomodifyalltheindividualsofavaryingspeciesthroughouttheareainthesamemannerinrelationtothesameconditions.Intercrosses,also,withtheindividualsofthesamespecies,whichotherwisewouldhaveinhabitedthesurroundinganddifferentlycircumstanceddistricts,willbeprevented.Butisolationprobablyactsmoreefficientlyincheckingtheimmigrationofbetteradaptedorganisms,afteranyphysicalchange,suchasofclimateorelevationoftheland,&c.;andthusnewplacesinthenaturaleconomyofthecountryareleftopenfortheoldinhabitantstostrugglefor,andbecomeadaptedto,throughmodificationsintheirstructureandconstitution.Lastly,isolation,bycheckingimmigrationandconsequentlycompetition,willgivetimeforanynewvarietytobeslowlyimproved;andthismaysometimesbeofimportanceintheproductionofnewspecies.If,however,anisolatedareabeverysmall,eitherfrombeingsurroundedbybarriers,orfromhavingverypeculiarphysicalconditions,thetotalnumberoftheindividualssupportedonitwillnecessarilybeverysmall;andfewnessofindividualswillgreatlyretardtheproductionofnewspeciesthroughnaturalselection,bydecreasingthechanceoftheappearanceoffavourablevariations.

Ifweturntonaturetotestthetruthoftheseremarks,andlookatanysmallisolatedarea,suchasanoceanicisland,althoughthetotalnumberofthespeciesinhabitingit,willbefoundtobesmall,asweshallseeinourchapterongeographicaldistribution;yetofthesespeciesaverylargeproportionareendemic,thatis,havebeenproducedthere,andnowhereelse.Henceanoceanicislandatfirstsightseemstohavebeenhighlyfavourablefortheproductionofnewspecies.Butwemaythusgreatlydeceiveourselves,fortoascertainwhetherasmallisolatedarea,oralargeopenarealikeacontinent,hasbeenmostfavourablefortheproductionofneworganicforms,weoughttomakethecomparisonwithinequaltimes;andthisweareincapableofdoing.

AlthoughIdonotdoubtthatisolationisofconsiderableimportanceintheproductionofnewspecies,onthewholeIaminclinedtobelievethatlargenessofareaisofmoreimportance,moreespeciallyintheproductionofspecies,whichwillprovecapableofenduringforalongperiod,andofspreadingwidely.Throughoutagreatandopenarea,notonlywilltherebeabetterchanceoffavourablevariationsarisingfromthelargenumberofindividualsofthesamespeciestheresupported,buttheconditionsoflifeareinfinitelycomplexfromthelargenumberofalreadyexistingspecies;andifsomeofthesemanyspeciesbecomemodifiedandimproved,otherswillhavetobeimprovedinacorrespondingdegreeortheywillbeexterminated.Eachnewform,also,assoonasithasbeenmuchimproved,willbeabletospreadovertheopenandcontinuousarea,andwillthuscomeintocompetitionwithmanyothers.Hencemorenew

placeswillbeformed,andthecompetitiontofillthemwillbemoresevere,onalargethanonasmallandisolatedarea.Moreover,greatareas,thoughnowcontinuous,owingtooscillationsoflevel,willoftenhaverecentlyexistedinabrokencondition,sothatthegoodeffectsofisolationwillgenerally,toacertainextent,haveconcurred.Finally,Iconcludethat,althoughsmallisolatedareasprobablyhavebeeninsomerespectshighlyfavourablefortheproductionofnewspecies,yetthatthecourseofmodificationwillgenerallyhavebeenmorerapidonlargeareas;andwhatismoreimportant,thatthenewformsproducedonlargeareas,whichalreadyhavebeenvictoriousovermanycompetitors,willbethosethatwillspreadmostwidely,willgiverisetomostnewvarietiesandspecies,andwillthusplayanimportantpartinthechanginghistoryoftheorganicworld.

Wecan,perhaps,ontheseviews,understandsomefactswhichwillbeagainalludedtoinourchapterongeographicaldistribution;forinstance,thattheproductionsofthesmallercontinentofAustraliahaveformerlyyielded,andapparentlyarenowyielding,beforethoseofthelargerEuropaeo‐Asiaticarea.Thus,also,itisthatcontinentalproductionshaveeverywherebecomesolargelynaturalisedonislands.Onasmallisland,theraceforlifewillhavebeenlesssevere,andtherewillhavebeenlessmodificationandlessextermination.Hence,perhaps,itcomesthatthefloraofMadeira,accordingtoOswaldHeer,resemblestheextincttertiaryfloraofEurope.Allfresh‐waterbasins,takentogether,makeasmallareacomparedwiththatoftheseaoroftheland;and,consequently,thecompetitionbetweenfresh‐waterproductionswillhavebeenlessseverethanelsewhere;newformswillhavebeenmoreslowlyformed,andoldformsmoreslowlyexterminated.AnditisinfreshwaterthatwefindsevengeneraofGanoidfishes,remnantsofaoncepreponderantorder:andinfreshwaterwefindsomeofthemostanomalousformsnowknownintheworld,astheOrnithorhynchusandLepidosiren,which,likefossils,connecttoacertainextentordersnowwidelyseparatedinthenaturalscale.Theseanomalousformsmayalmostbecalledlivingfossils;theyhaveenduredtothepresentday,fromhavinginhabitedaconfinedarea,andfromhavingthusbeenexposedtolessseverecompetition.

Tosumupthecircumstancesfavourableandunfavourabletonaturalselection,asfarastheextremeintricacyofthesubjectpermits.Iconclude,lookingtothefuture,thatforterrestrialproductionsalargecontinentalarea,whichwillprobablyundergomanyoscillationsoflevel,andwhichconsequentlywillexistforlongperiodsinabrokencondition,willbethemostfavourablefortheproductionofmanynewformsoflife,likelytoendurelongandtospreadwidely.Fortheareawillfirsthaveexistedasacontinent,andtheinhabitants,atthisperiodnumerousinindividualsandkinds,willhavebeensubjectedtoveryseverecompetition.Whenconvertedbysubsidenceintolargeseparateislands,therewillstillexistmanyindividualsofthesamespeciesoneachisland:intercrossingontheconfinesoftherangeofeachspecieswillthusbechecked:afterphysicalchangesofanykind,immigrationwillbeprevented,sothatnewplacesinthepolityofeachislandwillhavetobefilledupbymodificationsoftheoldinhabitants;andtimewillbeallowedforthevarietiesineachtobecomewellmodifiedandperfected.When,byrenewedelevation,theislandsshallbere‐convertedintoacontinentalarea,therewillagainbeseverecompetition:themostfavouredorimprovedvarietieswillbeenabledtospread:therewillbemuchextinctionofthelessimprovedforms,andtherelativeproportionalnumbersofthevariousinhabitantsoftherenewedcontinentwillagainbechanged;andagaintherewillbeafairfieldfornaturalselectiontoimprovestillfurthertheinhabitants,andthusproducenewspecies.

Thatnaturalselectionwillalwaysactwithextremeslowness,Ifullyadmit.Itsactiondependsontherebeingplacesinthepolityofnature,whichcanbebetteroccupiedbysomeoftheinhabitantsofthe

countryundergoingmodificationofsomekind.Theexistenceofsuchplaceswilloftendependonphysicalchanges,whicharegenerallyveryslow,andontheimmigrationofbetteradaptedformshavingbeenchecked.Buttheactionofnaturalselectionwillprobablystilloftenerdependonsomeoftheinhabitantsbecomingslowlymodified;themutualrelationsofmanyoftheotherinhabitantsbeingthusdisturbed.Nothingcanbeeffected,unlessfavourablevariationsoccur,andvariationitselfisapparentlyalwaysaveryslowprocess.Theprocesswilloftenbegreatlyretardedbyfreeintercrossing.Manywillexclaimthattheseseveralcausesareamplysufficientwhollytostoptheactionofnaturalselection.Idonotbelieveso.Ontheotherhand,Idobelievethatnaturalselectionwillalwaysactveryslowly,oftenonlyatlongintervalsoftime,andgenerallyononlyaveryfewoftheinhabitantsofthesameregionatthesametime.Ifurtherbelieve,thatthisveryslow,intermittentactionofnaturalselectionaccordsperfectlywellwithwhatgeologytellsusoftherateandmanneratwhichtheinhabitantsofthisworldhavechanged.

Slowthoughtheprocessofselectionmaybe,iffeeblemancandomuchbyhispowersofartificialselection,Icanseenolimittotheamountofchange,tothebeautyandinfinitecomplexityofthecoadaptationsbetweenallorganicbeings,onewithanotherandwiththeirphysicalconditionsoflife,whichmaybeeffectedinthelongcourseoftimebynature'spowerofselection.

Extinction

ThissubjectwillbemorefullydiscussedinourchapteronGeology;butitmustbeherealludedtofrombeingintimatelyconnectedwithnaturalselection.Naturalselectionactssolelythroughthepreservationofvariationsinsomewayadvantageous,whichconsequentlyendure.Butasfromthehighgeometricalpowersofincreaseofallorganicbeings,eachareaisalreadyfullystockedwithinhabitants,itfollowsthataseachselectedandfavouredformincreasesinnumber,sowillthelessfavouredformsdecreaseandbecomerare.Rarity,asgeologytellsus,istheprecursortoextinction.Wecan,also,seethatanyformrepresentedbyfewindividualswill,duringfluctuationsintheseasonsorinthenumberofitsenemies,runagoodchanceofutterextinction.Butwemaygofurtherthanthis;forasnewformsarecontinuallyandslowlybeingproduced,unlesswebelievethatthenumberofspecificformsgoesonperpetuallyandalmostindefinitelyincreasing,numbersinevitablymustbecomeextinct.Thatthenumberofspecificformshasnotindefinitelyincreased,geologyshowsusplainly;andindeedwecanseereasonwhytheyshouldnothavethusincreased,forthenumberofplacesinthepolityofnatureisnotindefinitelygreat,notthatwehaveanymeansofknowingthatanyoneregionhasasyetgotitsmaximumofspecies.probablynoregionisasyetfullystocked,forattheCapeofGoodHope,wheremorespeciesofplantsarecrowdedtogetherthaninanyotherquarteroftheworld,someforeignplantshavebecomenaturalised,withoutcausing,asfarasweknow,theextinctionofanynatives.

Furthermore,thespecieswhicharemostnumerousinindividualswillhavethebestchanceofproducingwithinanygivenperiodfavourablevariations.Wehaveevidenceofthis,inthefactsgiveninthesecondchapter,showingthatitisthecommonspecieswhichaffordthegreatestnumberofrecordedvarieties,orincipientspecies.Hence,rarespecieswillbelessquicklymodifiedorimprovedwithinanygivenperiod,andtheywillconsequentlybebeatenintheraceforlifebythemodifieddescendantsofthecommonerspecies.

FromtheseseveralconsiderationsIthinkitinevitablyfollows,thatasnewspeciesinthecourseoftimeareformedthroughnaturalselection,otherswillbecomerarerandrarer,andfinallyextinct.Theformswhichstandinclosestcompetitionwiththoseundergoingmodificationandimprovement,willnaturallysuffermost.AndwehaveseeninthechapterontheStruggleforExistencethatitisthemostclosely‐alliedforms,varietiesofthesamespecies,andspeciesofthesamegenusorofrelatedgenera,which,fromhavingnearlythesamestructure,constitution,andhabits,generallycomeintotheseverestcompetitionwitheachother.Consequently,eachnewvarietyorspecies,duringtheprogressofitsformation,willgenerallypresshardestonitsnearestkindred,andtendtoexterminatethem.Weseethesameprocessofexterminationamongstourdomesticatedproductions,throughtheselectionofimprovedformsbyman.Manycuriousinstancescouldbegivenshowinghowquicklynewbreedsofcattle,sheep,andotheranimals,andvarietiesofflowers,taketheplaceofolderandinferiorkinds.InYorkshire,itishistoricallyknownthattheancientblackcattleweredisplacedbythelong‐horns,andthatthese'weresweptawaybytheshort‐horns'(Iquotethewordsofanagriculturalwriter)'asifbysomemurderouspestilence.'

DivergenceofCharacter

Theprinciple,whichIhavedesignatedbythisterm,isofhighimportanceonmytheory,andexplains,asIbelieve,severalimportantfacts.Inthefirstplace,varieties,evenstrongly‐markedones,thoughhavingsomewhatofthecharacterofspeciesasisshownbythehopelessdoubtsinmanycaseshowtorankthemyetcertainlydifferfromeachotherfarlessthandogoodanddistinctspecies.Nevertheless,accordingtomyview,varietiesarespeciesintheprocessofformation,orare,asIhavecalledthem,incipientspecies.How,then,doesthelesserdifferencebetweenvarietiesbecomeaugmentedintothegreaterdifferencebetweenspecies?Thatthisdoeshabituallyhappen,wemustinferfrommostoftheinnumerablespeciesthroughoutnaturepresentingwell‐markeddifferences;whereasvarieties,thesupposedprototypesandparentsoffuturewell‐markedspecies,presentslightandill‐defineddifferences.Merechance,aswemaycallit,mightcauseonevarietytodifferinsomecharacterfromitsparents,andtheoffspringofthisvarietyagaintodifferfromitsparentintheverysamecharacterandinagreaterdegree;butthisalonewouldneveraccountforsohabitualandlargeanamountofdifferenceasthatbetweenvarietiesofthesamespeciesandspeciesofthesamegenus.

Ashasalwaysbeenmypractice,letusseeklightonthisheadfromourdomesticproductions.Weshallherefindsomethinganalogous.Afancierisstruckbyapigeonhavingaslightlyshorterbeak;anotherfancierisstruckbyapigeonhavingaratherlongerbeak;andontheacknowledgedprinciplethat'fanciersdonotandwillnotadmireamediumstandard,butlikeextremes,'theybothgoon(ashasactuallyoccurredwithtumbler‐pigeons)choosingandbreedingfrombirdswithlongerandlongerbeaks,orwithshorterandshorterbeaks.Again,wemaysupposethatatanearlyperiodonemanpreferredswifterhorses;anotherstrongerandmorebulkyhorses.Theearlydifferenceswouldbeveryslight;inthecourseoftime,fromthecontinuedselectionofswifterhorsesbysomebreeders,andofstrongeronesbyothers,thedifferenceswouldbecomegreater,andwouldbenotedasformingtwosub‐breeds;finally,afterthelapseofcenturies,thesub‐breedswouldbecomeconvertedintotwowell‐establishedanddistinctbreeds.Asthedifferencesslowlybecomegreater,theinferioranimalswithintermediatecharacters,beingneitherveryswiftnorverystrong,willhavebeenneglected,andwill

havetendedtodisappear.Here,then,weseeinman'sproductionstheactionofwhatmaybecalledtheprincipleofdivergence,causingdifferences,atfirstbarelyappreciable,steadilytoincrease,andthebreedstodivergeincharacterbothfromeachotherandfromtheircommonparent.

Buthow,itmaybeasked,cananyanalogousprincipleapplyinnature?Ibelieveitcananddoesapplymostefficiently,fromthesimplecircumstancethatthemorediversifiedthedescendantsfromanyonespeciesbecomeinstructure,constitution,andhabits,bysomuchwilltheybebetterenabledtoseizeonmanyandwidelydiversifiedplacesinthepolityofnature,andsobeenabledtoincreaseinnumbers.

Wecanclearlyseethisinthecaseofanimalswithsimplehabits.Takethecaseofacarnivorousquadruped,ofwhichthenumberthatcanbesupportedinanycountryhaslongagoarrivedatitsfullaverage.Ifitsnaturalpowersofincreasebeallowedtoact,itcansucceedinincreasing(thecountrynotundergoinganychangeinitsconditions)onlybyitsvaryingdescendantsseizingonplacesatpresentoccupiedbyotheranimals:someofthem,forinstance,beingenabledtofeedonnewkindsofprey,eitherdeadoralive;someinhabitingnewstations,climbingtrees,frequentingwater,andsomeperhapsbecominglesscarnivorous.Themorediversifiedinhabitsandstructurethedescendantsofourcarnivorousanimalbecame,themoreplacestheywouldbeenabledtooccupy.Whatappliestooneanimalwillapplythroughoutalltimetoallanimalsthatis,iftheyvaryforotherwisenaturalselectioncandonothing.Soitwillbewithplants.Ithasbeenexperimentallyproved,thatifaplotofgroundbesownwithseveraldistinctgeneraofgrasses,agreaternumberofplantsandagreaterweightofdryherbagecanthusberaised.Thesamehasbeenfoundtoholdgoodwhenfirstonevarietyandthenseveralmixedvarietiesofwheathavebeensownonequalspacesofground.Hence,ifanyonespeciesofgrassweretogoonvarying,andthosevarietieswerecontinuallyselectedwhichdifferedfromeachotherinatallthesamemannerasdistinctspeciesandgeneraofgrassesdifferfromeachother,agreaternumberofindividualplantsofthisspeciesofgrass,includingitsmodifieddescendants,wouldsucceedinlivingonthesamepieceofground.Andwewellknowthateachspeciesandeachvarietyofgrassisannuallysowingalmostcountlessseeds;andthus,asitmaybesaid,isstrivingitsutmosttoincreaseitsnumbers.Consequently,Icannotdoubtthatinthecourseofmanythousandsofgenerations,themostdistinctvarietiesofanyonespeciesofgrasswouldalwayshavethebestchanceofsucceedingandofincreasinginnumbers,andthusofsupplantingthelessdistinctvarieties;andvarieties,whenrenderedverydistinctfromeachother,taketherankofspecies.

Thetruthoftheprinciple,thatthegreatestamountoflifecanbesupportedbygreatdiversificationofstructure,isseenundermanynaturalcircumstances.Inanextremelysmallarea,especiallyiffreelyopentoimmigration,andwherethecontestbetweenindividualandindividualmustbesevere,wealwaysfindgreatdiversityinitsinhabitants.Forinstance,Ifoundthatapieceofturf,threefeetbyfourinsize,whichhadbeenexposedformanyyearstoexactlythesameconditions,supportedtwentyspeciesofplants,andthesebelongedtoeighteengeneraandtoeightorders,whichshowshowmuchtheseplantsdifferedfromeachother.Soitiswiththeplantsandinsectsonsmallanduniformislets;andsoinsmallpondsoffreshwater.Farmersfindthattheycanraisemostfoodbyarotationofplantsbelongingtothemostdifferentorders:naturefollowswhatmaybecalledasimultaneousrotation.Mostoftheanimalsandplantswhichlivecloseroundanysmallpieceofground,couldliveonit(supposingitnottobeinanywaypeculiarinitsnature),andmaybesaidtobestrivingtotheutmosttolivethere;but,itisseen,thatwheretheycomeintotheclosestcompetitionwitheachother,theadvantagesofdiversificationofstructure,withtheaccompanyingdifferencesofhabitandconstitution,determinethat

theinhabitants,whichthusjostleeachothermostclosely,shall,asageneralrule,belongtowhatwecalldifferentgeneraandorders.

Thesameprincipleisseeninthenaturalisationofplantsthroughman'sagencyinforeignlands.Itmighthavebeenexpectedthattheplantswhichhavesucceededinbecomingnaturalisedinanylandwouldgenerallyhavebeencloselyalliedtotheindigenes;forthesearecommonlylookedatasspeciallycreatedandadaptedfortheirowncountry.Itmight,also,perhapshavebeenexpectedthatnaturalisedplantswouldhavebelongedtoafewgroupsmoreespeciallyadaptedtocertainstationsintheirnewhomes.Butthecaseisverydifferent;andAlph.DeCandollehaswellremarkedinhisgreatandadmirablework,thatflorasgainbynaturalisation,proportionallywiththenumberofthenativegeneraandspecies,farmoreinnewgenerathaninnewspecies.Togiveasingleinstance:inthelasteditionofDrAsaGray's'ManualoftheFloraoftheNorthernUnitedStates,'260naturalisedplantsareenumerated,andthesebelongto162genera.Wethusseethatthesenaturalisedplantsareofahighlydiversifiednature.Theydiffer,moreover,toalargeextentfromtheindigenes,foroutofthe162genera,nolessthan100generaarenotthereindigenous,andthusalargeproportionaladditionismadetothegeneraoftheseStates.

Byconsideringthenatureoftheplantsoranimalswhichhavestruggledsuccessfullywiththeindigenesofanycountry,andhavetherebecomenaturalised,wecangainsomecrudeideainwhatmannersomeofthenativeswouldhavehadtobemodified,inordertohavegainedanadvantageovertheothernatives;andwemay,Ithink,atleastsafelyinferthatdiversificationofstructure,amountingtonewgenericdifferences,wouldhavebeenprofitabletothem.

Theadvantageofdiversificationintheinhabitantsofthesameregionis,infact,thesameasthatofthephysiologicaldivisionoflabourintheorgansofthesameindividualbodyasubjectsowellelucidatedbyMilneEdwards.Nophysiologistdoubtsthatastomachbybeingadaptedtodigestvegetablematteralone,orfleshalone,drawsmostnutrimentfromthesesubstances.Sointhegeneraleconomyofanyland,themorewidelyandperfectlytheanimalsandplantsarediversifiedfordifferenthabitsoflife,sowillagreaternumberofindividualsbecapableoftheresupportingthemselves.Asetofanimals,withtheirorganisationbutlittlediversified,couldhardlycompetewithasetmoreperfectlydiversifiedinstructure.Itmaybedoubted,forinstance,whethertheAustralianmarsupials,whicharedividedintogroupsdifferingbutlittlefromeachother,andfeeblyrepresenting,asMrWaterhouseandothershaveremarked,ourcarnivorous,ruminant,androdentmammals,couldsuccessfullycompetewiththesewell‐pronouncedorders.IntheAustralianmammals,weseetheprocessofdiversificationinanearlyandincompletestageofdevelopment.

Aftertheforegoingdiscussion,whichoughttohavebeenmuchamplified,wemay,Ithink,assumethatthemodifieddescendantsofanyonespecieswillsucceedbysomuchthebetterastheybecomemorediversifiedinstructure,andarethusenabledtoencroachonplacesoccupiedbyotherbeings.Nowletusseehowthisprincipleofgreatbenefitbeingderivedfromdivergenceofcharacter,combinedwiththeprinciplesofnaturalselectionandofextinction,willtendtoact.

Theaccompanyingdiagramwillaidusinunderstandingthisratherperplexingsubject.LetAtoLrepresentthespeciesofagenuslargeinitsowncountry;thesespeciesaresupposedtoresembleeachotherinunequaldegrees,asissogenerallythecaseinnature,andasisrepresentedinthediagrambythelettersstandingatunequaldistances.Ihavesaidalargegenus,becausewehaveseeninthesecondchapter,thatonanaveragemoreofthespeciesoflargegeneravarythanofsmallgenera;andthevaryingspeciesofthelargegenerapresentagreaternumberofvarieties.Wehave,also,seenthatthespecies,whicharethecommonestandthemostwidely‐diffused,varymorethanrarespecieswithrestrictedranges.Let(A)beacommon,widely‐diffused,andvaryingspecies,belongingtoagenuslargeinitsowncountry.Thelittlefanofdivergingdottedlinesofunequallengthsproceedingfrom(A),mayrepresentitsvaryingoffspring.Thevariationsaresupposedtobeextremelyslight,butofthemostdiversifiednature;theyarenotsupposedalltoappearsimultaneously,butoftenafterlongintervalsoftime;noraretheyallsupposedtoendureforequalperiods.Onlythosevariationswhichareinsomewayprofitablewillbepreservedornaturallyselected.Andheretheimportanceoftheprincipleofbenefitbeingderivedfromdivergenceofcharactercomesin;forthiswillgenerallyleadtothemostdifferentordivergentvariations(representedbytheouterdottedlines)beingpreservedandaccumulatedbynaturalselection.Whenadottedlinereachesoneofthehorizontallines,andistheremarkedbyasmallnumberedletter,asufficientamountofvariationissupposedtohavebeenaccumulatedtohaveformedafairlywell‐markedvariety,suchaswouldbethoughtworthyofrecordinasystematicwork.

Theintervalsbetweenthehorizontallinesinthediagram,mayrepresenteachathousandgenerations;butitwouldhavebeenbetterifeachhadrepresentedtenthousandgenerations.Afterathousandgenerations,species(A)issupposedtohaveproducedtwofairlywell‐markedvarieties,namelya1andm1.Thesetwovarietieswillgenerallycontinuetobeexposedtothesameconditionswhichmadetheirparentsvariable,andthetendencytovariabilityisinitselfhereditary,consequentlytheywilltendtovary,andgenerallytovaryinnearlythesamemannerastheirparentsvaried.Moreover,thesetwovarieties,beingonlyslightlymodifiedforms,willtendtoinheritthoseadvantageswhichmadetheircommonparent(A)morenumerousthanmostoftheotherinhabitantsofthesamecountry;theywilllikewisepartakeofthosemoregeneraladvantageswhichmadethegenustowhichtheparent‐speciesbelonged,alargegenusinitsowncountry.Andthesecircumstancesweknowtobefavourabletotheproductionofnewvarieties.

If,then,thesetwovarietiesbevariable,themostdivergentoftheirvariationswillgenerallybepreservedduringthenextthousandgenerations.Andafterthisinterval,varietya1issupposedinthediagramtohaveproducedvarietya2,whichwill,owingtotheprincipleofdivergence,differmorefrom(A)thandidvarietya1.Varietym1issupposedtohaveproducedtwovarieties,namelym2ands2,differingfromeachother,andmoreconsiderablyfromtheircommonparent(A).Wemaycontinuetheprocessbysimilarstepsforanylengthoftime;someofthevarieties,aftereachthousandgenerations,producingonlyasinglevariety,butinamoreandmoremodifiedcondition,someproducingtwoorthreevarieties,andsomefailingtoproduceany.Thusthevarietiesormodifieddescendants,proceedingfromthecommonparent(A),willgenerallygoonincreasinginnumberanddivergingincharacter.Inthediagramtheprocessisrepresenteduptotheten‐thousandthgeneration,andunderacondensedandsimplifiedformuptothefourteen‐thousandthgeneration.

ButImusthereremarkthatIdonotsupposethattheprocessevergoesonsoregularlyasisrepresentedinthediagram,thoughinitselfmadesomewhatirregular.Iamfarfromthinkingthatthemostdivergentvarietieswillinvariablyprevailandmultiply:amediumformmayoftenlongendure,andmayormaynotproducemorethanonemodifieddescendant;fornaturalselectionwillalwaysactaccordingtothenatureoftheplaceswhichareeitherunoccupiedornotperfectlyoccupiedbyotherbeings;andthiswilldependoninfinitelycomplexrelations.Butasageneralrule,themorediversifiedinstructurethedescendantsfromanyonespeciescanberendered,themoreplacestheywillbeenabledtoseizeon,andthemoretheirmodifiedprogenywillbeincreased.Inourdiagramthelineofsuccessionisbrokenatregularintervalsbysmallnumberedlettersmarkingthesuccessiveformswhichhavebecomesufficientlydistincttoberecordedasvarieties.Butthesebreaksareimaginary,andmighthavebeeninsertedanywhere,afterintervalslongenoughtohaveallowedtheaccumulationofaconsiderableamountofdivergentvariation.

Asallthemodifieddescendantsfromacommonandwidely‐diffusedspecies,belongingtoalargegenus,willtendtopartakeofthesameadvantageswhichmadetheirparentsuccessfulinlife,theywillgenerallygoonmultiplyinginnumberaswellasdivergingincharacter:thisisrepresentedinthediagrambytheseveraldivergentbranchesproceedingfrom(A).Themodifiedoffspringfromthelaterandmorehighlyimprovedbranchesinthelinesofdescent,will,itisprobable,oftentaketheplaceof,andsodestroy,theearlierandlessimprovedbranches:thisisrepresentedinthediagrambysomeofthelowerbranchesnotreachingtotheupperhorizontallines.InsomecasesIdonotdoubtthattheprocessofmodificationwillbeconfinedtoasinglelineofdescent,andthenumberofthedescendantswillnotbeincreased;althoughtheamountofdivergentmodificationmayhavebeenincreasedinthesuccessivegenerations.Thiscasewouldberepresentedinthediagram,ifallthelinesproceedingfrom(A)wereremoved,exceptingthatfroma1toa10Inthesameway,forinstance,theEnglishrace‐horseandEnglishpointerhaveapparentlybothgoneonslowlydivergingincharacterfromtheiroriginalstocks,withouteitherhavinggivenoffanyfreshbranchesorraces.

Aftertenthousandgenerations,species(A)issupposedtohaveproducedthreeforms,a10,f10,andm10,which,fromhavingdivergedincharacterduringthesuccessivegenerations,willhavecometodifferlargely,butperhapsunequally,fromeachotherandfromtheircommonparent.Ifwesupposetheamountofchangebetweeneachhorizontallineinourdiagramtobeexcessivelysmall,thesethreeformsmaystillbeonlywell‐markedvarieties;ortheymayhavearrivedatthedoubtfulcategoryofsub‐species;butwehaveonlytosupposethestepsintheprocessofmodificationtobemorenumerousorgreaterinamount,toconvertthesethreeformsintowell‐definedspecies:thusthediagramillustratesthestepsbywhichthesmalldifferencesdistinguishingvarietiesareincreasedintothelargerdifferencesdistinguishingspecies.Bycontinuingthesameprocessforagreaternumberofgenerations(asshowninthediagraminacondensedandsimplifiedmanner),wegeteightspecies,markedbythelettersbetweena14andm14,alldescendedfrom(A).Thus,asIbelieve,speciesaremultipliedandgeneraareformed.

Inalargegenusitisprobablethatmorethanonespecieswouldvary.InthediagramIhaveassumedthatasecondspecies(I)hasproduced,byanalogoussteps,aftertenthousandgenerations,eithertwowell‐markedvarieties(w10andz10)ortwospecies,accordingtotheamountofchangesupposedtoberepresentedbetweenthehorizontallines.Afterfourteenthousandgenerations,sixnewspecies,markedbythelettersn14toz14,aresupposedtohavebeenproduced.Ineachgenus,thespecies,whichare

alreadyextremelydifferentincharacter,willgenerallytendtoproducethegreatestnumberofmodifieddescendants;forthesewillhavethebestchanceoffillingnewandwidelydifferentplacesinthepolityofnature:henceinthediagramIhavechosentheextremespecies(A),andthenearlyextremespecies(I),asthosewhichhavelargelyvaried,andhavegivenrisetonewvarietiesandspecies.Theotherninespecies(markedbycapitalletters)ofouroriginalgenus,mayforalongperiodcontinuetransmittingunaltereddescendants;andthisisshowninthediagrambythedottedlinesnotprolongedfarupwardsfromwantofspace.

Butduringtheprocessofmodification,representedinthediagram,anotherofourprinciples,namelythatofextinction,willhaveplayedanimportantpart.Asineachfullystockedcountrynaturalselectionnecessarilyactsbytheselectedformhavingsomeadvantageinthestruggleforlifeoverotherforms,therewillbeaconstanttendencyintheimproveddescendantsofanyonespeciestosupplantandexterminateineachstageofdescenttheirpredecessorsandtheiroriginalparent.Foritshouldberememberedthatthecompetitionwillgenerallybemostseverebetweenthoseformswhicharemostnearlyrelatedtoeachotherinhabits,constitution,andstructure.Hencealltheintermediateformsbetweentheearlierandlaterstates,thatisbetweenthelessandmoreimprovedstateofaspecies,aswellastheoriginalparent‐speciesitself,willgenerallytendtobecomeextinct.Soitprobablywillbewithmanywholecollaterallinesofdescent,whichwillbeconqueredbylaterandimprovedlinesofdescent.If,however,themodifiedoffspringofaspeciesgetintosomedistinctcountry,orbecomequicklyadaptedtosomequitenewstation,inwhichchildandparentdonotcomeintocompetition,bothmaycontinuetoexist.

Ifthenourdiagrambeassumedtorepresentaconsiderableamountofmodification,species(A)andalltheearliervarietieswillhavebecomeextinct,havingbeenreplacedbyeightnewspecies(a14tom14);and(I)willhavebeenreplacedbysix(n14toz14)newspecies.

Butwemaygofurtherthanthis.Theoriginalspeciesofourgenusweresupposedtoresembleeachotherinunequaldegrees,asissogenerallythecaseinnature;species(A)beingmorenearlyrelatedtoB,C,andD,thantotheotherspecies;andspecies(I)moretoG,H,K,L,thantotheothers.Thesetwospecies(A)and(I),werealsosupposedtobeverycommonandwidelydiffusedspecies,sothattheymustoriginallyhavehadsomeadvantageovermostoftheotherspeciesofthegenus.Theirmodifieddescendants,fourteeninnumberatthefourteen‐thousandthgeneration,willprobablyhaveinheritedsomeofthesameadvantages:theyhavealsobeenmodifiedandimprovedinadiversifiedmannerateachstageofdescent,soastohavebecomeadaptedtomanyrelatedplacesinthenaturaleconomyoftheircountry.Itseems,therefore,tomeextremelyprobablethattheywillhavetakentheplacesof,andthusexterminated,notonlytheirparents(A)and(I),butlikewisesomeoftheoriginalspecieswhichweremostnearlyrelatedtotheirparents.Henceveryfewoftheoriginalspecieswillhavetransmittedoffspringtothefourteen‐thousandthgeneration.Wemaysupposethatonlyone(F),ofthetwospecieswhichwereleastcloselyrelatedtotheothernineoriginalspecies,hastransmitteddescendantstothislatestageofdescent.

Thenewspeciesinourdiagramdescendedfromtheoriginalelevenspecies,willnowbefifteeninnumber.Owingtothedivergenttendencyofnaturalselection,theextremeamountofdifferencein

characterbetweenspeciesa14andz14willbemuchgreaterthanthatbetweenthemostdifferentoftheoriginalelevenspecies.Thenewspecies,moreover,willbealliedtoeachotherinawidelydifferentmanner.Oftheeightdescendantsfrom(A)thethreemarkeda14,q14,p14,willbenearlyrelatedfromhavingrecentlybranchedofffroma14;b14andf14,fromhavingdivergedatanearlierperiodfroma5,willbeinsomedegreedistinctfromthethreefirst‐namedspecies;andlastly,o14,e14,andm14,willbenearlyrelatedonetotheother,butfromhavingdivergedatthefirstcommencementoftheprocessofmodification,willbewidelydifferentfromtheotherfivespecies,andmayconstituteasub‐genusorevenadistinctgenus.Thesixdescendantsfrom(I)willformtwosub‐generaorevengenera.Butastheoriginalspecies(I)differedlargelyfrom(A),standingnearlyattheextremepointsoftheoriginalgenus,thesixdescendantsfrom(I)will,owingtoinheritance,differconsiderablyfromtheeightdescendantsfrom(A);thetwogroups,moreover,aresupposedtohavegoneondivergingindifferentdirections.Theintermediatespecies,also(andthisisaveryimportantconsideration),whichconnectedtheoriginalspecies(A)and(I),haveallbecome,excepting(F),extinct,andhaveleftnodescendants.Hencethesixnewspeciesdescendedfrom(I),andtheeightdescendedfrom(A),willhavetoberankedasverydistinctgenera,orevenasdistinctsub‐families.

Thusitis,asIbelieve,thattwoormoregeneraareproducedbydescent,withmodification,fromtwoormorespeciesofthesamegenus.Andthetwoormoreparent‐speciesaresupposedtohavedescendedfromsomeonespeciesofanearliergenus.Inourdiagram,thisisindicatedbythebrokenlines,beneaththecapitalletters,converginginsub‐branchesdownwardstowardsasinglepoint;thispointrepresentingasinglespecies,thesupposedsingleparentofourseveralnewsub‐generaandgenera.

ItisworthwhiletoreflectforamomentonthecharacterofthenewspeciesF14,whichissupposednottohavedivergedmuchincharacter,buttohaveretainedtheformof(F),eitherunalteredoralteredonlyinaslightdegree.Inthiscase,itsaffinitiestotheotherfourteennewspecieswillbeofacuriousandcircuitousnature.Havingdescendedfromaformwhichstoodbetweenthetwoparent‐species(A)and(I),nowsupposedtobeextinctandunknown,itwillbeinsomedegreeintermediateincharacterbetweenthetwogroupsdescendedfromthesespecies.Butasthesetwogroupshavegoneondivergingincharacterfromthetypeoftheirparents,thenewspecies(F14)willnotbedirectlyintermediatebetweenthem,butratherbetweentypesofthetwogroups;andeverynaturalistwillbeabletobringsomesuchcasebeforehismind.

Inthediagram,eachhorizontallinehashithertobeensupposedtorepresentathousandgenerations,buteachmayrepresentamillionorhundredmilliongenerations,andlikewiseasectionofthesuccessivestrataoftheearth'scrustincludingextinctremains.Weshall,whenwecometoourchapteronGeology,havetoreferagaintothissubject,andIthinkweshallthenseethatthediagramthrowslightontheaffinitiesofextinctbeings,which,thoughgenerallybelongingtothesameorders,orfamilies,orgenera,withthosenowliving,yetareoften,insomedegree,intermediateincharacterbetweenexistinggroups;andwecanunderstandthisfact,fortheextinctspecieslivedatveryancientepochswhenthebranchinglinesofdescenthaddivergedless.

Iseenoreasontolimittheprocessofmodification,asnowexplained,totheformationofgeneraalone.If,inourdiagram,wesupposetheamountofchangerepresentedbyeachsuccessivegroupofdiverging

dottedlinestobeverygreat,theformsmarkeda214top14,thosemarkedb14andf14,andthosemarkedo14tom14,willformthreeverydistinctgenera.Weshallalsohavetwoverydistinctgeneradescendedfrom(I)andastheselattertwogenera,bothfromcontinueddivergenceofcharacterandfrominheritancefromadifferentparent,willdifferwidelyfromthethreegeneradescendedfrom(A),thetwolittlegroupsofgenerawillformtwodistinctfamilies,orevenorders,accordingtotheamountofdivergentmodificationsupposedtoberepresentedinthediagram.Andthetwonewfamilies,ororders,willhavedescendedfromtwospeciesoftheoriginalgenus;andthesetwospeciesaresupposedtohavedescendedfromonespeciesofastillmoreancientandunknowngenus.

Wehaveseenthatineachcountryitisthespeciesofthelargergenerawhichoftenestpresentvarietiesorincipientspecies.This,indeed,mighthavebeenexpected;forasnaturalselectionactsthroughoneformhavingsomeadvantageoverotherformsinthestruggleforexistence,itwillchieflyactonthosewhichalreadyhavesomeadvantage;andthelargenessofanygroupshowsthatitsspecieshaveinheritedfromacommonancestorsomeadvantageincommon.Hence,thestrugglefortheproductionofnewandmodifieddescendants,willmainlyliebetweenthelargergroups,whicharealltryingtoincreaseinnumber.Onelargegroupwillslowlyconqueranotherlargegroup,reduceitsnumbers,andthuslessenitschanceoffurthervariationandimprovement.Withinthesamelargegroup,thelaterandmorehighlyperfectedsub‐groups,frombranchingoutandseizingonmanynewplacesinthepolityofNature,willconstantlytendtosupplantanddestroytheearlierandlessimprovedsub‐groups.Smallandbrokengroupsandsub‐groupswillfinallytendtodisappear.Lookingtothefuture,wecanpredictthatthegroupsoforganicbeingswhicharenowlargeandtriumphant,andwhichareleastbrokenup,thatis,whichasyethavesufferedleastextinction,willforalongperiodcontinuetoincrease.Butwhichgroupswillultimatelyprevail,nomancanpredict;forwewellknowthatmanygroups,formerlymostextensivelydeveloped,havenowbecomeextinct.Lookingstillmoreremotelytothefuture,wemaypredictthat,owingtothecontinuedandsteadyincreaseofthelargergroups,amultitudeofsmallergroupswillbecomeutterlyextinct,andleavenomodifieddescendants;andconsequentlythatofthespecieslivingatanyoneperiod,extremelyfewwilltransmitdescendantstoaremotefuturity.IshallhavetoreturntothissubjectinthechapteronClassification,butImayaddthatonthisviewofextremelyfewofthemoreancientspecieshavingtransmitteddescendants,andontheviewofallthedescendantsofthesamespeciesmakingaclass,wecanunderstandhowitisthatthereexistbutveryfewclassesineachmaindivisionoftheanimalandvegetablekingdoms.Althoughextremelyfewofthemostancientspeciesmaynowhavelivingandmodifieddescendants,yetatthemostremotegeologicalperiod,theearthmayhavebeenaswellpeopledwithmanyspeciesofmanygenera,families,orders,andclasses,asatthepresentday.

SummaryofChapter

Ifduringthelongcourseofagesandundervaryingconditionsoflife,organicbeingsvaryatallintheseveralpartsoftheirorganisation,andIthinkthiscannotbedisputed;iftherebe,owingtothehighgeometricalpowersofincreaseofeachspecies,atsomeage,season,oryear,aseverestruggleforlife,andthiscertainlycannotbedisputed;then,consideringtheinfinitecomplexityoftherelationsofallorganicbeingstoeachotherandtotheirconditionsofexistence,causinganinfinitediversityinstructure,constitution,andhabits,tobeadvantageoustothem,Ithinkitwouldbeamostextraordinaryfactifnovariationeverhadoccurredusefultoeachbeing'sownwelfare,inthesamewayassomanyvariationshaveoccurredusefultoman.Butifvariationsusefultoanyorganicbeingdooccur,assuredly

individualsthuscharacterisedwillhavethebestchanceofbeingpreservedinthestruggleforlife;andfromthestrongprincipleofinheritancetheywilltendtoproduceoffspringsimilarlycharacterised.Thisprincipleofpreservation,Ihavecalled,forthesakeofbrevity,NaturalSelection.Naturalselection,ontheprincipleofqualitiesbeinginheritedatcorrespondingages,canmodifytheegg,seed,oryoung,aseasilyastheadult.Amongstmanyanimals,sexualselectionwillgiveitsaidtoordinaryselection,byassuringtothemostvigorousandbestadaptedmalesthegreatestnumberofoffspring.Sexualselectionwillalsogivecharactersusefultothemalesalone,intheirstruggleswithothermales.

Whethernaturalselectionhasreallythusactedinnature,inmodifyingandadaptingthevariousformsoflifetotheirseveralconditionsandstations,mustbejudgedofbythegeneraltenourandbalanceofevidencegiveninthefollowingchapters.Butwealreadyseehowitentailsextinction;andhowlargelyextinctionhasactedintheworld'shistory,geologyplainlydeclares.Naturalselection,also,leadstodivergenceofcharacter;formorelivingbeingscanbesupportedonthesameareathemoretheydivergeinstructure,habits,andconstitution,ofwhichweseeproofbylookingattheinhabitantsofanysmallspotoratnaturalisedproductions.Thereforeduringthemodificationofthedescendantsofanyonespecies,andduringtheincessantstruggleofallspeciestoincreaseinnumbers,themorediversifiedthesedescendantsbecome,thebetterwillbetheirchanceofsucceedinginthebattleoflife.Thusthesmalldifferencesdistinguishingvarietiesofthesamespecies,willsteadilytendtoincreasetilltheycometoequalthegreaterdifferencesbetweenspeciesofthesamegenus,orevenofdistinctgenera.

Wehaveseenthatitisthecommon,thewidely‐diffused,andwidely‐rangingspecies,belongingtothelargergenera,whichvarymost;andthesewilltendtotransmittotheirmodifiedoffspringthatsuperioritywhichnowmakesthemdominantintheirowncountries.Naturalselection,ashasjustbeenremarked,leadstodivergenceofcharacterandtomuchextinctionofthelessimprovedandintermediateformsoflife.Ontheseprinciples,Ibelieve,thenatureoftheaffinitiesofallorganicbeingsmaybeexplained.Itisatrulywonderfulfactthewonderofwhichweareapttooverlookfromfamiliaritythatallanimalsandallplantsthroughoutalltimeandspaceshouldberelatedtoeachotheringroupsubordinatetogroup,inthemannerwhichweeverywherebeholdnamely,varietiesofthesamespeciesmostcloselyrelatedtogether,speciesofthesamegenuslesscloselyandunequallyrelatedtogether,formingsectionsandsub‐genera,speciesofdistinctgeneramuchlesscloselyrelated,andgenerarelatedindifferentdegrees,formingsub‐families,families,orders,sub‐classes,andclasses.Theseveralsubordinategroupsinanyclasscannotberankedinasinglefile,butseemrathertobeclusteredroundpoints,andtheseroundotherpoints,andsooninalmostendlesscycles.Ontheviewthateachspecieshasbeenindependentlycreated,Icanseenoexplanationofthisgreatfactintheclassificationofallorganicbeings;but,tothebestofmyjudgment,itisexplainedthroughinheritanceandthecomplexactionofnaturalselection,entailingextinctionanddivergenceofcharacter,aswehaveseenillustratedinthediagram.

Theaffinitiesofallthebeingsofthesameclasshavesometimesbeenrepresentedbyagreattree.Ibelievethissimilelargelyspeaksthetruth.Thegreenandbuddingtwigsmayrepresentexistingspecies;andthoseproducedduringeachformeryearmayrepresentthelongsuccessionofextinctspecies.Ateachperiodofgrowthallthegrowingtwigshavetriedtobranchoutonallsides,andtoovertopandkillthesurroundingtwigsandbranches,inthesamemannerasspeciesandgroupsofspecieshavetriedtoovermasterotherspeciesinthegreatbattleforlife.Thelimbsdividedintogreatbranches,andtheseintolesserandlesserbranches,werethemselvesonce,whenthetreewassmall,

buddingtwigs;andthisconnexionoftheformerandpresentbudsbyramifyingbranchesmaywellrepresenttheclassificationofallextinctandlivingspeciesingroupssubordinatetogroups.Ofthemanytwigswhichflourishedwhenthetreewasamerebush,onlytwoorthree,nowgrownintogreatbranches,yetsurviveandbearalltheotherbranches;sowiththespecieswhichlivedduringlong‐pastgeologicalperiods,veryfewnowhavelivingandmodifieddescendants.Fromthefirstgrowthofthetree,manyalimbandbranchhasdecayedanddroppedoff;andtheselostbranchesofvarioussizesmayrepresentthosewholeorders,families,andgenerawhichhavenownolivingrepresentatives,andwhichareknowntousonlyfromhavingbeenfoundinafossilstate.Aswehereandthereseeathinstragglingbranchspringingfromaforklowdowninatree,andwhichbysomechancehasbeenfavouredandisstillaliveonitssummit,soweoccasionallyseeananimalliketheOrnithorhynchusorLepidosiren,whichinsomesmalldegreeconnectsbyitsaffinitiestwolargebranchesoflife,andwhichhasapparentlybeensavedfromfatalcompetitionbyhavinginhabitedaprotectedstation.Asbudsgiverisebygrowthtofreshbuds,andthese,ifvigorous,branchoutandovertoponallsidesmanyafeeblerbranch,sobygenerationIbelieveithasbeenwiththegreatTreeofLife,whichfillswithitsdeadandbrokenbranchesthecrustoftheearth,andcoversthesurfacewithitseverbranchingandbeautifulramifications.