forinstance，somekindsofmossesandlichenshaveneverbeenseentofructifyinFrance。
Someoftheseendemicandnaturalisedplantsareprobablyrenderedsterilefromexcessivemultiplicationbybuds，andtheirconsequentincapacitytoproduceandnourishseed。Butthesterilityofothersmoreprobablydependsonthepeculiarconditionsunderwhichtheylive，asinthecaseoftheivyinthenorthernpartofEurope，andofthetreesintheswampsoftheUnitedStates;yettheseplantsmustbeinsomerespectseminentlywelladaptedforthestationswhichtheyoccupy，fortheyholdtheirplacesagainstahostofcompetitors。]
Finally，thehighdegreeofsterilitywhichoftenaccompaniesthedoublingofflowers，oranexcessivedevelopmentoffruit，seldomsupervenesatonce。Anincipienttendencyisobserved，andcontinuedselectioncompletestheresult。
Theviewwhichseemsthemostprobable，andwhichconnectstogetheralltheforegoingfactsandbringsthemwithinourpresentsubject，is，thatchangedandunnaturalconditionsoflifefirstgiveatendencytosterility;andinconsequenceofthis，theorgansofreproductionbeingnolongerablefullytoperformtheirproperfunctions，asupplyoforganisedmatter，notrequiredforthedevelopmentoftheseed，flowseitherintotheseorgansandrendersthemfoliaceous，orintothefruit，stems，tubers，etc。，increasingtheirsizeandsucculency。Butitisprobablethatthereexists，independentlyofanyincipientsterility，anantagonismbetweenthetwoformsofreproduction，namely，byseedandbuds，wheneitheriscarriedtoanextremedegree。Thatincipientsterilityplaysanimportantpartinthedoublingofflowers，andintheothercasesjustspecified，Iinferchieflyfromthefollowingfacts。Whenfertilityislostfromawhollydifferentcause，namely，fromhybridism，thereisastrongtendency，asGartner（18/118。’Bastarderzeugung’s。565。Kolreuter’DritteFortsetzung’s。73，87，119）alsoshowsthatwhentwospecies，onesingleandtheotherdouble，arecrossed，thehybridsareapttobeextremelydouble。）affirms，forflowerstobecomedouble，andthistendencyisinherited。Moreover，itisnotoriousthatwithhybridsthemaleorgansbecomesterilebeforethefemaleorgans，andwithdoubleflowersthestamensfirstbecomefoliaceous。Thislatterfactiswellshownbythemaleflowersofdioeciousplants，which，accordingtoGallesio（18/119。’TeoriadellaRiproduzioneVeg。’1816page73。）firstbecomedouble。Again，Gartner（18/120。
’Bastarderzeugung’s。573。）ofteninsiststhattheflowersofevenutterlysterilehybrids，whichdonotproduceanyseed，generallyyieldperfectcapsulesorfruit，——afactwhichhaslikewisebeenrepeatedlyobservedbyNaudinwiththeCucurbitaceae;sothattheproductionoffruitbyplantsrenderedsterilethroughanycauseisintelligible。Kolreuterhasalsoexpressedhisunboundedastonishmentatthesizeanddevelopmentofthetubersincertainhybrids;andallexperimentalists（18/121。Ibids。527。）haveremarkedonthestrongtendencyinhybridstoincreasebyroots，runners，andsuckers。Seeingthathybridplants，whichfromtheirnaturearemoreorlesssterile，thustendtoproducedoubleflowers;thattheyhavethepartsincludingtheseed，thatisthefruit，perfectlydeveloped，evenwhencontainingnoseed;thattheysometimesyieldgiganticroots;thattheyalmostinvariablytendtoincreaselargelybysuckersandothersuchmeans;——seeingthis，andknowing，fromthemanyfactsgivenintheearlierpartsofthischapter，thatalmostallorganicbeingswhenexposedtounnaturalconditionstendtobecomemoreorlesssterile，itseemsmuchthemostprobableviewthatwithcultivatedplantssterilityistheexcitingcause，anddoubleflowers，richseedlessfruit，andinsomecaseslargely—developedorgansofvegetation，etc。，aretheindirectresults——theseresultshavingbeeninmostcaseslargelyincreasedthroughcontinuedselectionbyman。
CHAPTER2。XIX。
SUMMARYOFTHEFOURLASTCHAPTERS，WITHREMARKSONHYBRIDISM。
ONTHEEFFECTSOFCROSSING。
THEINFLUENCEOFDOMESTICATIONONFERTILITY。
CLOSEINTERBREEDING。
GOODANDEVILRESULTSFROMCHANGEDCONDITIONSOFLIFE。
VARIETIESWHENCROSSEDNOTINVARIABLYFERTILE。
ONTHEDIFFERENCEINFERTILITYBETWEENCROSSEDSPECIESANDVARIETIES。
CONCLUSIONSWITHRESPECTTOHYBRIDISM。
LIGHTTHROWNONHYBRIDISMBYTHEILLEGITIMATEPROGENYOFHETEROSTYLEDPLANTS。
STERILITYOFCROSSEDSPECIESDUETODIFFERENCESCONFINEDTOTHEREPRODUCTIVE
SYSTEM。
NOTACCUMULATEDTHROUGHNATURALSELECTION。
REASONSWHYDOMESTICVARIETIESARENOTMUTUALLYSTERILE。
TOOMUCHSTRESSHASBEENLAIDONTHEDIFFERENCEINFERTILITYBETWEENCROSSED
SPECIESANDCROSSEDVARIETIES。
CONCLUSION。
Itwasshowninthefifteenthchapterthatwhenindividualsofthesamevariety，orevenofadistinctvariety，areallowedfreelytointercross，uniformityofcharacterisultimatelyacquired。Somefewcharacters，however，areincapableoffusion，buttheseareunimportant，astheyareoftenofasemi—monstrousnature，andhavesuddenlyappeared。Hence，topreserveourdomesticatedbreedstrue，ortoimprovethembymethodicalselection，itisobviouslynecessarythattheyshouldbekeptseparate。Nevertheless，awholebodyofindividualsmaybeslowlymodified，throughunconsciousselection，asweshallseeinafuturechapter，withoutseparatingthemintodistinctlots。
Domesticraceshaveoftenbeenintentionallymodifiedbyoneortwocrosses，madewithsomealliedrace，andoccasionallyevenbyrepeatedcrosseswithverydistinctraces;butinalmostallsuchcases，long—continuedandcarefulselectionhasbeenabsolutelynecessary，owingtotheexcessivevariabilityofthecrossedoffspring，duetotheprincipleofreversion。Inafewinstances，however，mongrelshaveretainedauniformcharacterfromtheirfirstproduction。
Whentwovarietiesareallowedtocrossfreely，andoneismuchmorenumerousthantheother，theformerwillultimatelyabsorbthelatter。Shouldbothvarietiesexistinnearlyequalnumbers，itisprobablethataconsiderableperiodwouldelapsebeforetheacquirementofauniformcharacter;andthecharacterultimatelyacquiredwouldlargelydependonprepotencyoftransmissionandontheconditionsoflife;forthenatureoftheseconditionswouldgenerallyfavouronevarietymorethananother，sothatakindofnaturalselectionwouldcomeintoplay。Unlessthecrossedoffspringwereslaughteredbymanwithouttheleastdiscrimination，somedegreeofunmethodicalselectionwouldlikewisecomeintoaction。Fromtheseseveralconsiderationswemayinfer，thatwhentwoormorecloselyalliedspeciesfirstcameintothepossessionofthesametribe，theircrossingwillnothaveinfluenced，insogreatadegreeashasoftenbeensupposed，thecharacteroftheoffspringinfuturetimes;althoughinsomecasesitprobablyhashadaconsiderableeffect。
Domestication，asageneralrule，increasestheprolificnessofanimalsandplants。Iteliminatesthetendencytosterilitywhichiscommontospecieswhenfirsttakenfromastateofnatureandcrossed。Onthislatterheadwehavenodirectevidence;butasourracesofdogs，cattle，pigsetc。，arealmostcertainlydescendedfromaboriginallydistinctstocks，andastheseracesarenowfullyfertiletogether，oratleastincomparablymorefertilethanmostspecieswhencrossed，wemaywithentireconfidenceacceptthisconclusion。
Abundantevidencehasbeengiventhatcrossingaddstothesize，vigour，andfertilityoftheoffspring。Thisholdsgoodwhentherehasbeennopreviouscloseinterbreeding。Itappliestotheindividualsofthesamevarietybutbelongingtodifferentfamilies，todistinctvarieties，sub—species，andeventospecies。Inthelattercase，thoughsizeisgained，fertilityislost;buttheincreasedsize，vigour，andhardinessofmanyhybridscannotbeaccountedforsolelyontheprincipleofcompensationfromtheinactionofthereproductivesystem。Certainplantswhilstgrowingundertheirnaturalconditions，otherswhencultivated，andothersofhybridorigin，arecompletelyself—impotent，thoughperfectlyhealthy;andsuchplantscanbestimulatedtofertilityonlybybeingcrossedwithotherindividualsofthesameorofadistinctspecies。
Ontheotherhand，long—continuedcloseinterbreedingbetweenthenearestrelationsdiminishestheconstitutionalvigour，size，andfertilityoftheoffspring;andoccasionallyleadstomalformations，butnotnecessarilytogeneraldeteriorationofformorstructure。Thisfailureoffertilityshowsthattheevilresultsofinterbreedingareindependentoftheaugmentationofmorbidtendenciescommontobothparents，thoughthisaugmentationnodoubtisoftenhighlyinjurious。Ourbeliefthatevilfollowsfromcloseinterbreedingreststoacertainextentontheexperienceofpracticalbreeders，especiallyofthosewhohaverearedmanyanimalsofquicklypropagatingkinds;butitlikewiserestsonseveralcarefullyrecordedexperiments。Withsomeanimalscloseinterbreedingmaybecarriedonforalongperiodwithimpunitybytheselectionofthemostvigorousandhealthyindividuals;butsoonerorlaterevilfollows。Theevil，however，comesonsoslowlyandgraduallythatiteasilyescapesobservation，butcanberecognisedbythealmostinstantaneousmannerinwhichsize，constitutionalvigour，andfertilityareregainedwhenanimalsthathavelongbeeninterbredarecrossedwithadistinctfamily。
Thesetwogreatclassesoffacts，namely，thegoodderivedfromcrossing，andtheevilfromcloseinterbreeding，withtheconsiderationoftheinnumerableadaptationsthroughoutnatureforcompelling，orfavouring，oratleastpermitting，theoccasionalunionofdistinctindividuals，takentogether，leadtotheconclusionthatitisalawofnaturethatorganicbeingsshallnotfertilisethemselvesforperpetuity。Thislawwasfirstplainlyhintedatin1799，withrespecttoplants，byAndrewKnight（19/1。’TransactionsPhil。
Soc。’1799page202。ForKolreutersee’Mem。del’Acad。deSt。—Petersbourg’
tome31809published1811page197。InreadingC。K。Sprengel’sremarkablework，’DasentdeckteGeheimniss’etc。1793，itiscurioustoobservehowoftenthiswonderfullyacuteobserverfailedtounderstandthefullmeaningofthestructureoftheflowerswhichhehassowelldescribed，fromnotalwayshavingbeforehismindthekeytotheproblem，namely，thegoodderivedfromthecrossingofdistinctindividualplants。）and，notlongafterwards，thatsagaciousobserverKolreuter，aftershowinghowwelltheMalvaceaeareadaptedforcrossing，asks，\"anidaliquidinrecessuhabeat，quodhujuscemodifloresnunquampropriosuopulvere，sedsempereoaliarumsuspecieiimpregnentur，meritoquaritur？Certenaturanilfacitfrustra。\"AlthoughwemaydemurtoKolreuter’ssayingthatnaturedoesnothinginvain，seeinghowmanyrudimentaryanduselessorgansthereare，yetundoubtedlytheargumentfromtheinnumerablecontrivances，whichfavourcrossing，isofthegreatestweight。Themostimportantresultofthislawisthatitleadstouniformityofcharacterintheindividualsofthesamespecies。Inthecaseofcertainhermaphrodites，whichprobablyintercrossonlyatlongintervalsoftime，andwithunisexualanimalsinhabitingsomewhatseparatedlocalities，whichcanonlyoccasionallycomeintocontactandpair，thegreatervigourandfertilityofthecrossedoffspringwillultimatelytendtogiveuniformityofcharacter。
Butwhenwegobeyondthelimitsofthesamespecies，freeintercrossingisbarredbythelawofsterility。
Insearchingforfactswhichmightthrowlightonthecauseofthegoodeffectsfromcrossing，andoftheevileffectsfromcloseinterbreeding，wehaveseenthat，ontheonehand，itisawidelyprevalentandancientbelief，thatanimalsandplantsprofitfromslightchangesintheirconditionoflife;
anditwouldappearthatthegerm，inasomewhatanalogousmanner，ismoreeffectuallystimulatedbythemaleelement，whentakenfromadistinctindividual，andthereforeslightlymodifiedinnature，thanwhentakenfromamalehavingthesameidenticalconstitution。Ontheotherhand，numerousfactshavebeengiven，showingthatwhenanimalsarefirstsubjectedtocaptivity，evenintheirnativeland，andalthoughallowedmuchliberty，theirreproductivefunctionsareoftengreatlyimpairedorquiteannulled。Somegroupsofanimalsaremoreaffectedthanothers，butwithapparentlycapriciousexceptionsineverygroup。Someanimalsneverorrarelycoupleunderconfinement;somecouplefreely，butneverorrarelyconceive。Thesecondarymalecharacters，thematernalfunctionsandinstincts，areoccasionallyaffected。Withplants，whenfirstsubjectedtocultivation，analogousfactshavebeenobserved。Weprobablyoweourdoubleflowers，richseedlessfruits，andinsomecasesgreatlydevelopedtubers，etc。，toincipientsterilityoftheabovenaturecombinedwithacopioussupplyofnutriment。Animalswhichhavelongbeendomesticated，andplantswhichhavelongbeencultivated，cangenerallywithstand，withunimpairedfertility，greatchangesintheirconditionsoflife;thoughbotharesometimesslightlyaffected。Withanimalsthesomewhatrarecapacityofbreedingfreelyunderconfinement，togetherwiththeirutility，mainlydeterminethekindswhichhavebeendomesticated。
Wecaninnocasepreciselysaywhatisthecauseofthediminishedfertilityofananimalwhenfirstcaptured，orofaplantwhenfirstcultivated;wecanonlyinferthatitiscausedbyachangeofsomekindinthenaturalconditionsoflife。Theremarkablesusceptibilityofthereproductivesystemtosuchchanges，——asusceptibilitynotcommontoanyotherorgan，——apparentlyhasanimportantbearingonVariability，asweshallseeinafuturechapter。
Itisimpossiblenottobestruckwiththedoubleparallelismbetweenthetwoclassesoffactsjustalludedto。Ontheonehand，slightchangesintheconditionsoflife，andcrossesbetweenslightlymodifiedformsorvarieties，arebeneficialasfarasprolificnessandconstitutionalvigourareconcerned。
Ontheotherhand，changesintheconditionsgreaterindegree，orofadifferentnature，andcrossesbetweenformswhichhavebeenslowlyandgreatlymodifiedbynaturalmeans，——inotherwords，betweenspecies，——arehighlyinjurious，asfarasthereproductivesystemisconcerned，andinsomefewinstancesasfarasconstitutionalvigourisconcerned。Canthisparallelismbeaccidental？Doesitnotratherindicatesomerealbondofconnection？Asafiregoesoutunlessitbestirredup，sothevitalforcesarealwaystending，accordingtoMr。HerbertSpencer，toastateofequilibrium，unlessdisturbedandrenovatedthroughtheactionofotherforces。
Insomefewcasesvarietiestendtokeepdistinct，bybreedingatdifferentseasons，bygreatdifferenceinsize，orbysexualpreference。Butthecrossingofvarieties，farfromdiminishing，generallyaddstothefertilityofthefirstunionandofthemongreloffspring。Whetherallthemorewidelydistinctdomesticvarietiesareinvariablyquitefertilewhencrossed，wedonotpositivelyknow;muchtimeandtroublewouldberequisiteforthenecessaryexperiments，andmanydifficultiesoccur，suchasthedescentofthevariousracesfromaboriginallydistinctspecies，andthedoubtswhethercertainformsoughttoberankedasspeciesorvarieties。Nevertheless，thewideexperienceofpracticalbreedersprovesthatthegreatmajorityofvarieties，evenifsomeshouldhereafterprovenottobeindefinitelyfertileinterse，arefarmorefertilewhencrossed，thanthevastmajorityofcloselyalliednaturalspecies。Afewremarkablecaseshave，however，beengivenontheauthorityofexcellentobservers，showingthatwithplantscertainforms，whichundoubtedlymustberankedasvarieties，yieldfewerseedswhencrossedthanisnaturaltotheparent—species。Othervarietieshavehadtheirreproductivepowerssofarmodifiedthattheyareeithermoreorlessfertilethantheirparents，whencrossedwithadistinctspecies。
Nevertheless，thefactremainsindisputablethatdomesticatedvarieties，ofanimalsandofplants，whichdiffergreatlyfromoneanotherinstructure，butwhicharecertainlydescendedfromthesameaboriginalspecies，suchastheracesofthefowl，pigeon，manyvegetables，andahostofotherproductions，areextremelyfertilewhencrossed;andthisseemstomakeabroadandimpassablebarrierbetweendomesticvarietiesandnaturalspecies。But，asI
willnowattempttoshow，thedistinctionisnotsogreatandoverwhelminglyimportantasitatfirstappears。
ONTHEDIFFERENCEINFERTILITYBETWEENVARIETIESANDSPECIESWHENCROSSED。
Thisworkisnottheproperplaceforfullytreatingthesubjectofhybridism，andIhavealreadygiveninmy’OriginofSpecies’amoderatelyfullabstract。
Iwillheremerelyenumeratethegeneralconclusionswhichmaybereliedon，andwhichbearonourpresentpoint。
FIRSTLY。
Thelawsgoverningtheproductionofhybridsareidentical，ornearlyidentical，intheanimalandvegetablekingdoms。
SECONDLY。
Thesterilityofdistinctspecieswhenfirstunited，andthatoftheirhybridoffspring，graduate，byanalmostinfinitenumberofsteps，fromzero，whentheovuleisneverimpregnatedandaseed—capsuleisneverformed，uptocompletefertility。Wecanonlyescapetheconclusionthatsomespeciesarefullyfertilewhencrossed，bydeterminingtodesignateasvarietiesalltheformswhicharequitefertile。Thishighdegreeoffertilityis，however，rare。Nevertheless，plants，whichhavebeenexposedtounnaturalconditions，sometimesbecomemodifiedinsopeculiaramanner，thattheyaremuchmorefertilewhencrossedwithadistinctspeciesthanwhenfertilisedbytheirownpollen。Successineffectingafirstunionbetweentwospecies，andthefertilityoftheirhybrids，dependinaneminentdegreeontheconditionsoflifebeingfavourable。Theinnatesterilityofhybridsofthesameparentageandraisedfromthesameseed—capsuleoftendiffersmuchindegree。
THIRDLY。
Thedegreeofsterilityofafirstcrossbetweentwospeciesdoesnotalwaysrunstrictlyparallelwiththatoftheirhybridoffspring。Manycasesareknownofspecieswhichcanbecrossedwithease，butyieldhybridsexcessivelysterile;andconverselysomewhichcanbecrossedwithgreatdifficulty，butproducefairlyfertilehybrids。Thisisaninexplicablefact，ontheviewthatspecieshavebeenspeciallyendowedwithmutualsterilityinordertokeepthemdistinct。
FOURTHLY。
Thedegreeofsterilityoftendiffersgreatlyintwospecieswhenreciprocallycrossed;forthefirstwillreadilyfertilisethesecond;butthelatterisincapable，afterhundredsoftrials，offertilisingtheformer。Hybridsproducedfromreciprocalcrossesbetweenthesametwospecieslikewisesometimesdifferintheirdegreeofsterility。Thesecasesalsoareutterlyinexplicableontheviewofsterilitybeingaspecialendowment。
FIFTHLY。
Thedegreeofsterilityoffirstcrossesandofhybridsruns，toacertainextent，parallelwiththegeneralorsystematicaffinityoftheformswhichareunited。Forspeciesbelongingtodistinctgeneracanrarely，andthosebelongingtodistinctfamiliescannever，becrossed。Theparallelism，however，isfarfromcomplete;foramultitudeofcloselyalliedspecieswillnotunite，orunitewithextremedifficulty，whilstotherspecies，widelydifferentfromoneanother，canbecrossedwithperfectfacility。Nordoesthedifficultydependonordinaryconstitutionaldifferences，forannualandperennialplants，deciduousandevergreentrees，plantsfloweringatdifferentseasons，inhabitingdifferentstations，andnaturallylivingunderthemostoppositeclimates，canoftenbecrossedwithease。Thedifficultyorfacilityapparentlydependsexclusivelyonthesexualconstitutionofthespecieswhicharecrossed;orontheirsexualelectiveaffinity，i。e。WahlverwandtschaftofGartner。Asspeciesrarelyorneverbecomemodifiedinonecharacter，withoutbeingatthesametimemodifiedinmanycharacters，andassystematicaffinityincludesallvisiblesimilaritiesanddissimilarities，anydifferenceinsexualconstitutionbetweentwospecieswouldnaturallystandinmoreorlesscloserelationwiththeirsystematicposition。
SIXTHLY。
Thesterilityofspecieswhenfirstcrossed，andthatofhybrids，maypossiblydependtoacertainextentondistinctcauses。Withpurespeciesthereproductiveorgansareinaperfectcondition，whilstwithhybridstheyareoftenplainlydeteriorated。Ahybridembryowhichpartakesoftheconstitutionofitsfatherandmotherisexposedtounnaturalconditions，aslongasitisnourishedwithinthewomb，oregg，orseedofthemother—form;andasweknowthatunnaturalconditionsofteninducesterility，thereproductiveorgansofthehybridmightatthisearlyagebepermanentlyaffected。Butthiscausehasnobearingontheinfertilityoffirstunions。Thediminishednumberoftheoffspringfromfirstunionsmayoftenresult，asiscertainlysometimesthecase，fromtheprematuredeathofmostofthehybridembryos。Butweshallimmediatelyseethatalawofanunknownnatureapparentlyexists，whichleadstotheoffspringfromunions，whichareinfertile，beingthemselvesmoreorlessinfertile;andthisatpresentisallthatcanbesaid。
SEVENTHLY。
Hybridsandmongrelspresent，withtheonegreatexceptionoffertility，themoststrikingaccordanceinallotherrespects;namely，inthelawsoftheirresemblancetotheirtwoparents，intheirtendencytoreversion，intheirvariability，andinbeingabsorbedthroughrepeatedcrossesbyeitherparent—
form。
Afterarrivingattheseconclusions，Iwasledtoinvestigateasubjectwhichthrowsconsiderablelightonhybridism，namely，thefertilityofheterostyledordimorphicandtrimorphicplants，whenillegitimatelyunited。Ihavehadoccasionseveraltimestoalludetotheseplants，andImayheregiveabriefabstractofmyobservations。Severalplantsbelongingtodistinctorderspresenttwoforms，whichexistinaboutequalnumbers，andwhichdifferinnorespectexceptintheirreproductiveorgans;oneformhavingalongpistilwithshortstamens，theotherashortpistilwithlongstamens;bothwithdifferentlysizedpollen—grains。Withtrimorphicplantstherearethreeformslikewisedifferinginthelengthsoftheirpistilsandstamens，inthesizeandcolourofthepollen—grains，andinsomeotherrespects;andasineachofthethreeformstherearetwosetsofstamens，therearealtogethersixsetsofstamensandthreekindsofpistils。Theseorgansaresoproportionedinlengthtooneanotherthat，inanytwooftheforms，halfthestamensineachstandonalevelwiththestigmaofthethirdform。NowIhaveshown，andtheresulthasbeenconfirmedbyotherobservers，that，inordertoobtainfullfertilitywiththeseplants，itisnecessarythatthestigmaoftheoneformshouldbefertilisedbypollentakenfromthestamensofcorrespondingheightintheotherform。Sothatwithdimorphicspeciestwounions，whichmaybecalledlegitimate，arefullyfertile，andtwo，whichmaybecalledillegitimate，aremoreorlessinfertile。Withtrimorphicspeciessixunionsarelegitimate，orfullyfertile，andtwelveareillegitimate，ormoreorlessinfertile。（19/2。Myobservations’OntheCharacterandhybrid—likenatureoftheoffspringfromtheillegitimateunionofDimorphicandTrimorphicPlants’
werepublishedinthe’JournaloftheLinneanSoc。’volume10page393。Theabstractheregivenisnearlythesamewiththatwhichappearedinthe6theditionofmy’OriginofSpecies。’）
Theinfertilitywhichmaybeobservedinvariousdimorphicandtrimorphicplants，whenillegitimatelyfertilised，thatis，bypollentakenfromstamensnotcorrespondinginheightwiththepistil，differsmuchindegree，uptoabsoluteanduttersterility;justinthesamemannerasoccursincrossingdistinctspecies。Asthedegreeofsterilityinthelattercasedependsinaneminentdegreeontheconditionsoflifebeingmoreorlessfavourable，soI
havefounditwithillegitimateunions。Itiswellknownthatifpollenofadistinctspeciesbeplacedonthestigmaofaflower，anditsownpollenbeafterwards，evenafteraconsiderableintervaloftime，placedonthesamestigma，itsactionissostronglyprepotentthatitgenerallyannihilatestheeffectoftheforeignpollen;soitiswiththepollenoftheseveralformsofthesamespecies，forlegitimatepollenisstronglyprepotentoverillegitimatepollen，whenbothareplacedonthesamestigma。Iascertainedthisbyfertilisingseveralflowers，firstillegitimately，andtwenty—fourhoursafterwardslegitimately，withpollentakenfromapeculiarlycolouredvariety，andalltheseedlingsweresimilarlycoloured;thisshowsthatthelegitimatepollen，thoughappliedtwenty—fourhourssubsequently，hadwhollydestroyedorpreventedtheactionofthepreviouslyappliedillegitimatepollen。Again，as，inmakingreciprocalcrossesbetweenthesametwospecies，thereisoccasionallyagreatdifferenceintheresult，sothesamethingoccurswithtrimorphicplants;forinstance，themid—styledformofLythrumsalicariacouldbeillegitimatelyfertilisedwiththegreatesteasebypollenfromthelongerstamensoftheshort—styledform，andyieldedmanyseeds;buttheshort—styledformdidnotyieldasingleseedwhenfertilisedbythelongerstamensofthemid—styledform。
Inalltheserespectstheformsofthesameundoubtedspecies，whenillegitimatelyunited，behaveinexactlythesamemannerasdotwodistinctspecieswhencrossed。Thisledmecarefullytoobserveduringfouryearsmanyseedlings，raisedfromseveralillegitimateunions。Thechiefresultisthattheseillegitimateplants，astheymaybecalled，arenotfullyfertile。Itispossibletoraisefromdimorphicspecies，bothlong—styledandshort—styledillegitimateplants，andfromtrimorphicplantsallthreeillegitimateforms。
Thesecanthenbeproperlyunitedinalegitimatemanner。Whenthisisdone，thereisnoapparentreasonwhytheyshouldnotyieldasmanyseedsasdidtheirparentswhenlegitimatelyfertilised。Butsuchisnotthecase;theyareallinfertile，butinvariousdegrees;somebeingsoutterlyandincurablysterilethattheydidnotyieldduringfourseasonsasingleseedorevenseed—capsule。Theseillegitimateplants，whicharesosterile，althoughunitedwitheachotherinalegitimatemanner，maybestrictlycomparedwithhybridswhencrossedinterse，anditiswellknownhowsteriletheselattergenerallyare。When，ontheotherhand，ahybridiscrossedwitheitherpureparent—
species，thesterilityisusuallymuchlessened:andsoitiswhenanillegitimateplantisfertilisedbyalegitimateplant。Inthesamemannerasthesterilityofhybridsdoesnotalwaysrunparallelwiththedifficultyofmakingthefirstcrossbetweenthetwoparent—species，sothesterilityofcertainillegitimateplantswasunusuallygreat，whilstthesterilityoftheunionfromwhichtheywerederivedwasbynomeansgreat。Withhybridsraisedfromthesameseed—capsulethedegreeofsterilityisinnatelyvariable，soitisinamarkedmannerwithillegitimateplants。Lastly，manyhybridsareprofuseandpersistentflowerers，whilstotherandmoresterilehybridsproducefewflowers，andareweak，miserabledwarfs;exactlysimilarcasesoccurwiththeillegitimateoffspringofvariousdimorphicandtrimorphicplants。
Althoughthereistheclosestidentityincharacterandbehaviourbetweenillegitimateplantsandhybrids，itishardlyanexaggerationtomaintainthattheformerarehybrids，butproducedwithinthelimitsofthesamespeciesbytheimproperunionofcertainforms，whilstordinaryhybridsareproducedfromanimproperunionbetweenso—calleddistinctspecies。Wehavealreadyseenthatthereistheclosestsimilarityinallrespectsbetweenfirstillegitimateunions，andfirstcrossesbetweendistinctspecies。Thiswillperhapsbemademorefullyapparentbyanillustration:——wemaysupposethatabotanistfoundtwowell—markedvarieties（andsuchoccur）ofthelong—styledformofthetrimorphicLithrumsalicaria，andthathedeterminedtotrybycrossingwhethertheywerespecificallydistinct。Hewouldfindthattheyyieldedonlyaboutone—fifthofthepropernumberofseed，andthattheybehavedinalltheotherabove—specifiedrespectsasiftheyhadbeentwodistinctspecies。Buttomakethecasesure，hewouldraiseplantsfromhissupposedhybridisedseed，andhewouldfindthattheseedlingsweremiserablydwarfedandutterlysterile，andthattheybehavedinallotherrespectslikeordinaryhybrids，hemightthenmaintainthathehadactuallyproved，inaccordancewiththecommonview，thathistwovarietieswereasgoodandasdistinctspeciesasanyintheworld;buthewouldbecompletelymistaken。
Thefactsnowgivenondimorphicandtrimorphicplantsareimportant，becausetheyshowus，first，thatthephysiologicaltestoflessenedfertility，bothinfirstcrossesandinhybrids，isnocriterionofspecificdistinction;
secondly，becausewemayconcludethatthereissomeunknownbondwhichconnectstheinfertilityofillegitimateunionswiththatoftheirillegitimateoffspring，andweareledtoextendthesameviewtofirstcrossesandhybrids;thirdly，becausewefind，andthisseemstomeofespecialimportance，thattwoorthreeformsofthesamespeciesmayexistandmaydifferinnorespectwhatever，eitherinstructureorinconstitution，relativelytoexternalconditions，andyetbesterilewhenunitedincertainways。Forwemustrememberthatitistheunionofthesexualelementsofindividualsofthesameform，forinstance，oftwolong—styledforms，whichresultsinsterility;whilstitistheunionofthesexualelementpropertotwodistinctformswhichisfertile。Hencethecaseappearsatfirstsightexactlythereverseofwhatoccursintheordinaryunionsoftheindividualsofthesamespecies，andwithcrossesbetweendistinctspecies。Itis，however，doubtfulwhetherthisisreallyso;butIwillnotenlargeonthisobscuresubject。
Wemay，however，inferasprobablefromtheconsiderationofdimorphicandtrimorphicplants，thatthesterilityofdistinctspecieswhencrossed，andoftheirhybridprogeny，dependsexclusivelyonthenatureoftheirsexualelements，andnotonanydifferenceintheirstructureorgeneralconstitution。Wearealsoledtothissameconclusionbyconsideringreciprocalcrosses，inwhichthemaleofonespeciescannotbeunited，oronlywithgreatdifficulty，withthefemaleofasecondspecies，whilsttheconversecrosscanbeeffectedwithperfectfacility。Thatexcellentobserver，Gartner，likewiseconcludedthatspecieswhencrossedaresterileowingtodifferencesconfinedtotheirreproductivesystems。
Ontheprinciplewhichmakesitnecessaryforman，whilstheisselectingandimprovinghisdomesticvarieties，tokeepthemseparate，itwouldclearlybeadvantageoustovarietiesinastateofnature，thatistoincipientspecies，iftheycouldbekeptfromblending，eitherthroughsexualaversion，orbybecomingmutuallysterile。Henceitatonetimeappearedtomeprobable，asithastoothers，thatthissterilitymighthavebeenacquiredthroughnaturalselection。Onthisviewwemustsupposethatashadeoflessenedfertilityfirstspontaneouslyappeared，likeanyothermodification，incertainindividualsofaspecieswhencrossedwithotherindividualsofthesamespecies;andthatsuccessiveslightdegreesofinfertility，frombeingadvantageous，wereslowlyaccumulated。Thisappearsallthemoreprobable，ifweadmitthatthestructuraldifferencesbetweentheformsofdimorphicandtrimorphicplants，asthelengthandcurvatureofthepistil，etc。，havebeenco—adaptedthroughnaturalselection;forifthisbeadmitted，wecanhardlyavoidextendingthesameconclusiontotheirmutualinfertility。Sterility，moreover，hasbeenacquiredthroughnaturalselectionforotherandwidelydifferentpurposes，aswithneuterinsectsinreferencetotheirsocialeconomy。Inthecaseofplants，theflowersonthecircumferenceofthetrussintheguelderrose（Viburnumopulus）andthoseonthesummitofthespikeinthefeather—hyacinth（Muscaricomosum）havebeenrenderedconspicuous，andapparentlyinconsequencesterile，inorderthatinsectsmighteasilydiscoverandvisittheperfectflowers。Butwhenweendeavourtoapplytheprincipleofnaturalselectiontotheacquirementbydistinctspeciesofmutualsterility，wemeetwithgreatdifficulties。Inthefirstplace，itmayberemarkedthatseparateregionsareofteninhabitedbygroupsofspeciesorbysinglespecies，whichwhenbroughttogetherandcrossedarefoundtobemoreorlesssterile;nowitcouldclearlyhavebeennoadvantagetosuchseparatedspeciestohavebeenrenderedmutuallysterile，andconsequentlythiscouldnothavebeeneffectedthroughnaturalselection;butitmayperhapsbeargued，that，ifaspecieswererenderedsterilewithsomeonecompatriot，sterilitywithotherspecieswouldfollowasanecessaryconsequence。Inthesecondplace，itisasmuchopposedtothetheoryofnaturalselection，astothetheoryofspecialcreation，thatinreciprocalcrossesthemaleelementofoneformshouldhavebeenrenderedutterlyimpotentonasecondform，whilstatthesametimethemaleelementofthissecondformisenabledfreelytofertilisethefirstform;forthispeculiarstateofthereproductivesystemcouldnotpossiblyhavebeenadvantageoustoeitherspecies。
Inconsideringtheprobabilityofnaturalselectionhavingcomeintoactioninrenderingspeciesmutuallysterile，oneofthegreatestdifficultieswillbefoundtolieintheexistenceofmanygraduatedstepsfromslightlylessenedfertilitytoabsolutesterility。Itmaybeadmitted，ontheprincipleaboveexplained，thatitwouldprofitanincipientspeciesifitwererenderedinsomeslightdegreesterilewhencrossedwithitsparent—formorwithsomeothervariety;forthusfewerbastardisedanddeterioratedoffspringwouldbeproducedtocommingletheirbloodwiththenewspeciesinprocessofformation。Buthewhowilltakethetroubletoreflectonthestepsbywhichthisfirstdegreeofsterilitycouldbeincreasedthroughnaturalselectiontothathigherdegreewhichiscommontosomanyspecies，andwhichisuniversalwithspecieswhichhavebeendifferentiatedtoagenericorfamilyrank，willfindthesubjectextraordinarilycomplex。Aftermaturereflectionitseemstomethatthiscouldnothavebeeneffectedthroughnaturalselection。Takethecaseofanytwospecieswhich，whencrossed，producefewandsterileoffspring;now，whatistherewhichcouldfavourthesurvivalofthoseindividualswhichhappenedtobeendowedinaslightlyhigherdegreewithmutualinfertility，andwhichthusapproachedbyonesmallsteptowardsabsolutesterility？Yetanadvanceofthiskind，ifthetheoryofnaturalselectionbebroughttobear，musthaveincessantlyoccurredwithmanyspecies，foramultitudearemutuallyquitebarren。Withsterileneuterinsectswehavereasontobelievethatmodificationsintheirstructureandfertilityhavebeenslowlyaccumulatedbynaturalselection，fromanadvantagehavingbeenthusindirectlygiventothecommunitytowhichtheybelongedoverothercommunitiesofthesamespecies;butanindividualanimalnotbelongingtoasocialcommunity，ifrenderedslightlysterilewhencrossedwithsomeothervariety，wouldnotthusitselfgainanyadvantageorindirectlygiveanyadvantagetotheotherindividualsofthesamevariety，thusleadingtotheirpreservation。
Butitwouldbesuperfluoustodiscussthisquestionindetail;forwithplantswehaveconclusiveevidencethatthesterilityofcrossedspeciesmustbeduetosomeprinciple，quiteindependentofnaturalselection。BothGartnerandKolreuterhaveprovedthatingeneralincludingnumerousspecies，aseriescanbeformedfromspecieswhichwhencrossedyieldfewerandfewerseeds，tospecieswhichneverproduceasingleseed，butyetareaffectedbythepollenofcertainotherspecies，forthegermenswells。Itisheremanifestlyimpossibletoselectthemoresterileindividuals，whichhavealreadyceasedtoyieldseeds;sothatthisacmeofsterility，whenthegermenaloneisaffected，cannothavebeengainedthroughselection;andfromthelawsgoverningthevariousgradesofsterilitybeingsouniformthroughouttheanimalandvegetablekingdoms，wemayinferthatthecause，whateveritmaybe，isthesameornearlythesameinallcases。
Asspecieshavenotbeenrenderedmutuallyinfertilethroughtheaccumulativeactionofnaturalselection，andaswemaysafelyconclude，fromthepreviousaswellasfromotherandmoregeneralconsiderations，thattheyhavenotbeenendowedthroughanactofcreationwiththisquality，wemustinferthatithasarisenincidentallyduringtheirslowformationinconnectionwithotherandunknownchangesintheirorganisation。Byaqualityarisingincidentally，Irefertosuchcasesasdifferentspeciesofanimalsandplantsbeingdifferentlyaffectedbypoisonstowhichtheyarenotnaturallyexposed;andthisdifferenceinsusceptibilityisclearlyincidentalonotherandunknowndifferencesintheirorganisation。Soagainthecapacityindifferentkindsoftreestobegraftedoneachother，oronathirdspecies，differsmuch，andisofnoadvantagetothesetrees，butisincidentalonstructuralorfunctionaldifferencesintheirwoodytissues。Weneednotfeelsurpriseatsterilityincidentallyresultingfromcrossesbetweendistinctspecies，——themodifieddescendantsofacommonprogenitor，——whenwebearinmindhoweasilythereproductivesystemisaffectedbyvariouscauses——oftenbyextremelyslightchangesintheconditionsoflife，bytoocloseinterbreeding，andbyotheragencies。ItiswelltobearinmindsuchcasesasthatofthePassifloraalata，whichrecovereditsself—fertilityfrombeinggraftedonadistinctspecies——thecasesofplantswhichnormallyorabnormallyareself—impotent，butcanreadilybefertilisedbythepollenofadistinctspecies——andlastlythecasesofindividualdomesticatedanimalswhichevincetowardseachothersexualincompatibility。
Wenowatlastcometotheimmediatepointunderdiscussion:howisitthat，withsomefewexceptionsinthecaseofplants，domesticatedvarieties，suchasthoseofthedog，fowl，pigeon，severalfruit—trees，andculinaryvegetables，whichdifferfromeachotherinexternalcharactersmorethanmanyspecies，areperfectlyfertilewhencrossed，orevenfertileinexcess，whilstcloselyalliedspeciesarealmostinvariablyinsomedegreesterile？Wecan，toacertainextent，giveasatisfactoryanswertothisquestion。Passingoverthefactthattheamountofexternaldifferencebetweentwospeciesisnosureguidetotheirdegreeofmutualsterility，sothatsimilardifferencesinthecaseofvarietieswouldbenosureguide，weknowthatwithspeciesthecauseliesexclusivelyindifferencesintheirsexualconstitution。Nowtheconditionstowhichdomesticatedanimalsandcultivatedplantshavebeensubjectedhavehadsolittletendencytowardsmodifyingthereproductivesysteminamannerleadingtomutualsterility，thatwehaveverygoodgroundsforadmittingthedirectlyoppositedoctrineofPallas，namely，thatsuchconditionsgenerallyeliminatethistendency;sothatthedomesticateddescendantsofspecies，whichintheirnaturalstatewouldhavebeeninsomedegreesterilewhencrossed，becomeperfectlyfertiletogether。Withplants，sofariscultivationfromgivingatendencytowardsmutualsterility，thatinseveralwell—authenticatedcases，alreadyoftenalludedto，certainspecieshavebeenaffectedinaverydifferentmanner，fortheyhavebecomeself—
impotent，whilststillretainingthecapacityoffertilising，andbeingfertilisedby，distinctspecies。IfthePallasiandoctrineoftheeliminationofsterilitythroughlong—continueddomesticationbeadmitted，anditcanhardlyberejected，itbecomesinthehighestdegreeimprobablethatsimilarcircumstancesshouldcommonlybothinduceandeliminatethesametendency;
thoughincertaincases，withspecieshavingapeculiarconstitution，sterilitymightoccasionallybethusinduced。Thus，asIbelieve，wecanunderstandwhywithdomesticatedanimalsvarietieshavenotbeenproducedwhicharemutuallysterile;andwhywithplantsonlyafewsuchcaseshavebeenobserved，namely，byGartner，withcertainvarietiesofmaizeandverbascum，byotherexperimentalistswithvarietiesofthegourdandmelon，andbyKolreuterwithonekindoftobacco。
Withrespecttovarietieswhichhaveoriginatedinastateofnature，itisalmosthopelesstoexpecttoprovebydirectevidencethattheyhavebeenrenderedmutuallysterile;forifevenatraceofsterilitycouldbedetected，suchvarietieswouldatonceberaisedbyalmosteverynaturalisttotherankofdistinctspecies。If，forinstance，Gartner’sstatementwerefullyconfirmed，thattheblueandredfloweredformsofthepimpernel（Anagallisarvensis）aresterilewhencrossed，Ipresumethatallthebotanistswhonowmaintainonvariousgroundsthatthesetwoformsaremerelyfleetingvarieties，wouldatonceadmitthattheywerespecificallydistinct。
Therealdifficultyinourpresentsubjectisnot，asitappearstome，whydomesticvarietieshavenotbecomemutuallyinfertilewhencrossed，butwhythishassogenerallyoccurredwithnaturalvarietiesassoonastheyhavebeenmodifiedinasufficientandpermanentdegreetotakerankasspecies。Wearefarfrompreciselyknowingthecause;butwecanseethatthespecies，owingtotheirstruggleforexistencewithnumerouscompetitors，musthavebeenexposedtomoreuniformconditionsoflifeduringlongperiodsoftimethandomesticvarietieshavebeen，andthismaywellmakeawidedifferenceintheresult。Forweknowhowcommonlywildanimalsandplants，whentakenfromtheirnaturalconditionsandsubjectedtocaptivity，arerenderedsterile;andthereproductivefunctionsoforganicbeingswhichhavealwayslivedandbeenslowlymodifiedundernaturalconditionswouldprobablyinlikemannerbeeminentlysensitivetotheinfluenceofanunnaturalcross。Domesticatedproductions，ontheotherhand，which，asshownbythemerefactoftheirdomestication，werenotoriginallyhighlysensitivetochangesintheirconditionsoflife，andwhichcannowgenerallyresistwithundiminishedfertilityrepeatedchangesofconditions，mightbeexpectedtoproducevarieties，whichwouldbelittleliabletohavetheirreproductivepowersinjuriouslyaffectedbytheactofcrossingwithothervarietieswhichhadoriginatedinalikemanner。
Certainnaturalistshaverecentlylaidtoogreatstress，asitappearstome，onthedifferenceinfertilitybetweenvarietiesandspecieswhencrossed。
Somealliedspeciesoftreescannotbegraftedononeanother，whilstallvarietiescanbesografted。Somealliedanimalsareaffectedinaverydifferentmannerbythesamepoison，butwithvarietiesnosuchcaseuntilrecentlywasknown;whilstnowithasbeenprovedthatimmunityfromcertainpoisonssometimesstandsincorrelationwiththecolouroftheindividualsofthesamespecies。Theperiodofgestationgenerallydiffersmuchindistinctspecies，butwithvarietiesuntillatelynosuchdifferencehadbeenobserved。
Herewehavevariousphysiologicaldifferences，andnodoubtotherscouldbeadded，betweenonespeciesandanotherofthesamegenus，whichdonotoccur，oroccurwithextremerarity，inthecaseofvarieties;andthesedifferencesareapparentlywhollyorinchiefpartincidentalonotherconstitutionaldifferences，justinthesamemannerasthesterilityofcrossedspeciesisincidentalondifferencesconfinedtothesexualsystem。Why，then，shouldtheselatterdifferences，howeverserviceabletheymayindirectlybeinkeepingtheinhabitantsofthesamecountrydistinct，bethoughtofsuchparamountimportance，incomparisonwithotherincidentalandfunctionaldifferences？Nosufficientanswertothisquestioncanbegiven。Hencethefactthatwidelydistinctdomesticvarietiesare，withrareexceptions，perfectlyfertilewhencrossed，andproducefertileoffspring，whilstcloselyalliedspeciesare，withrareexceptions，moreorlesssterile，isnotnearlysoformidableanobjectionasitappearsatfirsttothetheoryofthecommondescentofalliedspecies。
CHAPTER2。XX。
SELECTIONBYMAN。
SELECTIONADIFFICULTART。
METHODICAL，UNCONSCIOUS，ANDNATURALSELECTION。
RESULTSOFMETHODICALSELECTION。
CARETAKENINSELECTION。
SELECTIONWITHPLANTS。
SELECTIONCARRIEDONBYTHEANCIENTSANDBYSEMI—CIVILISEDPEOPLE。
UNIMPORTANTCHARACTERSOFTENATTENDEDTO。
UNCONSCIOUSSELECTION。
ASCIRCUMSTANCESSLOWLYCHANGE，SOHAVEOURDOMESTICATEDANIMALSCHANGED
THROUGHTHEACTIONOFUNCONSCIOUSSELECTION。
INFLUENCEOFDIFFERENTBREEDERSONTHESAMESUB—VARIETY。
PLANTSASAFFECTEDBYUNCONSCIOUSSELECTION。
EFFECTSOFSELECTIONASSHOWNBYTHEGREATAMOUNTOFDIFFERENCEINTHEPARTS
MOSTVALUEDBYMAN。
ThepowerofSelection，whetherexercisedbyman，orbroughtintoplayundernaturethroughthestruggleforexistenceandtheconsequentsurvivalofthefittest，absolutelydependsonthevariabilityoforganicbeings。Withoutvariabilitynothingcanbeeffected;slightindividualdifferences，however，sufficeforthework，andareprobablythechieforsolemeansintheproductionofnewspecies。Henceourdiscussiononthecausesandlawsofvariabilityoughtinstrictordertohaveprecededthepresentsubject，aswellasinheritance，crossing，etc。;butpracticallythepresentarrangementhasbeenfoundthemostconvenient。Mandoesnotattempttocausevariability;
thoughheunintentionallyeffectsthisbyexposingorganismstonewconditionsoflife，andbycrossingbreedsalreadyformed。Butvariabilitybeinggranted，heworkswonders。Unlesssomedegreeofselectionbeexercised，thefreecomminglingoftheindividualsofthesamevarietysoonobliterates，aswehavepreviouslyseen，theslightdifferenceswhicharise，andgivesuniformityofcharactertothewholebodyofindividuals。Inseparateddistricts，long—
continuedexposuretodifferentconditionsoflifemayproducenewraceswithouttheaidofselection;buttothissubjectofthedirectactionoftheconditionsoflifeIshallrecurinafuturechapter。
Whenanimalsorplantsarebornwithsomeconspicuousandfirmlyinheritednewcharacter，selectionisreducedtothepreservationofsuchindividuals，andtothesubsequentpreventionofcrosses;sothatnothingmoreneedbesaidonthesubject。Butinthegreatmajorityofcasesanewcharacter，orsomesuperiorityinanoldcharacter，isatfirstfaintlypronounced，andisnotstronglyinherited;andthenthefulldifficultyofselectionisexperienced。