第6章

SummaryofChapter。Wehaveinthischapterdiscussedsomeofthedifficultiesandobjectionswhichmaybeurgedagainstmytheory。Manyofthemareverygrave。butIthinkthatinthediscussionlighthasbeenthrownonseveralfacts，whichonthetheoryofindependentactsofcreationareutterlyobscure。Wehaveseenthatspeciesatanyoneperiodarenotindefinitelyvariable，andarenotlinkedtogetherbyamultitudeofintermediategradations，partlybecausetheprocessofnaturalselectionwillalwaysbeveryslow，andwillact，atanyonetime，onlyonaveryfewforms。

andpartlybecausetheveryprocessofnaturalselectionalmostimpliesthecontinualsupplantingandextinctionofprecedingandintermediategradations。Closelyalliedspecies，nowlivingonacontinuousarea，mustoftenhavebeenformedwhentheareawasnotcontinuous，andwhentheconditionsoflifedidnotinsensiblygraduateawayfromoneparttoanother。Whentwovarietiesareformedintwodistrictsofacontinuousarea，anintermediatevarietywilloftenbeformed，fittedforanintermediatezone。butfromreasonsassigned，theintermediatevarietywillusuallyexistinlessernumbersthanthetwoformswhichitconnects。consequentlythetwolatter，duringthecourseoffurthermodification，fromexistingingreaternumbers，willhaveagreatadvantageoverthelessnumerousintermediatevariety，andwillthusgenerallysucceedinsupplantingandexterminatingit。

Wehaveseeninthischapterhowcautiousweshouldbeinconcludingthatthemostdifferenthabitsoflifecouldnotgraduateintoeachother。

thatabat，forinstance，couldnothavebeenformedbynaturalselectionfromananimalwhichatfirstcouldonlyglidethroughtheair。

Wehaveseenthataspeciesmayundernewconditionsoflifechangeitshabits，orhavediversifiedhabits，withsomehabitsveryunlikethoseofitsnearestcongeners。Hencewecanunderstandbearinginmindthateachorganicbeingistryingtolivewhereveritcanlive，howithasarisenthatthereareuplandgeesewithwebbedfeet，groundwoodpeckers，divingthrushes，andpetrelswiththehabitsofauks。

Althoughthebeliefthatanorgansoperfectastheeyecouldhavebeenformedbynaturalselection，ismorethanenoughtostaggeranyone。yetinthecaseofanyorgan，ifweknowofalongseriesofgradationsincomplexity，eachgoodforitspossessor，then，underchangingconditionsoflife，thereisnologicalimpossibilityintheacquirementofanyconceivabledegreeofperfectionthroughnaturalselection。Inthecasesinwhichweknowofnointermediateortransitionalstates，weshouldbeverycautiousinconcludingthatnonecouldhaveexisted，forthehomologiesofmanyorgansandtheirintermediatestatesshowthatwonderfulmetamorphosesinfunctionareatleastpossible。Forinstance，aswim-bladderhasapparentlybeenconvertedintoanair-breathinglung。Thesameorganhavingperformedsimultaneouslyverydifferentfunctions，andthenhavingbeenspecialisedforonefunction。andtwoverydistinctorganshavingperformedatthesametimethesamefunction，theonehavingbeenperfectedwhilstaidedbytheother，mustoftenhavelargelyfacilitatedtransitions。

Wearefartooignorant，inalmosteverycase，tobeenabledtoassertthatanypartororganissounimportantforthewelfareofaspecies，thatmodificationsinitsstructurecouldnothavebeenslowlyaccumulatedbymeansofnaturalselection。Butwemayconfidentlybelievethatmanymodifications，whollyduetothelawsofgrowth，andatfirstinnowayadvantageoustoaspecies，havebeensubsequentlytakenadvantageofbythestillfurthermodifieddescendantsofthisspecies。Wemay，also，believethatapartformerlyofhighimportancehasoftenbeenretained(asthetailofanaquaticanimalbyitsterrestrialdescendants)，thoughithasbecomeofsuchsmallimportancethatitcouldnot，initspresentstate，havebeenacquiredbynaturalselection，apowerwhichactssolelybythepreservationofprofitablevariationsinthestruggleforlife。

Naturalselectionwillproducenothinginonespeciesfortheexclusivegoodorinjuryofanother。thoughitmaywellproduceparts，organs，andexcretionshighlyusefulorevenindispensable，orhighlyinjurioustoanotherspecies，butinallcasesatthesametimeusefultotheowner。

Naturalselectionineachwell-stockedcountry，mustactchieflythroughthecompetitionoftheinhabitantsonewithanother，andconsequentlywillproduceperfection，orstrengthinthebattleforlife，onlyaccordingtothestandardofthatcountry。Hencetheinhabitantsofonecountry，generallythesmallerone，willoftenyield，asweseetheydoyield，totheinhabitantsofanotherandgenerallylargercountry。Forinthelargercountrytherewillhaveexistedmoreindividuals，andmorediversifiedforms，andthecompetitionwillhavebeenseverer，andthusthestandardofperfectionwillhavebeenrenderedhigher。Naturalselectionwillnotnecessarilyproduceabsoluteperfection。nor，asfaraswecanjudgebyourlimitedfaculties，canabsoluteperfectionbeeverywherefound。

Onthetheoryofnaturalselectionwecanclearlyunderstandthefullmeaningofthatoldcanoninnaturalhistory，’Naturanonfacitsaltum。’

Thiscanon，ifwelookonlytothepresentinhabitantsoftheworld，isnotstrictlycorrect，butifweincludeallthoseofpasttimes，itmustbymytheorybestrictlytrue。

ItisgenerallyacknowledgedthatallorganicbeingshavebeenformedontwogreatlawsUnityofType，andtheConditionsofExistence。Byunityoftypeismeantthatfundamentalagreementinstructure，whichweseeinorganicbeingsofthesameclass，andwhichisquiteindependentoftheirhabitsoflife。Onmytheory，unityoftypeisexplainedbyunityofdescent。Theexpressionofconditionsofexistence，soofteninsistedonbytheillustriousCuvier，isfullyembracedbytheprincipleofnaturalselection。Fornaturalselectionactsbyeithernowadaptingthevaryingpartsofeachbeingtoitsorganicandinorganicconditionsoflife。orbyhavingadaptedthemduringlong-pastperiodsoftime：theadaptationsbeingaidedinsomecasesbyuseanddisuse，beingslightlyaffectedbythedirectactionoftheexternalconditionsoflife，andbeinginallcasessubjectedtotheseverallawsofgrowth。Hence，infact，thelawoftheConditionsofExistenceisthehigherlaw。asitincludes，throughtheinheritanceofformeradaptations，thatofUnityofType。PreviousChapterNextChapterTheOriginofSpecies-Chapter7TheOriginofSpeciesbyCharlesDarwinPreviousChapterNextChapterChapter7-InstinctTHEsubjectofinstinctmighthavebeenworkedintothepreviouschapters。

butIhavethoughtthatitwouldbemoreconvenienttotreatthesubjectseparately，especiallyassowonderfulaninstinctasthatofthehive-beemakingitscellswillprobablyhaveoccurredtomanyreaders，asadifficultysufficienttooverthrowmywholetheory。Imustpremise，thatIhavenothingtodowiththeoriginoftheprimarymentalpowers，anymorethanIhavewiththatoflifeitself。Weareconcernedonlywiththediversitiesofinstinctandoftheothermentalqualitiesofanimalswithinthesameclass。

Iwillnotattemptanydefinitionofinstinct。Itwouldbeeasytoshowthatseveraldistinctmentalactionsarecommonlyembracedbythisterm。

buteveryoneunderstandswhatismeant，whenitissaidthatinstinctimpelsthecuckootomigrateandtolayhereggsinotherbirds’nests。

Anaction，whichweourselvesshouldrequireexperiencetoenableustoperform，whenperformedbyananimal，moreespeciallybyaveryyoungone，withoutanyexperience，andwhenperformedbymanyindividualsinthesameway，withouttheirknowingforwhatpurposeitisperformed，isusuallysaidtobeinstinctive。ButIcouldshowthatnoneofthesecharactersofinstinctareuniversal。Alittledose，asPierreHuberexpressesit，ofjudgmentorreason，oftencomesintoplay，eveninanimalsverylowinthescaleofnature。

FrederickCuvierandseveraloftheoldermetaphysicianshavecomparedinstinctwithhabit。Thiscomparisongives，Ithink，aremarkablyaccuratenotionoftheframeofmindunderwhichaninstinctiveactionisperformed，butnotofitsorigin。Howunconsciouslymanyhabitualactionsareperformed，indeednotrarelyindirectoppositiontoourconsciouswill！yettheymaybemodifiedbythewillorreason。Habitseasilybecomeassociatedwithotherhabits，andwithcertainperiodsoftimeandstatesofthebody。

Whenonceacquired，theyoftenremainconstantthroughoutlife。Severalotherpointsofresemblancebetweeninstinctsandhabitscouldbepointedout。Asinrepeatingawell-knownsong，soininstincts，oneactionfollowsanotherbyasortofrhythm。ifapersonbeinterruptedinasong，orinrepeatinganythingbyrote，heisgenerallyforcedtogobacktorecoverthehabitualtrainofthought：soP。Huberfounditwaswithacaterpillar，whichmakesaverycomplicatedhammock。forifhetookacaterpillarwhichhadcompleteditshammockupto，say，thesixthstageofconstruction，andputitintoahammockcompleteduponlytothethirdstage，thecaterpillarsimplyre-performedthefourth，fifth，andsixthstagesofconstruction。

If，however，acaterpillarweretakenoutofahammockmadeup，forinstance，tothethirdstage，andwereputintoonefinisheduptothesixthstage，sothatmuchofitsworkwasalreadydoneforit，farfromfeelingthebenefitofthis，itwasmuchembarrassed，and，inordertocompleteitshammock，seemedforcedtostartfromthethirdstage，whereithadleftoff，andthustriedtocompletethealreadyfinishedwork。

IfwesupposeanyhabitualactiontobecomeinheritedandIthinkitcanbeshownthatthisdoessometimeshappenthentheresemblancebetweenwhatoriginallywasahabitandaninstinctbecomessocloseasnottobedistinguished。IfMozart，insteadofplayingthepianoforteatthreeyearsoldwithwonderfullylittlepractice，hadplayedatunewithnopracticeatall，bemighttrulybesaidtohavedonesoinstinctively。Butitwouldbethemostseriouserrortosupposethatthegreaternumberofinstinctshavebeenacquiredbyhabitinonegeneration，andthentransmittedbyinheritancetosucceedinggenerations。Itcanbeclearlyshownthatthemostwonderfulinstinctswithwhichweareacquainted，namely，thoseofthehive-beeandofmanyants，couldnotpossiblyhavebeenthusacquired。

Itwillbeuniversallyadmittedthatinstinctsareasimportantascorporealstructureforthewelfareofeachspecies，underitspresentconditionsoflife。Underchangedconditionsoflife，itisatleastpossiblethatslightmodificationsofinstinctmightbeprofitabletoaspecies。andifitcanbeshownthatinstinctsdovaryeversolittle，thenIcanseenodifficultyinnaturalselectionpreservingandcontinuallyaccumulatingvariationsofinstincttoanyextentthatmaybeprofitable。Itisthus，asIbelieve，thatallthemostcomplexandwonderfulinstinctshaveoriginated。

Asmodificationsofcorporealstructurearisefrom，andareincreasedby，useorhabit，andarediminishedorlostbydisuse，soIdonotdoubtithasbeenwithinstincts。ButIbelievethattheeffectsofhabitareofquitesubordinateimportancetotheeffectsofthenaturalselectionofwhatmaybecalledaccidentalvariationsofinstincts。thatisofvariationsproducedbythesameunknowncauseswhichproduceslightdeviationsofbodilystructure。

Nocomplexinstinctcanpossiblybeproducedthroughnaturalselection，exceptbytheslowandgradualaccumulationofnumerous，slight，yetprofitable，variations。Hence，asinthecaseofcorporealstructures，weoughttofindinnature，nottheactualtransitionalgradationsbywhicheachcomplexinstincthasbeenacquiredforthesecouldbefoundonlyinthelinealancestorsofeachspeciesbutweoughttofindinthecollaterallinesofdescentsomeevidenceofsuchgradations。orweoughtatleasttobeabletoshowthatgradationsofsomekindarepossible。andthiswecertainlycando。Ihavebeensurprisedtofind，makingallowancefortheinstinctsofanimalshavingbeenbutlittleobservedexceptinEuropeandNorthAmerica，andfornoinstinctbeingknownamongstextinctspecies，howverygenerallygradations，leadingtothemostcomplexinstincts，canbediscovered。Thecanonof’Naturanonfacitsaltum’applieswithalmostequalforcetoinstinctsastobodilyorgans。Changesofinstinctmaysometimesbefacilitatedbythesamespecieshavingdifferentinstinctsatdifferentperiodsoflife，oratdifferentseasonsoftheyear，orwhenplacedunderdifferentcircumstances，&。c……inwhichcaseeitheroneortheotherinstinctmightbepreservedbynaturalselection。Andsuchinstancesofdiversityofinstinctinthesamespeciescanbeshowntooccurinnature。

Againasinthecaseofcorporealstructure，andconformablywithmytheory，theinstinctofeachspeciesisgoodforitself，buthasnever，asfaraswecanjudge，beenproducedfortheexclusivegoodofothers。

Oneofthestrongestinstancesofananimalapparentlyperforminganactionforthesolegoodofanother，withwhichIamacquainted，isthatofaphidesvoluntarilyyieldingtheirsweetexcretiontoants：thattheydosovoluntarily，thefollowingfactsshow。Iremovedalltheantsfromagroupofaboutadozenaphidesonadock-plant，andpreventedtheirattendanceduringseveralhours。Afterthisinterval，Ifeltsurethattheaphideswouldwanttoexcrete。Iwatchedthemforsometimethroughalens，butnotoneexcreted。Ithentickledandstrokedthemwithahairinthesamemanner，aswellasIcould，astheantsdowiththeirantennae。butnotoneexcreted。

AfterwardsIallowedananttovisitthem，anditimmediatelyseemed，byitseagerwayofrunningabout，tobewellawarewhatarichflockithaddiscovered。itthenbegantoplaywithitsantennaeontheabdomenfirstofoneaphisandthenofanother。andeachaphis，assoonasitfelttheantennae，immediatelyliftedupitsabdomenandexcretedalimpiddropofsweetjuice，whichwaseagerlydevouredbytheant。Eventhequiteyoungaphidesbehavedinthismanner，showingthattheactionwasinstinctive，andnottheresultofexperience。Butastheexcretionisextremelyviscid，itisprobablyaconveniencetotheaphidestohaveitremoved。andthereforeprobablytheaphidesdonotinstinctivelyexcreteforthesolegoodoftheants。AlthoughIdonotbelievethatanyanimalintheworldperformsanactionfortheexclusivegoodofanotherofadistinctspecies，yeteachspeciestriestotakeadvantageoftheinstinctsofothers，aseachtakesadvantageoftheweakerbodilystructureofothers。Soagain，insomefewcases，certaininstinctscannotbeconsideredasabsolutelyperfect。

butasdetailsonthisandothersuchpointsarenotindispensable，theymaybeherepassedover。

Assomedegreeofvariationininstinctsunderastateofnature，andtheinheritanceofsuchvariations，areindispensablefortheactionofnaturalselection，asmanyinstancesaspossibleoughttohavebeenheregiven。butwantofspacepreventsme。Icanonlyassert，thatinstinctscertainlydovaryforinstance，themigratoryinstinct，bothinextentanddirection，andinitstotalloss。Soitiswiththenestsofbirds，whichvarypartlyindependenceonthesituationschosen，andonthenatureandtemperatureofthecountryinhabited，butoftenfromcauseswhollyunknowntous：AudubonhasgivenseveralremarkablecasesofdifferencesinnestsofthesamespeciesinthenorthernandsouthernUnitedStates。

Fearofanyparticularenemyiscertainlyaninstinctivequality，asmaybeseeninnestlingbirds，thoughitisstrengthenedbyexperience，andbythesightoffearofthesameenemyinotheranimals。Butfearofmanisslowlyacquired，asIhaveelsewhereshown，byvariousanimalsinhabitingdesertislands。andwemayseeaninstanceofthis，eveninEngland，inthegreaterwildnessofallourlargebirdsthanofoursmallbirds。forthelargebirdshavebeenmostpersecutedbyman。Wemaysafelyattributethegreaterwildnessofourlargebirdstothiscause。forinuninhabitedislandslargebirdsarenotmorefearfulthansmall。andthemagpie，sowaryinEngland，istameinNorway，asisthehoodedcrowinEgypt。

Thatthegeneraldispositionofindividualsofthesamespecies，borninastateofnature，isextremelydiversified，canbeshownbyamultitudeoffacts。Severalcasesalso，couldbegiven，ofoccasionalandstrangehabitsincertainspecies，whichmight，ifadvantageoustothespecies，giverise，throughnaturalselection，toquitenewinstincts。ButIamwellawarethatthesegeneralstatements，withoutfactsgivenindetail，canproducebutafeebleeffectonthereader’smind。Icanonlyrepeatmyassurance，thatIdonotspeakwithoutgoodevidence。

Thepossibility，orevenprobability，ofinheritedvariationsofinstinctinastateofnaturewillbestrengthenedbybrieflyconsideringafewcasesunderdomestication。Weshallthusalsobeenabledtoseetherespectivepartswhichhabitandtheselectionofso-calledaccidentalvariationshaveplayedinmodifyingthementalqualitiesofourdomesticanimals。

Anumberofcuriousandauthenticinstancescouldbegivenoftheinheritanceofallshadesofdispositionandtastes，andlikewiseoftheoddesttricks，associatedwithcertainframesofmindorperiodsoftime。Butletuslooktothefamiliarcaseoftheseveralbreedsofdogs：itcannotbedoubtedthatyoungpointers(Ihavemyselfseenastrikinginstance)willsometimespointandevenbackotherdogstheveryfirsttimethattheyaretakenout。retrievingiscertainlyinsomedegreeinheritedbyretrievers。andatendencytorunround，insteadofat，aflockofsheep，byshepherd-dogs。

Icannotseethattheseactions，performedwithoutexperiencebytheyoung，andinnearlythesamemannerbyeachindividual，performedwitheagerdelightbyeachbreed，andwithouttheendbeingknown，fortheyoungpointercannomoreknowthathepointstoaidhismaster，thanthewhitebutterflyknowswhyshelayshereggsontheleafofthecabbage，Icannotseethattheseactionsdifferessentiallyfromtrueinstincts。Ifweweretoseeonekindofwolf，whenyoungandwithoutanytraining，assoonasitscenteditsprey，standmotionlesslikeastatue，andthenslowlycrawlforwardwithapeculiargait。andanotherkindofwolfrushinground，insteadofat，aherdofdeer，anddrivingthemtoadistantpoint，weshouldassuredlycalltheseactionsinstinctive。Domesticinstincts，astheymaybecalled，arecertifyfarlessfixedorinvariablethannaturalinstincts。buttheyhavebeenactedonbyfarlessrigorousselection，andhavebeentransmittedforanincomparablyshorterperiod，underlessfixedconditionsoflife。

Howstronglythesedomesticinstincts，habits，anddispositionsareinherited，andhowcuriouslytheybecomemingled，iswellshownwhendifferentbreedsofdogsarecrossed。Thusitisknownthatacrosswithabull-doghasaffectedformanygenerationsthecourageandobstinacyofgreyhounds。

andacrosswithagreyhoundhasgiventoawholefamilyofshepherd-dogsatendencytohunthares。Thesedomesticinstincts，whenthustestedbycrossing，resemblenaturalinstincts，whichinalikemannerbecomecuriouslyblendedtogether，andforalongperiodexhibittracesoftheinstinctsofeitherparent：forexample，LeRoydescribesadog，whosegreat-grandfatherwasawolf，andthisdogshowedatraceofitswildparentageonlyinoneway，bynotcominginastraightlinetohismasterwhencalled。

Domesticinstinctsaresometimesspokenofasactionswhichhavebecomeinheritedsolelyfromlong-continuedandcompulsoryhabit，butthis，I

think，isnottrue。Noonewouldeverhavethoughtofteaching，orprobablycouldhavetaught，thetumbler-pigeontotumble，anactionwhich，asI

havewitnessed，isperformedbyyoungbirds，thathaveneverseenapigeontumble。Wemaybelievethatsomeonepigeonshowedaslighttendencytothisstrangehabit，andthatthelong-continuedselectionofthebestindividualsinsuccessivegenerationsmadetumblerswhattheynoware。andnearGlasgowtherearehouse-tumblers，asIhearfromMrBrent，whichcannotflyeighteenincheshighwithoutgoingheadoverheels。Itmaybedoubtedwhetheranyonewouldhavethoughtoftrainingadogtopoint，hadnotsomeonedognaturallyshownatendencyinthisline。andthisisknownoccasionallytohappen，asIoncesawinapureterrier。Whenthefirsttendencywasoncedisplayed，methodicalselectionandtheinheritedeffectsofcompulsorytrainingineachsuccessivegenerationwouldsooncompletethework。andunconsciousselectionisstillatwork，aseachmantriestoprocure，withoutintendingtoimprovethebreed，dogswhichwillstandandhuntbest。Ontheotherhand，habitaloneinsomecaseshassufficed。noanimalismoredifficulttotamethantheyoungofthewildrabbit。scarcelyanyanimalistamerthantheyoungofthetamerabbit。butIdonotsupposethatdomesticrabbitshaveeverbeenselectedfortameness。andIpresumethatwemustattributethewholeoftheinheritedchangefromextremewildnesstoextremetameness，simplytohabitandlong-continuedcloseconfinement。

Naturalinstinctsarelostunderdomestication：aremarkableinstanceofthisisseeninthosebreedsoffowlswhichveryrarelyorneverbecome’broody，’thatis，neverwishtositontheireggs。Familiarityalonepreventsourseeinghowuniversallyandlargelythemindsofourdomesticanimalshavebeenmodifiedbydomestication。Itisscarcelypossibletodoubtthattheloveofmanhasbecomeinstinctiveinthedog。Allwolves，foxes，jackals，andspeciesofthecatgenus，whenkepttame，aremosteagertoattackpoultry，sheep，andpigs。andthistendencyhasbeenfoundincurableindogswhichhavebeenbroughthomeaspuppiesfromcountries，suchasTierradelFuegoandAustralia，wherethesavagesdonotkeepthesedomesticanimals。

Howrarely，ontheotherhand，doourciviliseddogs，evenwhenquiteyoung，requiretobetaughtnottoattackpoultry，sheep，andpigs！Nodoubttheyoccasionallydomakeanattack，andarethenbeaten。andifnotcured，theyaredestroyed。sothathabit，withsomedegreeofselection，hasprobablyconcurredincivilisingbyinheritanceourdogs。Ontheotherhand，youngchickenshavelost，whollybyhabit，thatfearofthedogandcatwhichnodoubtwasoriginallyinstinctiveinthem，inthesamewayasitissoplainlyinstinctiveinyoungpheasants，thoughrearedunderahen。Itisnotthatchickenshavelostallfear，butfearonlyofdogsandcats，forifthehengivesthedanger-chuckle，theywillrun(moreespeciallyyoungturkeys)fromunderher，andconcealthemselvesinthesurroundinggrassorthickets。andthisisevidentlydonefortheinstinctivepurposeofallowing，asweseeinwildground-birds，theirmothertoflyaway。Butthisinstinctretainedbyourchickenshasbecomeuselessunderdomestication，forthemother-henhasalmostlostbydisusethepowerofflight。

Hence，wemayconclude，thatdomesticinstinctshavebeenacquiredandnaturalinstinctshavebeenlostpartlybyhabit，andpartlybymanselectingandaccumulatingduringsuccessivegenerations，peculiarmentalhabitsandactions，whichatfirstappearedfromwhatwemustinourignorancecallanaccident。Insomecasescompulsoryhabitalonehassufficedtoproducesuchinheritedmentalchanges。inothercasescompulsoryhabithasdonenothing，andallhasbeentheresultofselection，pursuedbothmethodicallyandunconsciously。butinmostcases，probably，habitandselectionhaveactedtogether。

Weshall，perhaps，bestunderstandhowinstinctsinastateofnaturehavebecomemodifiedbyselection，byconsideringafewcases。Iwillselectonlythree，outoftheseveralwhichIshallhavetodiscussinmyfuturework，namely，theinstinctwhichleadsthecuckootolayhereggsinotherbirds’nests。theslave-makinginstinctofcertainants。andthecomb-makingpowerofthehive-bee：thesetwolatterinstinctshavegenerally，andmostjustly，beenrankedbynaturalistsasthemostwonderfulofallknowninstincts。

Itisnowcommonlyadmittedthatthemoreimmediateandfinalcauseofthecuckoo’sinstinctis，thatshelayshereggs，notdaily，butatintervalsoftwoorthreedays。sothat，ifsheweretomakeherownnestandsitonherowneggs，thosefirstlaidwouldhavetobeleftforsometimeunincubated，ortherewouldbeeggsandyoungbirdsofdifferentagesinthesamenest。Ifthiswerethecase，theprocessoflayingandhatchingmightbeinconvenientlylong，moreespeciallyasshehastomigrateataveryearlyperiod。andthefirsthatchedyoungwouldprobablyhavetobefedbythemalealone。ButtheAmericancuckooisinthispredicament。

forshemakesherownnestandhaseggsandyoungsuccessivelyhatched，allatthesametime。IthasbeenassertedthattheAmericancuckoooccasionallylayshereggsinotherbirds’nests。butIhearonthehighauthorityofDr。Brewer，thatthisisamistake。Nevertheless，Icouldgiveseveralinstancesofvariousbirdswhichhavebeenknownoccasionallytolaytheireggsinotherbirds’nests。NowletussupposethattheancientprogenitorofourEuropeancuckoohadthehabitsoftheAmericancuckoo。butthatoccasionallyshelaidanegginanotherbird’snest。Iftheoldbirdprofitedbythisoccasionalhabit，oriftheyoungweremademorevigorousbyadvantagehavingbeentakenofthemistakenmaternalinstinctofanotherbird，thanbytheirownmother’scare，encumberedasshecanhardlyfailtobebyhavingeggsandyoungofdifferentagesatthesametime。thentheoldbirdsorthefosteredyoungwouldgainanadvantage。Andanalogywouldleadmetobelieve，thattheyoungthusrearedwouldbeapttofollowbyinheritancetheoccasionalandaberranthabitoftheirmother，andintheirturnwouldbeapttolaytheireggsinotherbirds’nests，andthusbesuccessfulinrearingtheiryoung。Byacontinuedprocessofthisnature，Ibelievethatthestrangeinstinctofourcuckoocouldbe，andhasbeen，generated。Imayaddthat，accordingtoDr。Grayandtosomeotherobservers，theEuropeancuckoohasnotutterlylostallmaternalloveandcareforherownoffspring。

Theoccasionalhabitofbirdslayingtheireggsinotherbirds’nests，eitherofthesameorofadistinctspecies，isnotveryuncommonwiththeGallinaceae。andthisperhapsexplainstheoriginofasingularinstinctinthealliedgroupofostriches。Forseveralhenostriches，atleastinthecaseoftheAmericanspecies，uniteandlayfirstafeweggsinonenestandtheninanother。andthesearehatchedbythemales。Thisinstinctmayprobablybeaccountedforbythefactofthehenslayingalargenumberofeggs。but，asinthecaseofthecuckoo，atintervalsoftwoorthreedays。Thisinstinct，however，oftheAmericanostrichhasnotasyetbeenperfected。forasurprisingnumberofeggsliestrewedovertheplains，sothatinoneday’shuntingIpickedupnolessthantwentylostandwastedeggs。

Manybeesareparasitic，andalwayslaytheireggsinthenestsofbeesofotherkinds。Thiscaseismoreremarkablethanthatofthecuckoo。forthesebeeshavenotonlytheirinstinctsbuttheirstructuremodifiedinaccordancewiththeirparasitichabits。fortheydonotpossessthepollen-collectingapparatuswhichwouldbenecessaryiftheyhadtostorefoodfortheirownyoung。Somespecies，likewise，ofSphegidae(wasp-likeinsects)areparasiticonotherspecies。andM。FabrehaslatelyshowngoodreasonforbelievingthatalthoughtheTachytesnigragenerallymakesitsownburrowandstoresitwithparalysedpreyforitsownlarvaetofeedon，yetthatwhenthisinsectfindsaburrowalreadymadeandstoredbyanothersphex，ittakesadvantageoftheprize，andbecomesfortheoccasionparasitic。

Inthiscase，aswiththesupposedcaseofthecuckoo，Icanseenodifficultyinnaturalselectionmakinganoccasionalhabitpermanent，ifofadvantagetothespecies，andiftheinsectwhosenestandstoredfoodarethusfeloniouslyappropriated，benotthusexterminated。

Slave-makinginstinct。ThisremarkableinstinctwasfirstdiscoveredintheFormica(Polyerges)rufescensbyPierreHuber，abetterobservereventhanhiscelebratedfather。Thisantisabsolutelydependentonitsslaves。withouttheiraid，thespecieswouldcertainlybecomeextinctinasingleyear。Themalesandfertilefemalesdonowork。Theworkersorsterilefemales，thoughmostenergeticandcourageousincapturingslaves，donootherwork。Theyareincapableofmakingtheirownnests，oroffeedingtheirownlarvae。Whentheoldnestisfoundinconvenient，andtheyhavetomigrate，itistheslaveswhichdeterminethemigration，andactuallycarrytheirmastersintheirjaws。Soutterlyhelplessarethemasters，thatwhenHubershutupthirtyofthemwithoutaslave，butwithplentyofthefoodwhichtheylikebest，andwiththeirlarvaeandpupaetostimulatethemtowork，theydidnothing。theycouldnotevenfeedthemselves，andmanyperishedofhunger。Huberthenintroducedasingleslave(F。fusca)，andsheinstantlysettowork，fedandsavedthesurvivors。madesomecellsandtendedthelarvae，andputalltorights。Whatcanbemoreextraordinarythanthesewell-ascertainedfacts？Ifwehadnotknownofanyotherslave-makingant，itwouldhavebeenhopelesstohavespeculatedhowsowonderfulaninstinctcouldhavebeenperfected。

FormicasanguineawaslikewisefirstdiscoveredbyP。Hubertobeaslave-makingant。ThisspeciesisfoundinthesouthernpartsofEngland，anditshabitshavebeenattendedtobyMr。F。Smith，oftheBritishMuseum，towhomIammuchindebtedforinformationonthisandothersubjects。

AlthoughfullytrustingtothestatementsofHuberandMr。Smith，Itriedtoapproachthesubjectinascepticalframeofmind，asanyonemaywellbeexcusedfordoubtingthetruthofsoextraordinaryandodiousaninstinctasthatofmakingslaves。HenceIwillgivetheobservationswhichIhavemademyselfmade，insomelittledetail。IopenedfourteennestsofF。

sanguinea，andfoundafewslavesinall。Malesandfertilefemalesoftheslave-speciesarefoundonlyintheirownpropercommunities，andhaveneverbeenobservedinthenestsofF。sanguinea。Theslavesareblackandnotabovehalfthesizeoftheirredmasters，sothatthecontrastintheirappearanceisverygreat。Whenthenestisslightlydisturbed，theslavesoccasionallycomeout，andliketheirmastersaremuchagitatedanddefendtheirnest：whenthenestismuchdisturbedandthelarvaeandpupaeareexposed，theslavesworkenergeticallywiththeirmastersincarryingthemawaytoaplaceofsafety。Hence，itisclear，thattheslavesfeelquiteathome。DuringthemonthsofJuneandJuly，onthreesuccessiveyears，IhavewatchedformanyhoursseveralnestsinSurreyandSussex，andneversawaslaveeitherleaveorenteranest。As，duringthesemonths，theslavesareveryfewinnumber，Ithoughtthattheymightbehavedifferentlywhenmorenumerous。butMr。SmithinformsmethathehaswatchedthenestsatvarioushoursduringMay，JuneandAugust，bothinSurreyandHampshire，andhasneverseentheslaves，thoughpresentinlargenumbersinAugust，eitherleaveorenterthenest。Henceheconsidersthemasstrictlyhouseholdslaves。Themasters，ontheotherhand，maybeconstantlyseenbringinginmaterialsforthenest，andfoodofallkinds。Duringthepresentyear，however，inthemonthofJuly，Icameacrossacommunitywithanunusuallylargestockofslaves，andIobservedafewslavesmingledwiththeirmastersleavingthenest，andmarchingalongthesameroadtoatallScotch-fir-tree，twenty-fiveyardsdistant，whichtheyascendedtogether，probablyinsearchofaphidesorcocci。AccordingtoHuber，whohadampleopportunitiesforobservation，inSwitzerlandtheslaveshabituallyworkwiththeirmastersinmakingthenest，andtheyaloneopenandclosethedoorsinthemorningandevening。and，asHuberexpresslystates，theirprincipalofficeistosearchforaphides。Thisdifferenceintheusualhabitsofthemastersandslavesinthetwocountries，probablydependsmerelyontheslavesbeingcapturedingreaternumbersinSwitzerlandthaninEngland。

OnedayIfortunatelychancedtowitnessamigrationfromonenesttoanother，anditwasamostinterestingspectacletobeholdthemasterscarefullycarrying，asHuberhasdescribed，theirslavesintheirjaws。

Anotherdaymyattentionwasstruckbyaboutascoreoftheslave-makershauntingthesamespot，andevidentlynotinsearchoffood。theyapproachedandwerevigorouslyrepulsedbyanindependentcommunityoftheslavespecies(F。fusca)。sometimesasmanyasthreeoftheseantsclingingtothelegsoftheslave-makingF。sanguinea。Thelatterruthlesslykilledtheirsmallopponents，andcarriedtheirdeadbodiesasfoodtotheirnest，twenty-nineyardsdistant。buttheywerepreventedfromgettinganypupaetorearasslaves。IthendugupasmallparcelofthepupaeofF。fuscafromanothernest，andputthemdownonabarespotneartheplaceofcombat。theywereeagerlyseized，andcarriedoffbythetyrants，whoperhapsfanciedthat，afterall，theyhadbeenvictoriousintheirlatecombat。

AtthesametimeIlaidonthesameplaceasmallparcelofthepupaeofanotherspecies，F。flava，withafewoftheselittleyellowantsstillclingingtothefragmentsofthenest。Thisspeciesissometimes，thoughrarely，madeintoslaves，ashasbeendescribedbyMrSmith。Althoughsosmallaspecies，itisverycourageous，andIhaveseenitferociouslyattackotherants。InoneinstanceIfoundtomysurpriseanindependentcommunityofF。flavaunderastonebeneathanestoftheslave-makingF。sanguinea。andwhenIhadaccidentallydisturbedbothnests，thelittleantsattackedtheirbigneighbourswithsurprisingcourage。NowIwascurioustoascertainwhetherF。sanguineacoulddistinguishthepupaeofF。fusca，whichtheyhabituallymakeintoslaves，fromthoseofthelittleandfuriousF。flava，whichtheyrarelycapture，anditwasevidentthattheydidatoncedistinguishthem：forwehaveseenthattheyeagerlyandinstantlyseizedthepupaeofF。fusca，whereastheyweremuchterrifiedwhentheycameacrossthepupae，oreventheearthfromthenestofF。flava，andquicklyranaway。butinaboutaquarterofanhour，shortlyafterallthelittleyellowantshadcrawledaway，theytookheartandcarriedoffthepupae。

OneeveningIvisitedanothercommunityofF。sanguinea，andfoundanumberoftheseantsenteringtheirnest，carryingthedeadbodiesofF。

fusca(showingthatitwasnotamigration)andnumerouspupae。Itracedthereturningfileburthenedwithbooty，foraboutfortyyards，toaverythickclumpofheath。whenceIsawthelastindividualofF。sanguineaemerge，carryingapupa。butIwasnotabletofindthedesolatednestinthethickheath。Thenest，however，musthavebeencloseathand，fortwoorthreeindividualsofF。fuscawererushingaboutinthegreatestagitation，andonewasperchedmotionlesswithitsownpupainitsmouthonthetopofasprayofheathoveritsravagedhome。

Sucharethefacts，thoughtheydidnotneedconfirmationbyme，inregardtothewonderfulinstinctofmakingslaves。LetitbeobservedwhatacontrasttheinstinctivehabitsofF。sanguineapresentwiththoseoftheF。rufescens。Thelatterdoesnotbuilditsownnest，doesnotdetermineitsownmigrations，doesnotcollectfoodforitselforitsyoung，andcannotevenfeeditself：itisabsolutelydependentonitsnumerousslaves。

Formicasanguinea，ontheotherhand，possessesmuchfewerslaves，andintheearlypartofthesummerextremelyfew。Themastersdeterminewhenandwhereanewnestshallbeformed，andwhentheymigrate，themasterscarrytheslaves。BothinSwitzerlandandEnglandtheslavesseemtohavetheexclusivecareofthelarvae，andthemastersalonegoonslave-makingexpeditions。InSwitzerlandtheslavesandmastersworktogether，makingandbringingmaterialsforthenest：both，butchieflytheslaves，tend，andmilkasitmaybecalled，theiraphides。andthusbothcollectfoodforthecommunity。InEnglandthemastersaloneusuallyleavethenesttocollectbuildingmaterialsandfoodforthemselves，theirslavesandlarvae。SothatthemastersinthiscountryreceivemuchlessservicefromtheirslavesthantheydoinSwitzerland。

BywhatstepstheinstinctofF。sanguineaoriginatedIwillnotpretendtoconjecture。Butasants，whicharenotslave-makers，will，asIhaveseen，carryoffpupaeofotherspecies，ifscatteredneartheirnests，itispossiblethatpupaeoriginallystoredasfoodmightbecomedeveloped。

andtheantsthusunintentionallyrearedwouldthenfollowtheirproperinstincts，anddowhatworktheycould。Iftheirpresenceprovedusefultothespecieswhichhadseizedthemifitweremoreadvantageoustothisspeciestocaptureworkersthantoprocreatethemthehabitofcollectingpupaeoriginallyforfoodmightbynaturalselectionbestrengthenedandrenderedpermanentfortheverydifferentpurposeofraisingslaves。Whentheinstinctwasonceacquired，ifcarriedouttoamuchlessextenteventhaninourBritishF。sanguinea，which，aswehaveseen，islessaidedbyitsslavesthanthesamespeciesinSwitzerland，IcanseenodifficultyinnaturalselectionincreasingandmodifyingtheinstinctalwayssupposingeachmodificationtobeofusetothespeciesuntilanantwasformedasabjectlydependentonitsslavesasistheFormicarufescens。

Cell-makinginstinctoftheHive-Bee。Iwillnothereenteronminutedetailsonthissubject，butwillmerelygiveanoutlineoftheconclusionsatwhichIhavearrived。Hemustbeadullmanwhocanexaminetheexquisitestructureofacomb，sobeautifullyadaptedtoitsend，withoutenthusiasticadmiration。Wehearfrommathematiciansthatbeeshavepracticallysolvedareconditeproblem，andhavemadetheircellsofthepropershapetoholdthegreatestpossibleamountofhoney，withtheleastpossibleconsumptionofpreviouswaxintheirconstruction。Ithasbeenremarkedthataskilfulworkman，withfittingtoolsandmeasures，wouldfinditverydifficulttomakecellsofwaxofthetrueform，thoughthisisperfectlyeffectedbyacrowdofbeesworkinginadarkhive。Grantwhateverinstinctsyouplease，anditseemsatfirstquiteinconceivablehowtheycanmakeallthenecessaryanglesandplanes，orevenperceivewhentheyarecorrectlymade。Butthedifficultyisnotnearlysogreatasitatfirstappears：

allthisbeautifulworkcanbeshown，Ithink，tofollowfromafewverysimpleinstincts。

IwasledtoinvestigatethissubjectbyMr。Waterhouse，whohasshownthattheformofthecellstandsincloserelationtothepresenceofadjoiningcells。andthefollowingviewmay，perhaps，beconsideredonlyasamodificationofthistheory。Letuslooktothegreatprincipleofgradation，andseewhetherNaturedoesnotrevealtoushermethodofwork。Atoneendofashortserieswehavehumble-bees，whichusetheiroldcocoonstoholdhoney，sometimesaddingtothemshorttubesofwax，andlikewisemakingseparateandveryirregularroundedcellsofwax。Attheotherendoftheserieswehavethecellsofthehive-bee，placedinadoublelayer：eachcell，asiswellknow，isanhexagonalprism，withthebasaledgesofitssixsidesbevelledsoastojoinontoapyramid，formedofthreerhombs。

Theserhombshavecertainangles，andthethreewhichformthepyramidalbaseofasinglecellononesideofthecomb，enterintothecompositionofthebasesofthreeadjoiningcellsontheoppositeside。Intheseriesbetweentheextremeperfectionofthecellsofthehive-beeandthesimplicityofthoseofthehumble-bee，wehavethecellsoftheMexicanMeliponadomestica，carefullydescribedandfiguredbyPierreHuber。TheMeliponaitselfisintermediateinstructurebetweenthehiveandhumblebee，butmorenearlyrelatedtothelatter：itformsanearlyregularwaxencombofcylindricalcells，inwhichtheyoungarehatched，and，inaddition，somelargecellsofwaxforholdinghoney。Theselattercellsarenearlysphericalandofnearlyequalsizes，andareaggregatedintoanirregularmass。Buttheimportantpointtonotice，isthatthesecellsarealwaysmadeatthatdegreeofnearnesstoeachother，thattheywouldhaveintersectedorbrokenintoeachother，ifthesphereshadbeencompleted。butthisisneverpermitted，thebeesbuildingperfectlyflatwallsofwaxbetweenthesphereswhichthustendtointersect。Henceeachcellconsistsofanoutersphericalportionandoftwo，three，ormoreperfectlyflatsurfaces，accordingasthecelladjoinstwo，threeormoreothercells。Whenonecellcomesintocontactwiththreeothercells，which，fromthespheresbeingnearlyofthesamesize，isveryfrequentlyandnecessarilythecase，thethreeflatsurfacesareunitedintoapyramid。andthispyramid，asHuberhasremarked，ismanifestlyagrossimitationofthethree-sidedpyramidalbasisofthecellofthehive-bee。Asinthecellsofthehive-bee，sohere，thethreeplanesurfacesinanyonecellnecessarilyenterintotheconstructionofthreeadjoiningcells。ItisobviousthattheMeliponasaveswaxbythismannerofbuilding。fortheflatwallsbetweentheadjoiningcellsarenotdouble，butareofthesamethicknessastheoutersphericalportions，andyeteachflatportionformsapartoftwocells。

Reflectingonthiscase，itoccurredtomethatiftheMeliponahadmadeitsspheresatsomegivendistancefromeachother，andhadmadethemofequalsizesandhadarrangedthemsymmetricallyinadoublelayer，theresultingstructurewouldprobablyhavebeenasperfectasthecombofthehive-bee。AccordinglyIwrotetoProfessorMiller，ofCambridge，andthisgeometerhaskindlyreadoverthefollowingstatement，drawnupfromhisinformation，andtellsmethatitisstrictlycorrect：-

Ifanumberofequalspheresbedescribedwiththeircentresplacedintwoparallellayers。withthecentreofeachsphereatthedistanceofradiusX/sqrt[2]orradiusX1。41421(oratsomelesserdistance)，fromthecentresofthesixsurroundingspheresinthesamelayer。andatthesamedistancefromthecentresoftheadjoiningspheresintheotherandparallellayer。then，ifplanesofintersectionbetweentheseveralspheresinbothlayersbeformed，therewillresultadoublelayerofhexagonalprismsunitedtogetherbypyramidalbasesformedofthreerhombs。andtherhombsandthesidesofthehexagonalprismswillhaveeveryangleidenticallythesamewiththebestmeasurementswhichhavebeenmadeofthecellsofthehive-bee。

HencewemaysafelyconcludethatifwecouldslightlymodifytheinstinctsalreadypossessedbytheMelipona，andinthemselvesnotverywonderful，thisbeewouldmakeastructureaswonderfullyperfectasthatofthehive-bee。

WemustsupposetheMeliponatomakehercellstrulyspherical，andofequalsizes。andthiswouldnotbeverysurprising，seeingthatshealreadydoessotoacertainextent，andseeingwhatperfectlycylindricalburrowsinwoodmanyinsectscanmake，apparentlybyturningroundonafixedpoint。

WemustsupposetheMeliponatoarrangehercellsinlevellayers，asshealreadydoeshercylindricalcells。andwemustfurthersuppose，andthisisthegreatestdifficulty，thatshecansomehowjudgeaccuratelyatwhatdistancetostandfromherfellow-labourerswhenseveralaremakingtheirspheres。butsheisalreadysofarenabledtojudgeofdistance，thatshealwaysdescribesherspheressoastointersectlargely。andthensheunitesthepointsofintersectionbyperfectlyflatsurfaces。Wehavefurthertosuppose，butthisisnodifficulty，thatafterhexagonalprismshavebeenformedbytheintersectionofadjoiningspheresinthesamelayer，shecanprolongthehexagontoanylengthrequisitetoholdthestockofhoney。inthesamewayastherudehumble-beeaddscylindersofwaxtothecircularmouthsofheroldcocoons。Bysuchmodificationsofinstinctsinthemselvesnotverywonderful，hardlymorewonderfulthanthosewhichguideabirdtomakeitsnest，Ibelievethatthehive-beehasacquired，throughnaturalselection，herinimitablearchitecturalpowers。

Butthistheorycanbetestedbyexperiment。FollowingtheexampleofMrTegetmeier，Iseparatedtwocombs，andputbetweenthemalong，thick，squarestripofwax：thebeesinstantlybegantoexcavateminutecircularpitsinit。andastheydeepenedtheselittlepits，theymadethemwiderandwideruntiltheywereconvertedintoshallowbasins，appearingtotheeyeperfectlytrueorpartsofasphere，andofaboutthediameterofacell。Itwasmostinterestingtometoobservethatwhereverseveralbeeshadbeguntoexcavatethesebasinsneartogether，theyhadbeguntheirworkatsuchadistancefromeachother，thatbythetimethebasinshadacquiredtheabovestatedwidth(i。e。aboutthewidthofanordinarycell)，andwereindepthaboutonesixthofthediameterofthesphereofwhichtheyformedapart，therimsofthebasinsintersectedorbrokeintoeachother。Assoonasthisoccurred，thebeesceasedtoexcavate，andbegantobuildupflatwallsofwaxonthelinesofintersectionbetweenthebasins，sothateachhexagonalprismwasbuiltuponthefestoonededgeofasmoothbasin，insteadofonthestraightedgesofathree-sidedpyramidasinthecaseofordinarycells。

Ithenputintothehive，insteadofathick，squarepieceofwax，athinandnarrow，knife-edgedridge，colouredwithvermilion。Thebeesinstantlybeganonbothsidestoexcavatelittlebasinsneartoeachother，inthesamewayasbefore。buttheridgeofwaxwassothin，thatthebottomsofthebasins，iftheyhadbeenexcavatedtothesamedepthasintheformerexperiment，wouldhavebrokenintoeachotherfromtheoppositesides。

Thebees，however，didnotsufferthistohappen，andtheystoppedtheirexcavationsinduetime。sothatthebasins，assoonastheyhadbeenalittledeepened，cametohaveflatbottoms。andtheseflatbottoms，formedbythinlittleplatesofthevermilionwaxhavingbeenleftungnawed，weresituated，asfarastheeyecouldjudge，exactlyalongtheplanesofimaginaryintersectionbetweenthebasinsontheoppositesidesoftheridgeofwax。

Inparts，onlylittlebits，inotherparts，largeportionsofarhombicplatehadbeenleftbetweentheopposedbasins，butthework，fromtheunnaturalstateofthings，hadnotbeenneatlyperformed。Thebeesmusthaveworkedatverynearlythesamerateontheoppositesideoftheridgeofvermilionwax，astheycircularlygnawedawayanddeepenedthebasinsonbothsides，inordertohavesucceededinthusleavingflatplatesbetweenthebasins，bystoppingworkalongtheintermediateplanesorplanesofintersection。

Consideringhowflexiblethinwaxis，Idonotseethatthereisanydifficultyinthebees，whilstatworkonthetwosidesofastripofwax，perceivingwhentheyhavegnawedthewaxawaytotheproperthinness，andthenstoppingtheirwork。Inordinarycombsithasappearedtomethatthebeesdonotalwayssucceedinworkingatexactlythesameratefromtheoppositesides。forIhavenoticedhalf-completedrhombsatthebaseofajust-commencedcell，whichwereslightlyconcaveononeside，whereIsupposethatthebeeshadexcavatedtooquickly，andconvexontheopposedside，wherethebeeshadworkedlessquickly。Inonewell-markedinstance，Iputthecombbackintothehiveandallowedthebeestogoonworkingforashorttimeandagainexaminedthecell，andIfoundthattherhombicplatehadbeencompleted，andhadbecomeperfectlyflat：itwasabsolutelyimpossible，fromtheextremethinnessofthelittlerhombicplate，thattheycouldhaveaffectedthisbygnawingawaytheconvexside。

andIsuspectthatthebeesinsuchcasesstandintheopposedcellsandpushandbendtheductileandwarmwax(whichasIhavetriediseasilydone)intoitsproperintermediateplane，andthusflattenit。

Fromtheexperimentoftheridgeofvermilionwax，wecanclearlyseethatifthebeesweretobuildforthemselvesathinwallofwax，theycouldmaketheircellsofthepropershape，bystandingattheproperdistancefromeachother，byexcavatingatthesamerate，andbyendeavouringtomakeequalsphericalhollows，butneverallowingthespherestobreakintoeachother。Nowbees，asmaybeclearlyseenbyexaminingtheedgeofagrowingcomb，domakearough，circumferentialwallorrimallroundthecomb。andtheygnawintothisfromtheoppositesides，alwaysworkingcircularlyastheydeepeneachcell。Theydonotmakethewholethree-sidedpyramidalbaseofanyonecellatthesametime，butonlytheonerhombicplatewhichstandsontheextremegrowingmargin，orthetwoplates，asthecasemaybe。andtheynevercompletetheupperedgesoftherhombicplates，untilthehexagonalwallsarecommenced。SomeofthesestatementsdifferfromthosemadebythejustlycelebratedelderHuber，butIamconvincedoftheiraccuracy。andifIhadspace，Icouldshowthattheyareconformablewithmytheory。

Huber’sstatementthattheveryfirstcellisexcavatedoutofalittleparallel-sidedwallofwax，isnot，asfarasIhaveseen，strictlycorrect。

thefirstcommencementhavingalwaysbeenalittlehoodofwax。butIwillnothereenteronthesedetails。Weseehowimportantapartexcavationplaysintheconstructionofthecells。butitwouldbeagreaterrortosupposethatthebeescannotbuilduparoughwallofwaxintheproperpositionthatis，alongtheplaneofintersectionbetweentwoadjoiningspheres。Ihaveseveralspecimensshowingclearlythattheycandothis。

Evenintherudecircumferentialrimorwallofwaxroundagrowingcomb，flexuresmaysometimesbeobserved，correspondinginpositiontotheplanesoftherhombicbasalplatesoffuturecells。Buttheroughwallofwaxhasineverycasetobefinishedoff，bybeinglargelygnawedawayonbothsides。Themannerinwhichthebeesbuildiscurious。theyalwaysmakethefirstroughwallfromtentotwentytimesthickerthantheexcessivelythinfinishedwallofthecell，whichwillultimatelybeleft。Weshallunderstandhowtheywork，bysupposingmasonsfirsttopileupabroadridgeofcement，andthentobegincuttingitawayequallyonbothsidesneartheground，tillasmooth，verythinwallisleftinthemiddle。themasonsalwayspilingupthecut-awaycement，andaddingfreshcement，onthesummitoftheridge。Weshallthushaveathinwallsteadilygrowingupward。butalwayscrownedbyagiganticcoping。Fromallthecells，boththosejustcommencedandthosecompleted，beingthuscrownedbyastrongcopingofwax，thebeescanclusterandcrawloverthecombwithoutinjuringthedelicatehexagonalwalls，whichareonlyaboutonefour-hundredthofaninchinthickness。theplatesofthepyramidalbasisbeingabouttwiceasthick。Bythissingularmannerofbuilding，strengthiscontinuallygiventothecomb，withtheutmostultimateeconomyofwax。

Itseemsatfirsttoaddtothedifficultyofunderstandinghowthecellsaremade，thatamultitudeofbeesallworktogether。onebeeafterworkingashorttimeatonecellgoingtoanother，sothat，asHuberhasstated，ascoreofindividualsworkevenatthecommencementofthefirstcell。Iwasablepracticallytoshowthisfact，bycoveringtheedgesofthehexagonalwallsofasinglecell，ortheextrememarginofthecircumferentialrimofagrowingcomb，withanextremelythinlayerofmeltedvermilionwax。andIinvariablyfoundthatthecolourwasmostdelicatelydiffusedbythebeesasdelicatelyasapaintercouldhavedonewithhisbrushbyatomsofthecolouredwaxhavingbeentakenfromthespotonwhichithadbeenplaced，andworkedintothegrowingedgesofthecellsallround。

Theworkofconstructionseemstobeasortofbalancestruckbetweenmanybees，allinstinctivelystandingatthesamerelativedistancefromeachother，alltryingtosweepequalspheres，andthenbuildingup，orleavingungnawed，theplanesofintersectionbetweenthesespheres。Itwasreallycurioustonoteincasesofdifficulty，aswhentwopiecesofcombmetatanangle，howoftenthebeeswouldentirelypulldownandrebuildindifferentwaysthesamecell，sometimesrecurringtoashapewhichtheyhadatfirstrejected。

Whenbeeshaveaplaceonwhichtheycanstandintheirproperpositionsforworking，forinstance，onaslipofwood，placeddirectlyunderthemiddleofacombgrowingdownwardssothatthecombhastobebuiltoveronefaceoftheslipinthiscasethebeescanlaythefoundationsofonewallofanewhexagon，initsstrictlyproperplace，projectingbeyondtheothercompletedcells。Itsufficesthatthebeesshouldbeenabledtostandattheirproperrelativedistancesfromeachotherandfromthewallsofthelastcompletedcells，andthen，bystrikingimaginaryspheres，theycanbuildupawallintermediatebetweentwoadjoiningspheres。but，asfarasIhaveseen，theynevergnawawayandfinishofftheanglesofacelltillalargepartbothofthatcellandoftheadjoiningcellshasbeenbuilt。Thiscapacityinbeesoflayingdownundercertaincircumstancesaroughwallinitsproperplacebetweentwojust-commencedcells，isimportant，asitbearsonafact，whichseemsatfirstquitesubversiveoftheforegoingtheory。namely，thatthecellsontheextrememarginofwasp-combsaresometimesstrictlyhexagonal。butIhavenotspaceheretoenteronthissubject。Nordoesthereseemtomeanygreatdifficultyinasingleinsect(asinthecaseofaqueen-wasp)makinghexagonalcells，ifsheworkalternatelyontheinsideandoutsideoftwoorthreecellscommencedatthesametime，alwaysstandingattheproperrelativedistancefromthepartsofthecellsjustbegun，sweepingspheresorcylinders，andbuildingupintermediateplanes。Itisevenconceivablethataninsectmight，byfixingonapointatwhichtocommenceacell，andthenmovingoutside，firsttoonepoint，andthentofiveotherpoints，attheproperrelativedistancesfromthecentralpointandfromeachother，striketheplanesofintersection，andsomakeanisolatedhexagon：butIamnotawarethatanysuchcasehasbeenobserved。norwouldanygoodbederivedfromasinglehexagonbeingbuilt，asinitsconstructionmorematerialswouldberequiredthanforacylinder。

Asnaturalselectionactsonlybytheaccumulationofslightmodificationsofstructureorinstinct，eachprofitabletotheindividualunderitsconditionsoflife，itmayreasonablybeasked，howalongandgraduatedsuccessionofmodifiedarchitecturalinstincts，alltendingtowardsthepresentperfectplanofconstruction，couldhaveprofitedtheprogenitorsofthehive-bee？

Ithinktheanswerisnotdifficult：itisknownthatbeesareoftenhardpressedtogetsufficientnectar。andIaminformedbyMr。Tegetmeierthatithasbeenexperimentallyfoundthatnolessthanfromtwelvetofifteenpoundsofdrysugarareconsumedbyahiveofbeesforthesecretionofeachpoundofwax。sothataprodigiousquantityoffluidnectarmustbecollectedandconsumedbythebeesinahiveforthesecretionofthewaxnecessaryfortheconstructionoftheircombs。Moreover，manybeeshavetoremainidleformanydaysduringtheprocessofsecretion。Alargestoreofhoneyisindispensabletosupportalargestockofbeesduringthewinter。

andthesecurityofthehiveisknownmainlytodependonalargenumberofbeesbeingsupported。Hencethesavingofwaxbylargelysavinghoneymustbeamostimportantelementofsuccessinanyfamilyofbees。Ofcoursethesuccessofanyspeciesofbeemaybedependentonthenumberofitsparasitesorotherenemies，oronquitedistinctcauses，andsobealtogetherindependentofthequantityofhoneywhichthebeescouldcollect。Butletussupposethatthislattercircumstancedetermined，asitprobablyoftendoesdetermine，thenumbersofahumble-beewhichcouldexistinacountry。andletusfurthersupposethatthecommunitylivedthroughoutthewinter，andconsequentlyrequiredastoreofhoney：therecaninthiscasebenodoubtthatitwouldbeanadvantagetoourhumble-bee，ifaslightmodificationofherinstinctledhertomakeherwaxencellsneartogether，soastointersectalittle。forawallincommoneventotwoadjoiningcells，wouldsavesomelittlewax。Henceitwouldcontinuallybemoreandmoreadvantageoustoourhumble-bee，ifsheweretomakehercellsmoreandmoreregular，nearertogether，andaggregatedintoamass，likethecellsoftheMelipona。forinthiscasealargepartoftheboundingsurfaceofeachcellwouldservetoboundothercells，andmuchwaxwouldbesaved。Again，fromthesamecause，itwouldbeadvantageoustotheMelipona，ifsheweretomakehercellsclosertogether，andmoreregularineverywaythanatpresent。forthen，aswehaveseen，thesphericalsurfaceswouldwhollydisappear，andwouldallbereplacedbyplanesurfaces。andtheMeliponawouldmakeacombasperfectasthatofthehive-bee。Beyondthisstageofperfectioninarchitecture，naturalselectioncouldnotlead。

forthecombofthehive-bee，asfaraswecansee，isabsolutelyperfectineconomisingwax。

Thus，asIbelieve，themostwonderfulofallknowninstincts，thatofthehive-bee，canbeexplainedbynaturalselectionhavingtakenadvantageofnumerous，successive，slightmodificationsofsimplerinstincts。naturalselectionhavingbyslowdegrees，moreandmoreperfectly，ledthebeestosweepequalspheresatagivendistancefromeachotherinadoublelayer，andtobuildupandexcavatethewaxalongtheplanesofintersection。

Thebees，ofcourse，nomoreknowingthattheyswepttheirspheresatoneparticulardistancefromeachother，thantheyknowwhataretheseveralanglesofthehexagonalprismsandofthebasalrhombicplates。Themotivepoweroftheprocessofnaturalselectionhavingbeeneconomyofwax。thatindividualswarmwhichwastedleasthoneyinthesecretionofwax，havingsucceededbest，andhavingtransmittedbyinheritanceitsnewlyacquiredeconomicalinstincttonewswarms，whichintheirturnwillhavehadthebestchanceofsucceedinginthestruggleforexistence。

Nodoubtmanyinstinctsofverydifficultexplanationcouldbeopposedtothetheoryofnaturalselection，cases，inwhichwecannotseehowaninstinctcouldpossiblyhaveoriginated。cases，inwhichnointermediategradationsareknowntoexist。casesofinstinctofapparentlysuchtriflingimportance，thattheycouldhardlyhavebeenactedonbynaturalselection。

casesofinstinctsalmostidenticallythesameinanimalssoremoteinthescaleofnature，thatwecannotaccountfortheirsimilaritybyinheritancefromacommonparent，andmustthereforebelievethattheyhavebeenacquiredbyindependentactsofnaturalselection。Iwillnothereenterontheseseveralcases，butwillconfinemyselftoonespecialdifficulty，whichatfirstappearedtomeinsuperable，andactuallyfataltomywholetheory。

Ialludetotheneutersorsterilefemalesininsect-communities：fortheseneutersoftendifferwidelyininstinctandinstructurefromboththemalesandfertilefemales，andyet，frombeingsterile，theycannotpropagatetheirkind。

Thesubjectwelldeservestobediscussedatgreatlength，butIwillheretakeonlyasinglecase，thatofworkingorsterileants。Howtheworkershavebeenrenderedsterileisadifficulty。butnotmuchgreaterthanthatofanyotherstrikingmodificationofstructure。foritcanbeshownthatsomeinsectsandotherarticulateanimalsinastateofnatureoccasionallybecomesterile。andifsuchinsectshadbeensocial，andithadbeenprofitabletothecommunitythatanumbershouldhavebeenannuallyborncapableofwork，butincapableofprocreation，Icanseenoverygreatdifficultyinthisbeingeffectedbynaturalselection。ButImustpassoverthispreliminarydifficulty。Thegreatdifficultyliesintheworkingantsdifferingwidelyfromboththemalesandthefertilefemalesinstructure，asintheshapeofthethoraxandinbeingdestituteofwingsandsometimesofeyes，andininstinct。Asfarasinstinctaloneisconcerned，theprodigiousdifferenceinthisrespectbetweentheworkersandtheperfectfemales，wouldhavebeenfarbetterexemplifiedbythehive-bee。Ifaworkingantorotherneuterinsecthadbeenananimalintheordinarystate，Ishouldhaveunhesitatinglyassumedthatallitscharactershadbeenslowlyacquiredthroughnaturalselection。namely，byanindividualhavingbeenbornwithsomeslightprofitablemodificationofstructure，thisbeinginheritedbyitsoffspring，whichagainvariedandwereagainselected，andsoonwards。

Butwiththeworkingantwehaveaninsectdifferinggreatlyfromitsparents，yetabsolutelysterile。sothatitcouldneverhavetransmittedsuccessivelyacquiredmodificationsofstructureorinstincttoitsprogeny。Itmaywellbeaskedhowisitpossibletoreconcilethiscasewiththetheoryofnaturalselection？

First，letitberememberedthatwehaveinnumerableinstances，bothinourdomesticproductionsandinthoseinastateofnature，ofallsortsofdifferencesofstructurewhichhavebecomecorrelatedtocertainages，andtoeithersex。Wehavedifferencescorrelatednotonlytoonesex，buttothatshortperiodalonewhenthereproductivesystemisactive，asinthenuptialplumageofmanybirds，andinthehookedjawsofthemalesalmon。Wehaveevenslightdifferencesinthehornsofdifferentbreedsofcattleinrelationtoanartificiallyimperfectstateofthemalesex。foroxenofcertainbreedshavelongerhornsthaninotherbreeds，incomparisonwiththehornsofthebullsorcowsofthesesamebreeds。

HenceIcanseenorealdifficultyinanycharacterhavingbecomecorrelatedwiththesterileconditionofcertainmembersofinsect-communities：thedifficultyliesinunderstandinghowsuchcorrelatedmodificationsofstructurecouldhavebeenslowlyaccumulatedbynaturalselection。

Thisdifficulty，thoughappearinginsuperable，islessened，or，asI

believe，disappears，whenitisrememberedthatselectionmaybeappliedtothefamily，aswellastotheindividual，andmaythusgainthedesiredend。Thus，awell-flavouredvegetableiscooked，andtheindividualisdestroyed。butthehorticulturistsowsseedsofthesamestock，andconfidentlyexpectstogetnearlythesamevariety。breedersofcattlewishthefleshandfattobewellmarbledtogether。theanimalhasbeenslaughtered，butthebreedergoeswithconfidencetothesamefamily。Ihavesuchfaithinthepowersofselection，thatIdonotdoubtthatabreedofcattle，alwaysyieldingoxenwithextraordinarilylonghorns，couldbeslowlyformedbycarefullywatchingwhichindividualbullsandcows，whenmatched，producedoxenwiththelongesthorns。andyetnooneoxcouldeverhavepropagateditskind。ThusIbelieveithasbeenwithsocialinsects：aslightmodificationofstructure，orinstinct，correlatedwiththesterileconditionofcertainmembersofthecommunity，hasbeenadvantageoustothecommunity：consequentlythefertilemalesandfemalesofthesamecommunityflourished，andtransmittedtotheirfertileoffspringatendencytoproducesterilemembershavingthesamemodification。AndIbelievethatthisprocesshasbeenrepeated，untilthatprodigiousamountofdifferencebetweenthefertileandsterilefemalesofthesamespecieshasbeenproduced，whichweseeinmanysocialinsects。

Butwehavenotasyettouchedontheclimaxofthedifficulty。namely，thefactthattheneutersofseveralantsdiffer，notonlyfromthefertilefemalesandmales，butfromeachother，sometimestoanalmostincredibledegree，andarethusdividedintotwooreventhreecastes。Thecastes，moreover，donotgenerallygraduateintoeachother，butareperfectlywelldefined。beingasdistinctfromeachother，asareanytwospeciesofthesamegenus，orratherasanytwogeneraofthesamefamily。ThusinEciton，thereareworkingandsoldierneuters，withjawsandinstinctsextraordinarilydifferent：inCryptocerus，theworkersofonecastealonecarryawonderfulsortofshieldontheirheads，theuseofwhichisquiteunknown：intheMexicanMyrmecocystus，theworkersofonecasteneverleavethenest。theyarefedbytheworkersofanothercaste，andtheyhaveanenormouslydevelopedabdomenwhichsecretesasortofhoney，supplyingtheplaceofthatexcretedbytheaphides，orthedomesticcattleastheymaybecalled，whichourEuropeanantsguardorimprison。

ItwillindeedbethoughtthatIhaveanoverweeningconfidenceintheprincipleofnaturalselection，whenIdonotadmitthatsuchwonderfulandwell-establishedfactsatonceannihilatemytheory。Inthesimplercaseofneuterinsectsallofonecasteorofthesamekind，whichhavebeenrenderedbynaturalselection，asIbelievetobequitepossible，differentfromthefertilemalesandfemales，inthiscase，wemaysafelyconcludefromtheanalogyofordinaryvariations，thateachsuccessive，slight，profitablemodificationdidnotprobablyatfirstappearinalltheindividualneutersinthesamenest，butinafewalone。andthatbythelong-continuedselectionofthefertileparentswhichproducedmostneuterswiththeprofitablemodification，alltheneutersultimatelycametohavethedesiredcharacter。Onthisviewweoughtoccasionallytofindneuter-insectsofthesamespecies，inthesamenest，presentinggradationsofstructure。andthiswedofind，evenoften，consideringhowfewneuter-insectsoutofEuropehavebeencarefullyexamined。MrF。SmithhasshownhowsurprisinglytheneutersofseveralBritishantsdifferfromeachotherinsizeandsometimesincolour。andthattheextremeformscansometimesbeperfectlylinkedtogetherbyindividualstakenoutofthesamenest：Ihavemyselfcomparedperfectgradationsofthiskind。Itoftenhappensthatthelargerorthesmallersizedworkersarethemostnumerous。orthatbothlargeandsmallarenumerous，withthoseofanintermediatesizescantyinnumbers。

Formicaflavahaslargerandsmallerworkers，withsomeofintermediatesize。and，inthisspecies，asMrF。Smithhasobserved，thelargerworkershavesimpleeyes(ocelli)，whichthoughsmallcanbeplainlydistinguished，whereasthesmallerworkershavetheirocellirudimentary。Havingcarefullydissectedseveralspecimensoftheseworkers，Icanaffirmthattheeyesarefarmorerudimentaryinthesmallerworkersthancanbeaccountedformerelybytheirproportionallylessersize。andIfullybelieve，thoughIdarenotassertsopositively，thattheworkersofintermediatesizehavetheirocelliinanexactlyintermediatecondition。Sothatweherehavetwobodiesofsterileworkersinthesamenest，differingnotonlyinsize，butintheirorgansofvision，yetconnectedbysomefewmembersinanintermediatecondition。Imaydigressbyadding，thatifthesmallerworkershadbeenthemostusefultothecommunity，andthosemalesandfemaleshadbeencontinuallyselected，whichproducedmoreandmoreofthesmallerworkers，untilalltheworkershadcometobeinthiscondition。

weshouldthenhavehadaspeciesofantwithneutersverynearlyinthesameconditionwiththoseofMyrmica。FortheworkersofMyrmicahavenotevenrudimentsofocelli，thoughthemaleandfemaleantsofthisgenushavewell-developedocelli。

Imaygiveoneothercase：soconfidentlydidIexpecttofindgradationsinimportantpointsofstructurebetweenthedifferentcastesofneutersinthesamespecies，thatIgladlyavailedmyselfofMrF。Smith’sofferofnumerousspecimensfromthesamenestofthedriverant(Anomma)ofWestAfrica。Thereaderwillperhapsbestappreciatetheamountofdifferenceintheseworkers，bymygivingnottheactualmeasurements，butastrictlyaccurateillustration：thedifferencewasthesameasifweweretoseeasetofworkmenbuildingahouseofwhommanywerefivefeetfourincheshigh，andmanysixteenfeethigh。butwemustsupposethatthelargerworkmenhadheadsfourinsteadofthreetimesasbigasthoseofthesmallermen，andjawsnearlyfivetimesasbig。Thejaws，moreover，oftheworkingantsoftheseveralsizesdifferedwonderfullyinshape，andintheformandnumberoftheteeth。Buttheimportantfactforusis，thatthoughtheworkerscanbegroupedintocastesofdifferentsizes，yettheygraduateinsensiblyintoeachother，asdoesthewidely-differentstructureoftheirjaws。Ispeakconfidentlyonthislatterpoint，asMrLubbockmadedrawingsformewiththecameralucidaofthejawswhichIhaddissectedfromtheworkersoftheseveralsizes。

Withthesefactsbeforeme，Ibelievethatnaturalselection，byactingonthefertileparents，couldformaspecieswhichshouldregularlyproduceneuters，eitheralloflargesizewithoneformofjaw，orallofsmallsizewithjawshavingawidelydifferentstructure。orlastly，andthisisourclimaxofdifficulty，onesetofworkersofonesizeandstructure，andsimultaneouslyanothersetofworkersofadifferentsizeandstructure。

agraduatedserieshavingbeenfirstformed，asinthecaseofthedriverant，andthentheextremeforms，frombeingthemostusefultothecommunity，havingbeenproducedingreaterandgreaternumbersthroughthenaturalselectionoftheparentswhichgeneratedthem。untilnonewithanintermediatestructurewereproduced。

Thus，asIbelieve，thewonderfulfactoftwodistinctlydefinedcastesofsterileworkersexistinginthesamenest，bothwidelydifferentfromeachotherandfromtheirparents，hasoriginated。Wecanseehowusefultheirproductionmayhavebeentoasocialcommunityofinsects，onthesameprinciplethatthedivisionoflabourisusefultocivilisedman。

Asantsworkbyinheritedinstinctsandbyinheritedtoolsorweapons，andnotbyacquiredknowledgeandmanufacturedinstruments，aperfectdivisionoflabourcouldbeeffectedwiththemonlybytheworkersbeingsterile。

forhadtheybeenfertile，theywouldhaveintercrossed，andtheirinstinctsandstructurewouldhavebecomeblended。Andnaturehas，asIbelieve，effectedthisadmirabledivisionoflabourinthecommunitiesofants，bythemeansofnaturalselection。ButIamboundtoconfess，that，withallmyfaithinthisprinciple，Ishouldneverhaveanticipatedthatnaturalselectioncouldhavebeenefficientinsohighadegree，hadnotthecaseoftheseneuterinsectsconvincedmeofthefact。Ihave，therefore，discussedthiscase，atsomelittlebutwhollyinsufficientlength，inordertoshowthepowerofnaturalselection，andlikewisebecausethisisbyfarthemostseriousspecialdifficulty，whichmytheoryhasencountered。Thecase，also，isveryinteresting，asitprovesthatwithanimals，aswithplants，anyamountofmodificationinstructurecanbeeffectedbytheaccumulationofnumerous，slight，andaswemustcallthemaccidental，variations，whichareinanymannerprofitable，withoutexerciseorhabithavingcomeintoplay。Fornoamountofexercise，orhabit，orvolition，intheutterlysterilemembersofacommunitycouldpossiblyhaveaffectedthestructureorinstinctsofthefertilemembers，whichaloneleavedescendants。Iamsurprisedthatnoonehasadvancedthisdemonstrativecaseofneuterinsects，againstthewell-knowndoctrineofLamarck。

Summary。Ihaveendeavouredbrieflyinthischaptertoshowthatthementalqualitiesofourdomesticanimalsvary，andthatthevariationsareinherited。StillmorebrieflyIhaveattemptedtoshowthatinstinctsvaryslightlyinastateofnature。Noonewilldisputethatinstinctsareofthehighestimportancetoeachanimal。ThereforeIcanseenodifficulty，underchangingconditionsoflife，innaturalselectionaccumulatingslightmodificationsofinstincttoanyextent，inanyusefuldirection。Insomecaseshabitoruseanddisusehaveprobablycomeintoplay。Idonotpretendthatthefactsgiveninthischapterstrengtheninanygreatdegreemytheory。butnoneofthecasesofdifficulty，tothebestofmyjudgment，annihilateit。Ontheotherhand，thefactthatinstinctsarenotalwaysabsolutelyperfectandareliabletomistakes。thatnoinstincthasbeenproducedfortheexclusivegoodofotheranimals，butthateachanimaltakesadvantageoftheinstinctsofothers。thatthecanoninnaturalhistory，of’naturanonfacitsaltum’isapplicabletoinstinctsaswellastocorporealstructure，andisplainlyexplicableontheforegoingviews，butisotherwiseinexplicable，alltendtocorroboratethetheoryofnaturalselection。

Thistheoryis，also，strengthenedbysomefewotherfactsinregardtoinstincts。asbythatcommoncaseofcloselyallied，butcertainlydistinct，species，wheninhabitingdistantpartsoftheworldandlivingunderconsiderablydifferentconditionsoflife，yetoftenretainingnearlythesameinstincts。

Forinstance，wecanunderstandontheprincipleofinheritance，howitisthatthethrushofSouthAmericalinesitsnestwithmud，inthesamepeculiarmannerasdoesourBritishthrush：howitisthatthemalewrens(Troglodytes)ofNorthAmerica，build’cock-nests，’toroostin，likethemalesofourdistinctKitty-wrens，ahabitwhollyunlikethatofanyotherknownbird。Finally，itmaynotbealogicaldeduction，buttomyimaginationitisfarmoresatisfactorytolookatsuchinstinctsastheyoungcuckooejectingitsfoster-brothers，antsmakingslaves，——thelarvaeofichneumonidaefeedingwithinthelivebodiesofcaterpillars，notasspeciallyendowedorcreatedinstincts，butassmallconsequencesofonegenerallaw，leadingtotheadvancementofallorganicbeings，namely，multiply，vary，letthestrongestliveandtheweakestdie。PreviousChapterNextChapterTheOriginofSpecies-Chapter8TheOriginofSpeciesbyCharlesDarwinPreviousChapterNextChapterChapter8-HybridismTHEviewgenerallyentertainedbynaturalistsisthatspecies，whenintercrossed，havebeenspeciallyendowedwiththequalityofsterility，inordertopreventtheconfusionofallorganicforms。Thisviewcertainlyseemsatfirstprobable，forspecieswithinthesamecountrycouldhardlyhavekeptdistincthadtheybeencapableofcrossingfreely。Theimportanceofthefactthathybridsareverygenerallysterile，has，Ithink，beenmuchunderratedbysomelatewriters。Onthetheoryofnaturalselectionthecaseisespeciallyimportant，inasmuchasthesterilityofhybridscouldnotpossiblybeofanyadvantagetothem，andthereforecouldnothavebeenacquiredbythecontinuedpreservationofsuccessiveprofitabledegreesofsterility。Ihope，however，tobeabletoshowthatsterilityisnotaspeciallyacquiredorendowedquality，butisincidentalonotheracquireddifferences。

Intreatingthissubject，twoclassesoffacts，toalargeextentfundamentallydifferent，havegenerallybeenconfoundedtogether。namely，thesterilityoftwospecieswhenfirstcrossed，andthesterilityofthehybridsproducedfromthem。

Purespecieshaveofcoursetheirorgansofreproductioninaperfectcondition，yetwhenintercrossedtheyproduceeitherfewornooffspring。

Hybrids，ontheotherhand，havetheirreproductiveorgansfunctionallyimpotent，asmaybeclearlyseeninthestateofthemaleelementinbothplantsandanimals。thoughtheorgansthemselvesareperfectinstructure，asfarasthemicroscopereveals。Inthefirstcasethetwosexualelementswhichgotoformtheembryoareperfect。inthesecondcasetheyareeithernotatalldeveloped，orareimperfectlydeveloped。Thisdistinctionisimportant，whenthecauseofthesterility，whichiscommontothetwocases，hastobeconsidered。Thedistinctionhasprobablybeenslurredover，owingtothesterilityinbothcasesbeinglookedonasaspecialendowment，beyondtheprovinceofourreasoningpowers。

Thefertilityofvarieties，thatisoftheformsknownorbelievedtohavedescendedfromcommonparents，whenintercrossed，andlikewisethefertilityoftheirmongreloffspring，is，onmytheory，ofequalimportancewiththesterilityofspecies。foritseemstomakeabroadandcleardistinctionbetweenvarietiesandspecies。

First，forthesterilityofspecieswhencrossedandoftheirhybridoffspring。Itisimpossibletostudytheseveralmemoirsandworksofthosetwoconscientiousandadmirableobservers，Kö。lreuterandGä。rtner，whoalmostdevotedtheirlivestothissubject，withoutbeingdeeplyimpressedwiththehighgeneralityofsomedegreeofsterility。Kö。lreutermakestheruleuniversal。butthenhecutstheknot，forintencasesinwhichhefoundtwoforms，consideredbymostauthorsasdistinctspecies，quitefertiletogether，heunhesitatinglyranksthemasvarieties。Gä。rtner，also，makestheruleequallyuniversal。andhedisputestheentirefertilityofKö。lreuter’stencases。Butintheseandinmanyothercases，Gä。rtnerisobligedcarefullytocounttheseeds，inordertoshowthatthereisanydegreeofsterility。Healwayscomparesthemaximumnumberofseedsproducedbytwospecieswhencrossedandbytheirhybridoffspring，withtheaveragenumberproducedbybothpureparent-speciesinastateofnature。

Butaseriouscauseoferrorseemstometobehereintroduced：aplanttobehybridisedmustbecastrated，and，whatisoftenmoreimportant，mustbesecludedinordertopreventpollenbeingbroughttoitbyinsectsfromotherplants。NearlyalltheplantsexperimentisedonbyGä。rtnerwerepotted，andapparentlywerekeptinachamberinhishouse。Thattheseprocessesareofteninjurioustothefertilityofaplantcannotbedoubted。

forGä。rtnergivesinhistableaboutascoreofcasesofplantswhichhecastrated，andartificiallyfertilisedwiththeirownpollen，and(excludingallcasessuchastheLeguminosae，inwhichthereisanacknowledgeddifficultyinthemanipulation)halfofthesetwentyplantshadtheirfertilityinsomedegreeimpaired。Moreover，asGä。rtnerduringseveralyearsrepeatedlycrossedtheprimroseandcowslip，whichwehavesuchgoodreasontobelievetobevarieties，andonlyonceortwicesucceededingettingfertileseed。

ashefoundthecommonredandbluepimpernels(Anagallisarvensisandcoerulea)，whichthebestbotanistsrankasvarieties，absolutelysteriletogether。andashecametothesameconclusioninseveralotheranalogouscases。itseemstomethatwemaywellbepermittedtodoubtwhethermanyotherspeciesarereallysosterile，whenintercrossed，asGä。rtnerbelieves。

Itiscertain，ontheonehand，thatthesterilityofvariousspecieswhencrossedissodifferentindegreeandgraduatesawaysoinsensibly，and，ontheotherhand，thatthefertilityofpurespeciesissoeasilyaffectedbyvariouscircumstances，thatforallpracticalpurposesitismostdifficulttosaywhereperfectfertilityendsandsterilitybegins。