第10章

SirC。Lyellinastrikingpassagehasspeculated，inlanguagealmostidenticalwithmine，ontheeffectsofgreatalterationsofclimateongeographicaldistribution。Ibelievethattheworldhasrecentlyfeltoneofhisgreatcyclesofchange。andthatonthisview，combinedwithmodificationthroughnaturalselection，amultitudeoffactsinthepresentdistributionbothofthesameandofalliedformsoflifecanbeexplained。Thelivingwatersmaybesaidtohaveflowedduringoneshortperiodfromthenorthandfromthesouth，andtohavecrossedattheequator。buttohaveflowedwithgreaterforcefromthenorthsoastohavefreelyinundatedthesouth。

Asthetideleavesitsdriftinhorizontallines，thoughrisinghigherontheshoreswherethetideriseshighest，sohavethelivingwaterslefttheirlivingdriftonourmountain-summits，inalinegentlyrisingfromthearcticlowlandstoagreatheightundertheequator。Thevariousbeingsthusleftstrandedmaybecomparedwithsavageracesofman，drivenupandsurvivinginthemountain-fastnessesofalmosteveryland，whichserveasarecord，fullofinteresttous，oftheformerinhabitantsofthesurroundinglowlands。PreviousChapterNextChapterTheOriginofSpecies-Chapter12TheOriginofSpeciesbyCharlesDarwinPreviousChapterNextChapterChapter12-GeographicalDistributioncontinuedAslakesandriver-systemsareseparatedfromeachotherbybarriersofland，itmighthavebeenthoughtthatfresh-waterproductionswouldnothaverangedwidelywithinthesamecountry，andastheseaisapparentlyastillmoreimpassablebarrier，thattheyneverwouldhaveextendedtodistantcountries。Butthecaseisexactlythereverse。Notonlyhavemanyfresh-waterspecies，belongingtoquitedifferentclasses，anenormousrange，butalliedspeciesprevailinaremarkablemannerthroughouttheworld。Iwellremember，whenfirstcollectinginthefreshwatersofBrazil，feelingmuchsurpriseatthesimilarityofthefresh-waterinsects，shells，&。c。，andatthedissimilarityofthesurroundingterrestrialbeings，comparedwiththoseofBritain。

Butthispowerinfresh-waterproductionsofrangingwidely，thoughsounexpected，can，Ithink，inmostcasesbeexplainedbytheirhavingbecomefitted，inamannerhighlyusefultothem，forshortandfrequentmigrationsfrompondtopond，orfromstreamtostream。andliabilitytowidedispersalwouldfollowfromthiscapacityasanalmostnecessaryconsequence。

Wecanhereconsideronlyafewcases。Inregardtofish，Ibelievethatthesamespeciesneveroccurinthefreshwatersofdistantcontinents。

Butonthesamecontinentthespeciesoftenrangewidelyandalmostcapriciously。

fortworiver-systemswillhavesomefishincommonandsomedifferent。

Afewfactsseemtofavourthepossibilityoftheiroccasionaltransportbyaccidentalmeans。likethatofthelivefishnotrarelydroppedbywhirlwindsinIndia，andthevitalityoftheirovawhenremovedfromthewater。ButIaminclinedtoattributethedispersaloffresh-waterfishmainlytoslightchangeswithintherecentperiodintheleveloftheland，havingcausedriverstoflowintoeachother。Instances，also，couldbegivenofthishavingoccurredduringfloods，withoutanychangeoflevel。WehaveevidenceintheloessoftheRhineofconsiderablechangesoflevelinthelandwithinaveryrecentgeologicalperiod，andwhenthesurfacewaspeopledbyexistinglandandfresh-watershells。Thewidedifferenceofthefishonoppositesidesofcontinuousmountain-ranges，whichfromanearlyperiodmusthavepartedriver-systemsandcompletelypreventedtheirinosculation，seemstoleadtothissameconclusion。Withrespecttoalliedfresh-waterfishoccurringatverydistantpointsoftheworld，nodoubttherearemanycaseswhichcannotatpresentbeexplained：butsomefresh-waterfishbelongtoveryancientforms，andinsuchcasestherewillhavebeenampletimeforgreatgeographicalchanges，andconsequentlytimeandmeansformuchmigration。Inthesecondplace，salt-waterfishcanwithcarebeslowlyaccustomedtoliveinfreshwater。and，accordingtoValenciennes，thereishardlyasinglegroupoffishesconfinedexclusivelytofreshwater，sothatwemayimaginethatamarinememberofafresh-watergroupmighttravelfaralongtheshoresofthesea，andsubsequentlybecomemodifiedandadaptedtothefreshwatersofadistantland。

Somespeciesoffresh-watershellshaveaverywiderange，andalliedspecies，which，onmytheory，aredescendedfromacommonparentandmusthaveproceededfromasinglesource，prevailthroughouttheworld。Theirdistributionatfirstperplexedmemuch，astheirovaarenotlikelytobetransportedbybirds，andtheyareimmediatelykilledbyseawater，asaretheadults。Icouldnotevenunderstandhowsomenaturalisedspecieshaverapidlyspreadthroughoutthesamecountry。Buttwofacts，whichI

haveobservedandnodoubtmanyothersremaintobeobservedthrowsomelightonthissubject。Whenaducksuddenlyemergesfromapondcoveredwithduck-weed，Ihavetwiceseentheselittleplantsadheringtoitsback。

andithashappenedtome，inremovingalittleduck-weedfromoneaquariumtoanother，thatIhavequiteunintentionallystockedtheonewithfresh-watershellsfromtheother。Butanotheragencyisperhapsmoreeffectual：I

suspendedaduck’sfeet，whichmightrepresentthoseofabirdsleepinginanaturalpond，inanaquarium，wheremanyovaoffresh-watershellswerehatching。andIfoundthatnumbersoftheextremelyminuteandjusthatchedshellscrawledonthefeet，andclungtothemsofirmlythatwhentakenoutofthewatertheycouldnotbejarredoff，thoughatasomewhatmoreadvancedagetheywouldvoluntarilydropoff。Thesejusthatchedmolluscs，thoughaquaticintheirnature，survivedontheduck’sfeet，indampair，fromtwelvetotwentyhours。andinthislengthoftimeaduckorheronmightflyatleastsixorsevenhundredmiles，andwouldbesuretoalightonapoolorrivulet，ifblownacrossseatoanoceanicislandortoanyotherdistantpoint。SirCharlesLyellalsoinformsmethataDyticushasbeencaughtwithanAncylus(afresh-watershelllikealimpet)firmlyadheringtoit。andawater-beetleofthesamefamily，aColymbetes，onceflewonboardthe`Beagle，’whenforty-fivemilesdistantfromthenearestland：howmuchfartheritmighthaveflownwithafavouringgalenoonecantell。

Withrespecttoplants，ithaslongbeenknownwhatenormousrangesmanyfresh-waterandevenmarsh-specieshave，bothovercontinentsandtothemostremoteoceanicislands。Thisisstrikinglyshown，asremarkedbyAlph。deCandolle，inlargegroupsofterrestrialplants，whichhaveonlyaveryfewaquaticmembers。fortheselatterseemimmediatelytoacquire，asifinconsequence，averywiderange。Ithinkfavourablemeansofdispersalexplainthisfact。Ihavebeforementionedthatearthoccasionally，thoughrarely，adheresinsomequantitytothefeetandbeaksofbirds。Wadingbirds，whichfrequentthemuddyedgesofponds，ifsuddenlyflushed，wouldbethemostlikelytohavemuddyfeet。BirdsofthisorderIcanshowarethegreatestwanderers，andareoccasionallyfoundonthemostremoteandbarrenislandsintheopenocean。theywouldnotbelikelytoalightonthesurfaceofthesea，sothatthedirtwouldnotbewashedofftheirfeet。whenmakingland，theywouldbesuretoflytotheirnaturalfresh-waterhaunts。Idonotbelievethatbotanistsareawarehowchargedthemudofpondsiswithseeds：Ihavetriedseverallittleexperiments，butwillheregiveonlythemoststrikingcase：ItookinFebruarythreetable-spoonfulsofmudfromthreedifferentpoints，beneathwater，ontheedgeofalittlepond。thismudwhendryweighedonly63/4ounces。Ikeptitcoveredupinmystudyforsixmonths，pullingupandcountingeachplantasitgrew。

theplantswereofmanykinds，andwerealtogether537innumber。andyettheviscidmudwasallcontainedinabreakfastcup！Consideringthesefacts，Ithinkitwouldbeaninexplicablecircumstanceifwater-birdsdidnottransporttheseedsoffresh-waterplantstovastdistances，andifconsequentlytherangeoftheseplantswasnotverygreat。Thesameagencymayhavecomeintoplaywiththeeggsofsomeofthesmallerfresh-wateranimals。

Otherandunknownagenciesprobablyhavealsoplayedapart。Ihavestatedthatfresh-waterfisheatsomekindsofseeds，thoughtheyrejectmanyotherkindsafterhavingswallowedthem。evensmallfishswallowseedsofmoderatesize，asoftheyellowwater-lilyandPotamogeton。Heronsandotherbirds，centuryaftercentury，havegoneondailydevouringfish。

theythentakeflightandgotootherwaters，orareblownacrossthesea。

andwehaveseenthatseedsretaintheirpowerofgermination，whenrejectedinpelletsorinexcrement，manyhoursafterwards。WhenIsawthegreatsizeoftheseedsofthatfinewater-lily，theNelumbium，andrememberedAlph。deCandolle’sremarksonthisplant，Ithoughtthatitsdistributionmustremainquiteinexplicable。butAudubonstatesthathefoundtheseedsofthegreatsouthernwater-lily(probably，accordingtoDrHooker，theNelumbiumluteum)inaheron’sstomach。althoughIdonotknowthefact，yetanalogymakesmebelievethataheronflyingtoanotherpondandgettingaheartymealoffish，wouldprobablyrejectfromitsstomachapelletcontainingtheseedsoftheNelumbiumundigested。ortheseedsmightbedroppedbythebirdwhilstfeedingitsyoung，inthesamewayasfishareknownsometimestobedropped。

Inconsideringtheseseveralmeansofdistribution，itshouldberememberedthatwhenapondorstreamisfirstformed，forinstance，onarisingislet，itwillbeunoccupied。andasingleseedoreggwillhaveagoodchanceofsucceeding。Althoughtherewillalwaysbeastruggleforlifebetweentheindividualsofthespecies，howeverfew，alreadyoccupyinganypond，yetasthenumberofkindsissmall，comparedwiththoseontheland，thecompetitionwillprobablybelessseverebetweenaquaticthanbetweenterrestrialspecies。consequentlyanintruderfromthewatersofaforeigncountry，wouldhaveabetterchanceofseizingonaplace，thaninthecaseofterrestrialcolonists。Weshould，also，rememberthatsome，perhapsmany，fresh-waterproductionsarelowinthescaleofnature，andthatwehavereasontobelievethatsuchlowbeingschangeorbecomemodifiedlessquicklythanthehigh。andthiswillgivelongertimethantheaverageforthemigrationofthesameaquaticspecies。Weshouldnotforgettheprobabilityofmanyspecieshavingformerlyrangedascontinuouslyasfresh-waterproductionsevercanrange，overimmenseareas，andhavingsubsequentlybecomeextinctinintermediateregions。Butthewidedistributionoffresh-waterplantsandoftheloweranimals，whetherretainingthesameidenticalformorinsomedegreemodified，Ibelievemainlydependsonthewidedispersaloftheirseedsandeggsbyanimals，moreespeciallybyfresh-waterbirds，whichhavelargepowersofflight，andnaturallytravelfromonetoanotherandoftendistantpieceofwater。Nature，likeacarefulgardener，thustakesherseedsfromabedofaparticularnature，anddropstheminanotherequallywellfittedforthem。OntheInhabitantsofOceanicIslandsWenowcometothelastofthethreeclassesoffacts，whichIhaveselectedaspresentingthegreatestamountofdifficulty，ontheviewthatalltheindividualsbothofthesameandofalliedspecieshavedescendedfromasingleparent。andthereforehaveallproceededfromacommonbirthplace，notwithstandingthatinthecourseoftimetheyhavecometoinhabitdistantpointsoftheglobe。IhavealreadystatedthatIcannothonestlyadmitForbes’sviewoncontinentalextensions，which，iflegitimatelyfollowedout，wouldleadtothebeliefthatwithintherecentperiodallexistingislandshavebeennearlyorquitejoinedtosomecontinent。Thisviewwouldremovemanydifficulties，butitwouldnot，Ithink，explainallthefactsinregardtoinsularproductions。InthefollowingremarksIshallnotconfinemyselftothemerequestionofdispersal。butshallconsidersomeotherfacts，whichbearonthetruthofthetwotheoriesofindependentcreationandofdescentwithmodification。

Thespeciesofallkindswhichinhabitoceanicislandsarefewinnumbercomparedwiththoseonequalcontinentalareas：Alph。deCandolleadmitsthisforplants，andWollastonforinsects。IfwelooktothelargesizeandvariedstationsofNewZealand，extendingover780milesoflatitude，andcompareitsfloweringplants，only750innumber，withthoseonanequalareaattheCapeofGoodHopeorinAustralia，wemust，Ithink，admitthatsomethingquiteindependentlyofanydifferenceinphysicalconditionshascausedsogreatadifferenceinnumber。EventheuniformcountyofCambridgehas847plants，andthelittleislandofAnglesea764，butafewfernsandafewintroducedplantsareincludedinthesenumbers，andthecomparisoninsomeotherrespectsisnotquitefair。WehaveevidencethatthebarrenislandofAscensionaboriginallypossessedunderhalf-a-dozenfloweringplants。yetmanyhavebecomenaturalisedonit，astheyhaveonNewZealandandoneveryotheroceanicislandwhichcanbenamed。InSt。Helenathereisreasontobelievethatthenaturalisedplantsandanimalshavenearlyorquiteexterminatedmanynativeproductions。Hewhoadmitsthedoctrineofthecreationofeachseparatespecies，willhavetoadmit，thatasufficientnumberofthebestadaptedplantsandanimalshavenotbeencreatedonoceanicislands。formanhasunintentionallystockedthemfromvarioussourcesfarmorefullyandperfectlythanhasnature。

Althoughinoceanicislandsthenumberofkindsofinhabitantsisscanty，theproportionofendemicspecies(i。e。thosefoundnowhereelseintheworld)isoftenextremelylarge。Ifwecompare，forinstance，thenumberoftheendemicland-shellsinMadeira，oroftheendemicbirdsintheGalapagosArchipelago，withthenumberfoundonanycontinent，andthencomparetheareaoftheislandswiththatofthecontinent，weshallseethatthisistrue。Thisfactmighthavebeenexpectedonmytheoryfor，asalreadyexplained，speciesoccasionallyarrivingafterlongintervalsinanewandisolateddistrict，andhavingtocompetewithnewassociates，willbeeminentlyliabletomodification，andwilloftenproducegroupsofmodifieddescendants。Butitbynomeansfollows，that，becauseinanislandnearlyallthespeciesofoneclassarepeculiar，thoseofanotherclass，orofanothersectionofthesameclass，arepeculiar。andthisdifferenceseemstodependonthespecieswhichdonotbecomemodifiedhavingimmigratedwithfacilityandinabody，sothattheirmutualrelationshavenotbeenmuchdisturbed。ThusintheGalapagosIslandsnearlyeveryland-bird，butonlytwooutoftheelevenmarinebirds，arepeculiar。anditisobviousthatmarinebirdscouldarriveattheseislandsmoreeasilythanland-birds。Bermuda，ontheotherhand，whichliesataboutthesamedistancefromNorthAmericaastheGalapagosIslandsdofromSouthAmerica，andwhichhasaverypeculiarsoil，doesnotpossessoneendemiclandbird。

andweknowfromMr。J。M。Jones’sadmirableaccountofBermuda，thatverymanyNorthAmericanbirds，duringtheirgreatannualmigrations，visiteitherperiodicallyoroccasionallythisisland。Madeiradoesnotpossessonepeculiarbird，andmanyEuropeanandAfricanbirdsarealmosteveryyearblownthere，asIaminformedbyMr。E。V。Harcourt。SothatthesetwoislandsofBermudaandMadeirahavebeenstockedbybirds，whichforlongageshavestruggledtogetherintheirformerhomes，andhavebecomemutuallyadaptedtoeachother。andwhensettledintheirnewhomes，eachkindwillhavebeenkeptbytheotherstotheirproperplacesandhabits，andwillconsequentlyhavebeenlittleliabletomodification。Madeira，again，isinhabitedbyawonderfulnumberofpeculiarland-shells，whereasnotonespeciesofsea-shellisconfinedtoitsshores：now，thoughwedonotknowhowseashellsaredispersed，yetwecanseethattheireggsorlarvae，perhapsattachedtoseaweedorfloatingtimber，ortothefeetofwading-birds，mightbetransportedfarmoreeasilythanland-shells，acrossthreeorfourhundredmilesofopensea。ThedifferentordersofinsectsinMadeiraapparentlypresentanalogousfacts。

Oceanicislandsaresometimesdeficientincertainclasses，andtheirplacesareapparentlyoccupiedbytheotherinhabitants。intheGalapagosIslandsreptiles，andinNewZealandgiganticwinglessbirds，taketheplaceofmammals。IntheplantsoftheGalapagosIslands，Dr。Hookerhasshownthattheproportionalnumbersofthedifferentordersareverydifferentfromwhattheyareelsewhere。Suchcasesaregenerallyaccountedforbythephysicalconditionsoftheislands。butthisexplanationseemstomenotalittledoubtful。Facilityofimmigration，Ibelieve，hasbeenatleastasimportantasthenatureoftheconditions。

Manyremarkablelittlefactscouldbegivenwithrespecttotheinhabitantsofremoteislands。Forinstance，incertainislandsnottenantedbymammals，someoftheendemicplantshavebeautifullyhookedseeds。yetfewrelationsaremorestrikingthantheadaptationofhookedseedsfortransportalbythewoolandfurofquadrupeds。Thiscasepresentsnodifficultyonmyview，forahookedseedmightbetransportedtoanislandbysomeothermeans。andtheplantthenbecomingslightlymodified，butstillretainingitshookedseeds，wouldformanendemicspecies，havingasuselessanappendageasanyrudimentaryorgan，forinstance，astheshrivelledwingsunderthesolderedelytraofmanyinsularbeetles。Again，islandsoftenpossesstreesorbushesbelongingtoorderswhichelsewhereincludeonlyherbaceousspecies。

nowtrees，asAlph。deCandollehasshown，generallyhave，whateverthecausemaybe，confinedranges。Hencetreeswouldbelittlelikelytoreachdistantoceanicislands。andanherbaceousplant，thoughitwouldhavenochanceofsuccessfullycompetinginstaturewithafullydevelopedtree，whenestablishedonanislandandhavingtocompetewithherbaceousplantsalone，mightreadilygainanadvantagebygrowingtallerandtallerandovertoppingtheotherplants。Ifso，naturalselectionwouldoftentendtoaddtothestatureofherbaceousplantswhengrowingonanisland，towhateverordertheybelonged，andthusconvertthemfirstintobushesandultimatelyintotrees。

Withrespecttotheabsenceofwholeordersonoceanicislands，BorySt。VincentlongagoremarkedthatBatrachians(frogs，toads，newts)haveneverbeenfoundonanyofthemanyislandswithwhichthegreatoceansarestudded。Ihavetakenpainstoverifythisassertion，andIhavefounditstrictlytrue。Ihave，however，beenassuredthatafrogexistsonthemountainsofthegreatislandofNewZealand。butIsuspectthatthisexception(iftheinformationbecorrect)maybeexplainedthroughglacialagency。

Thisgeneralabsenceoffrogs，toads，andnewtsonsomanyoceanicislandscannotbeaccountedforbytheirphysicalconditions。indeeditseemsthatislandsarepeculiarlywellfittedfortheseanimals。forfrogshavebeenintroducedintoMadeira，theAzores，andMauritius，andhavemultipliedsoastobecomeanuisance。Butastheseanimalsandtheirspawnareknowntobeimmediatelykilledbysea-water，onmyviewwecanseethattherewouldbegreatdifficultyintheirtransportalacrossthesea，andthereforewhytheydonotexistonanyoceanicisland。Butwhy，onthetheoryofcreation，theyshouldnothavebeencreatedthere，itwouldbeverydifficulttoexplain。

Mammalsofferanotherandsimilarcase。Ihavecarefullysearchedtheoldestvoyages，buthavenotfinishedmysearch。asyetIhavenotfoundasingleinstance，freefromdoubt，ofaterrestrialmammal(excludingdomesticatedanimalskeptbythenatives)inhabitinganislandsituatedabove300milesfromacontinentorgreatcontinentalisland。andmanyislandssituatedatamuchlessdistanceareequallybarren。TheFalklandIslands，whichareinhabitedbyawolf-likefox，comenearesttoanexception。

butthisgroupcannotbeconsideredasoceanic，asitliesonabankconnectedwiththemainland。moreover，icebergsformerlybroughtboulderstoitswesternshores，andtheymayhaveformerlytransportedfoxes，assofrequentlynowhappensinthearcticregions。Yetitcannotbesaidthatsmallislandswillnotsupportsmallmammals，fortheyoccurinmanypartsoftheworldonverysmallislands，ifclosetoacontinent。andhardlyanislandcanbenamedonwhichoursmallerquadrupedshavenotbecomenaturalisedandgreatlymultiplied。Itcannotbesaid，ontheordinaryviewofcreation，thattherehasnotbeentimeforthecreationofmammals。manyvolcanicislandsaresufficientlyancient，asshownbythestupendousdegradationwhichtheyhavesufferedandbytheirtertiarystrata：therehasalsobeentimefortheproductionofendemicspeciesbelongingtootherclasses。

andoncontinentsitisthoughtthatmammalsappearanddisappearataquickerratethanotherandloweranimals。Thoughterrestrialmammalsdonotoccuronoceanicislands，aë。rialmammalsdooccuronalmosteveryisland。NewZealandpossessestwobatsfoundnowhereelseintheworld：

NorfolkIsland，theVitiArchipelago，theBoninIslands，theCarolineandMarianneArchipelagoes，andMauritius，allpossesstheirpeculiarbats。

Why，itmaybeasked，hasthesupposedcreativeforceproducedbatsandnoothermammalsonremoteislands？Onmyviewthisquestioncaneasilybeanswered。fornoterrestrialmammalcanbetransportedacrossawidespaceofsea，butbatscanflyacross。BatshavebeenseenwanderingbydayfarovertheAtlanticOcean。andtwoNorthAmericanspecieseitherregularlyoroccasionallyvisitBermuda，atthedistanceof600milesfromthemainland。IhearfromMr。Tomes，whohasspeciallystudiedthisfamily，thatmanyofthesamespecieshaveenormousranges，andarefoundoncontinentsandonfardistantislands。Hencewehaveonlytosupposethatsuchwanderingspecieshavebeenmodifiedthroughnaturalselectionintheirnewhomesinrelationtotheirnewposition，andwecanunderstandthepresenceofendemicbatsonislands，withtheabsenceofallterrestrialmammals。

Besidestheabsenceofterrestrialmammalsinrelationtotheremotenessofislandsfromcontinents，thereisalsoarelation，toacertainextentindependentofdistance，betweenthedepthoftheseaseparatinganislandfromtheneighbouringmainland，andthepresenceinbothofthesamemammiferousspeciesorofalliedspeciesinamoreorlessmodifiedcondition。Mr。

WindsorEarlhasmadesomestrikingobservationsonthisheadinregardtothegreatMalayArchipelago，whichistraversednearCelebesbyaspaceofdeepocean。andthisspaceseparatestwowidelydistinctmammalianfaunas。

Oneithersidetheislandsaresituatedonmoderatelydeepsubmarinebanks，andtheyareinhabitedbycloselyalliedoridenticalquadrupeds。Nodoubtsomefewanomaliesoccurinthisgreatarchipelago，andthereismuchdifficultyinformingajudgmentinsomecasesowingtotheprobablenaturalisationofcertainmammalsthroughman’sagency。butweshallsoonhavemuchlightthrownonthenaturalhistoryofthisarchipelagobytheadmirablezealandresearchesofMrWallace。Ihavenotasyethadtimetofollowupthissubjectinallotherquartersoftheworld。butasfarasIhavegone，therelationgenerallyholdsgood。WeseeBritainseparatedbyashallowchannelfromEurope，andthemammalsarethesameonbothsides。wemeetwithanalogousfactsonmanyislandsseparatedbysimilarchannelsfromAustralia。TheWestIndianIslandsstandonadeeplysubmergedbank，nearly1000fathomsindepth，andherewefindAmericanforms，butthespeciesandeventhegeneraaredistinct。Astheamountofmodificationinallcasesdependstoacertaindegreeonthelapseoftime，andasduringchangesoflevelitisobviousthatislandsseparatedbyshallowchannelsaremorelikelytohavebeencontinuouslyunitedwithinarecentperiodtothemainlandthanislandsseparatedbydeeperchannels，wecanunderstandthefrequentrelationbetweenthedepthoftheseaandthedegreeofaffinityofthemammalianinhabitantsofislandswiththoseofaneighbouringcontinent，anexplicablerelationontheviewofindependentactsofcreation。

Alltheforegoingremarksontheinhabitantsofoceanicislands，namely，thescarcityofkinds——therichnessinendemicformsinparticularclassesorsectionsofclasses，theabsenceofwholegroups，asofbatrachians，andofterrestrialmammalsnotwithstandingthepresenceofaë。rialbats，thesingularproportionsofcertainordersofplants，herbaceousformshavingbeendevelopedintotrees，&。c。，seemtometoaccordbetterwiththeviewofoccasionalmeansoftransporthavingbeenlargelyefficientinthelongcourseoftime，thanwiththeviewofallouroceanicislandshavingbeenformerlyconnectedbycontinuouslandwiththenearestcontinent。

foronthislatterviewthemigrationwouldprobablyhavebeenmorecomplete。

andifmodificationbeadmitted，alltheformsoflifewouldhavebeenmoreequallymodified，inaccordancewiththeparamountimportanceoftherelationoforganismtoorganism。

Idonotdenythattherearemanyandgravedifficultiesinunderstandinghowseveraloftheinhabitantsofthemoreremoteislands，whetherstillretainingthesamespecificformormodifiedsincetheirarrival，couldhavereachedtheirpresenthomes。Buttheprobabilityofmanyislandshavingexistedashalting-places，ofwhichnotawrecknowremains，mustnotbeoverlooked。Iwillheregiveasingleinstanceofoneofthecasesofdifficulty。

Almostalloceanicislands，eventhemostisolatedandsmallest，areinhabitedbyland-shells，generallybyendemicspecies，butsometimesbyspeciesfoundelsewhere。Dr。Aug。A。Gouldhasgivenseveralinterestingcasesinregardtotheland-shellsoftheislandsofthePacific。Nowitisnotoriousthatland-shellsareveryeasilykilledbysalt。theireggs，atleastsuchasIhavetried，sinkinsea-waterandarekilledbyit。Yettheremustbe，onmyview，someunknown，buthighlyefficientmeansfortheirtransportal。

Wouldthejust-hatchedyoungoccasionallycrawlonandadheretothefeetofbirdsroostingontheground，andthusgettransported？Itoccurredtomethatland-shells，whenhybernatingandhavingamembranousdiaphragmoverthemouthoftheshell，mightbefloatedinchinksofdriftedtimberacrossmoderatelywidearmsofthesea。AndIfoundthatseveralspeciesdidinthisstatewithstanduninjuredanimmersioninsea-waterduringsevendays：oneoftheseshellswastheHelixpomatia，andafterithadagainhybernatedIputitinsea-waterfortwentydays，anditperfectlyrecovered。Asthisspecieshasathickcalcareousoperculum，Iremovedit，andwhenithadformedanewmembranousone，Iimmerseditforfourteendaysinsea-water，anditrecoveredandcrawledaway：butmoreexperimentsarewantedonthishead。

Themoststrikingandimportantfactforusinregardtotheinhabitantsofislands，istheiraffinitytothoseofthenearestmainland，withoutbeingactuallythesamespecies。Numerousinstancescouldbegivenofthisfact。Iwillgiveonlyone，thatoftheGalapagosArchipelago，situatedundertheequator，between500and600milesfromtheshoresofSouthAmerica。

HerealmosteveryproductofthelandandwaterbearstheunmistakeablestampoftheAmericancontinent。Therearetwenty-sixlandbirds，andtwenty-fiveofthosearerankedbyMrGouldasdistinctspecies，supposedtohavebeencreatedhere。yetthecloseaffinityofmostofthesebirdstoAmericanspeciesineverycharacter，intheirhabits，gestures，andtonesofvoice，wasmanifest。Soitiswiththeotheranimals，andwithnearlyalltheplants，asshownbyDr。HookerinhisadmirablememoirontheFloraofthisarchipelago。Thenaturalist，lookingattheinhabitantsofthesevolcanicislandsinthePacific，distantseveralhundredmilesfromthecontinent，yetfeelsthatheisstandingonAmericanland。Whyshouldthisbeso？

whyshouldthespecieswhicharesupposedtohavebeencreatedintheGalapagosArchipelago，andnowhereelse，bearsoplainastampofaffinitytothosecreatedinAmerica？Thereisnothingintheconditionsoflife，inthegeologicalnatureoftheislands，intheirheightorclimate，orintheproportionsinwhichtheseveralclassesareassociatedtogether，whichresemblescloselytheconditionsoftheSouthAmericancoast：infactthereisaconsiderabledissimilarityinalltheserespects。Ontheotherhand，thereisaconsiderabledegreeofresemblanceinthevolcanicnatureofthesoil，inclimate，height，andsizeoftheislands，betweentheGalapagosandCapedeVerdeArchipelagos：butwhatanentireandabsolutedifferenceintheirinhabitants！TheinhabitantsoftheCapedeVerdeIslandsarerelatedtothoseofAfrica，likethoseoftheGalapagostoAmerica。Ibelievethisgrandfactcanreceivenosortofexplanationontheordinaryviewofindependentcreation。whereasontheviewheremaintained，itisobviousthattheGalapagosIslandswouldbelikelytoreceivecolonists，whetherbyoccasionalmeansoftransportorbyformerlycontinuousland，fromAmerica。

andtheCapedeVerdeIslandsfromAfrica。andthatsuchcolonistswouldbeliabletomodifications。theprincipleofinheritancestillbetrayingtheiroriginalbirthplace。

Manyanalogousfactscouldbegiven：indeeditisanalmostuniversalrulethattheendemicproductionsofislandsarerelatedtothoseofthenearestcontinent，orofothernearislands。Theexceptionsarefew，andmostofthemcanbeexplained。ThustheplantsofKerguelenLand，thoughstandingnearertoAfricathantoAmerica，arerelated，andthatveryclosely，asweknowfromDr。Hooker’saccount，tothoseofAmerica：butontheviewthatthisislandhasbeenmainlystockedbyseedsbroughtwithearthandstonesonicebergs，driftedbytheprevailingcurrents，thisanomalydisappears。

NewZealandinitsendemicplantsismuchmorecloselyrelatedtoAustralia，thenearestmainland，thantoanyotherregion：andthisiswhatmighthavebeenexpected。butitisalsoplainlyrelatedtoSouthAmerica，which，althoughthenextnearestcontinent，issoenormouslyremote，thatthefactbecomesananomaly。ButthisdifficultyalmostdisappearsontheviewthatbothNewZealand，SouthAmerica，andothersouthernlandswerelongagopartiallystockedfromanearlyintermediatethoughdistantpoint，namelyfromtheantarcticislands，whentheywereclothedwithvegetation，beforethecommencementoftheGlacialperiod。Theaffinity，which，thoughfeeble，IamassuredbyDr。Hookerisreal，betweenthefloraofthesouth-westerncornerofAustraliaandoftheCapeofGoodHope，isafarmoreremarkablecase，andisatpresentinexplicable：butthisaffinityisconfinedtotheplants，andwill，Idonotdoubt，besomedayexplained。

Thelawwhichcausestheinhabitantsofanarchipelago，thoughspecificallydistinct，tobecloselyalliedtothoseofthenearestcontinent，wesometimesseedisplayedonasmallscale，yetinamostinterestingmanner，withinthelimitsofthesamearchipelago。ThustheseveralislandsoftheGalapagosArchipelagoaretenanted，asIhaveelsewhereshown，inaquitemarvellousmanner，byverycloselyrelatedspecies。sothattheinhabitantsofeachseparateisland，thoughmostlydistinct，arerelatedinanincomparablycloserdegreetoeachotherthantotheinhabitantsofanyotherpartoftheworld。Andthisisjustwhatmighthavebeenexpectedonmyview，fortheislandsaresituatedsoneareachotherthattheywouldalmostcertainlyreceiveimmigrantsfromthesameoriginalsource，orfromeachother。Butthisdissimilaritybetweentheendemicinhabitantsoftheislandsmaybeusedasanargumentagainstmyviews。foritmaybeasked，howhasithappenedintheseveralislandssituatedwithinsightofeachother，havingthesamegeologicalnature，thesameheight，climate，&。c。，thatmanyoftheimmigrantsshouldhavebeendifferentlymodified，thoughonlyinasmalldegree。Thislongappearedtomeagreatdifficulty：butitarisesinchiefpartfromthedeeply-seatederrorofconsideringthephysicalconditionsofacountryasthemostimportantforitsinhabitants。whereasitcannot，Ithink，bedisputedthatthenatureoftheotherinhabitants，withwhicheachhastocompete，isatleastasimportant，andgenerallyafarmoreimportantelementofsuccess。NowifwelooktothoseinhabitantsoftheGalapagosArchipelagowhicharefoundinotherpartsoftheworld(havingononesideforthemomenttheendemicspecies，whichcannotbeherefairlyincluded，asweareconsideringhowtheyhavecometobemodifiedsincetheirarrival)，wefindaconsiderableamountofdifferenceintheseveralislands。Thisdifferencemightindeedhavebeenexpectedontheviewoftheislandshavingbeenstockedbyoccasionalmeansoftransportaseed，forinstance，ofoneplanthavingbeenbroughttooneisland，andthatofanotherplanttoanotherisland。Hencewheninformertimesanimmigrantsettledonanyoneormoreoftheislands，orwhenitsubsequentlyspreadfromoneislandtoanother，itwouldundoubtedlybeexposedtodifferentconditionsoflifeinthedifferentislands，foritwouldhavetocompetewithdifferentsetsoforganisms：aplant，forinstance，wouldfindthebest-fittedgroundmoreperfectlyoccupiedbydistinctplantsinoneislandthaninanother，anditwouldbeexposedtotheattacksofsomewhatdifferentenemies。Ifthenitvaried，naturalselectionwouldprobablyfavourdifferentvarietiesinthedifferentislands。Somespecies，however，mightspreadandyetretainthesamecharacterthroughoutthegroup，justasweseeoncontinentssomespecies’spreadingwidelyandremainingthesame。

ThereallysurprisingfactinthiscaseoftheGalapagosArchipelago，andinalesserdegreeinsomeanalogousinstances，isthatthenewspeciesformedintheseparateislandshavenotquicklyspreadtotheotherislands。

Buttheislands，thoughinsightofeachother，areseparatedbydeeparmsofthesea，inmostcaseswiderthantheBritishChannel，andthereisnoreasontosupposethattheyhaveatanyformerperiodbeencontinuouslyunited。Thecurrentsoftheseaarerapidandsweepacrossthearchipelago，andgalesofwindareextraordinarilyrare。sothattheislandsarefarmoreeffectuallyseparatedfromeachotherthantheyappeartobeonamap。Neverthelessagoodmanyspecies，boththosefoundinotherpartsoftheworldandthoseconfinedtothearchipelago，arecommontotheseveralislands，andwemayinferfromcertainfactsthatthesehaveprobablyspreadfromsomeoneislandtotheothers。Butweoftentake，Ithink，anerroneousviewoftheprobabilityofcloselyalliedspeciesinvadingeachother’sterritory，whenputintofreeintercommunication。Undoubtedlyifonespecieshasanyadvantagewhateveroveranother，itwillinaverybrieftimewhollyorinpartsupplantit。butifbothareequallywellfittedfortheirownplacesinnature，bothprobablywillholdtheirownplacesandkeepseparateforalmostanylengthoftime。Beingfamiliarwiththefactthatmanyspecies，naturalisedthroughman’sagency，havespreadwithastonishingrapidityovernewcountries，weareapttoinferthatmostspecieswouldthusspread。

butweshouldrememberthattheformswhichbecomenaturalisedinnewcountriesarenotgenerallycloselyalliedtotheaboriginalinhabitants，butareverydistinctspecies，belonginginalargeproportionofcases，asshownbyAlph。deCandolle，todistinctgenera。IntheGalapagosArchipelago，manyevenofthebirds，thoughsowelladaptedforflyingfromislandtoisland，aredistinctoneach。thustherearethreeclosely-alliedspeciesofmocking-thrush，eachconfinedtoitsownisland。Nowletussupposethemocking-thrushofChathamIslandtobeblowntoCharlesIsland，whichhasitsownmocking-thrush：whyshoulditsucceedinestablishingitselfthere？WemaysafelyinferthatCharlesIslandiswellstockedwithitsownspecies，forannuallymoreeggsarelaidtherethancanpossiblybereared。andwemayinferthatthemocking-thrushpeculiartoCharlesIslandisatleastaswellfittedforitshomeasisthespeciespeculiartoChathamIsland。SirC。LyellandMr。Wollastonhavecommunicatedtomearemarkablefactbearingonthissubject。namely，thatMadeiraandtheadjoiningisletofPortoSantopossessmanydistinctbutrepresentativeland-shells，someofwhichliveincrevicesofstone。andalthoughlargequantitiesofstoneareannuallytransportedfromPortoSantotoMadeira，yetthislatterislandhasnotbecomecolonisedbythePortoSantospecies：neverthelessbothislandshavebeencolonisedbysomeEuropeanland-shells，whichnodoubthadsomeadvantageovertheindigenousspecies。FromtheseconsiderationsIthinkweneednotgreatlymarvelattheendemicandrepresentativespecies，whichinhabittheseveralislandsoftheGalapagosArchipelago，nothavinguniversallyspreadfromislandtoisland。Inmanyotherinstances，asintheseveraldistrictsofthesamecontinent，pre-occupationhasprobablyplayedanimportantpartincheckingthecomminglingofspeciesunderthesameconditionsoflife。Thus，thesouth-eastandsouth-westcornersofAustraliahavenearlythesamephysicalconditions，andareunitedbycontinuousland，yettheyareinhabitedbyavastnumberofdistinctmammals，birds，andplants。

Theprinciplewhichdeterminesthegeneralcharacterofthefaunaandfloraofoceanicislands，namely，thattheinhabitants，whennotidenticallythesame，yetareplainlyrelatedtotheinhabitantsofthatregionwhencecolonistscouldmostreadilyhavebeenderived，thecolonistshavingbeensubsequentlymodifiedandbetterfittedtotheirnewhomes，isofthewidestapplicationthroughoutnature。Weseethisoneverymountain，ineverylakeandmarsh。ForAlpinespecies，exceptinginsofarasthesameforms，chieflyofplants，havespreadwidelythroughouttheworldduringtherecentGlacialepoch，arerelatedtothoseofthesurroundinglowlands。thuswehaveinSouthAmerica，Alpinehumming-birds，Alpinerodents，Alpineplants，&。c。，allofstrictlyAmericanforms，anditisobviousthatamountain，asitbecameslowlyupheaved，wouldnaturallybecolonisedfromthesurroundinglowlands。Soitiswiththeinhabitantsoflakesandmarshes，exceptinginsofarasgreatfacilityoftransporthasgiventhesamegeneralformstothewholeworld。WeseethissameprincipleintheblindanimalsinhabitingthecavesofAmericaandofEurope。Otheranalogousfactscouldbegiven。

Anditwill，Ibelieve，beuniversallyfoundtobetrue，thatwhereverintworegions，letthembeeversodistant，manycloselyalliedorrepresentativespeciesoccur，therewilllikewisebefoundsomeidenticalspecies，showing，inaccordancewiththeforegoingview，thatatsomeformerperiodtherehasbeenintercommunicationormigrationbetweenthetworegions。Andwherevermanyclosely-alliedspeciesoccur，therewillbefoundmanyformswhichsomenaturalistsrankasdistinctspecies，andsomeasvarieties。thesedoubtfulformsshowingusthestepsintheprocessofmodification。

Thisrelationbetweenthepowerandextentofmigrationofaspecies，eitheratthepresenttimeoratsomeformerperiodunderdifferentphysicalconditions，andtheexistenceatremotepointsoftheworldofotherspeciesalliedtoit，isshowninanotherandmoregeneralway。Mr。Gouldremarkedtomelongago，thatinthosegeneraofbirdswhichrangeovertheworld，manyofthespecieshaveverywideranges。Icanhardlydoubtthatthisruleisgenerallytrue，thoughitwouldbedifficulttoproveit。Amongstmammals，weseeitstrikinglydisplayedinBats，andinalesserdegreeintheFelidaeandCanidae。Weseeit，ifwecomparethedistributionofbutterfliesandbeetles。Soitiswithmostfresh-waterproductions，inwhichsomanygenerarangeovertheworld，andmanyindividualspecieshaveenormousranges。Itisnotmeantthatinworld-ranginggeneraallthespecieshaveawiderange，oreventhattheyhaveonanaverageawiderange。butonlythatsomeofthespeciesrangeverywidely。forthefacilitywithwhichwidely-rangingspeciesvaryandgiverisetonewformswilllargelydeterminetheiraveragerange。Forinstance，twovarietiesofthesamespeciesinhabitAmericaandEurope，andthespeciesthushasanimmenserange。but，ifthevariationhadbeenalittlegreater，thetwovarietieswouldhavebeenrankedasdistinctspecies，andthecommonrangewouldhavebeengreatlyreduced。Stilllessisitmeant，thataspecieswhichapparentlyhasthecapacityofcrossingbarriersandrangingwidely，asinthecaseofcertainpowerfully-wingedbirds，willnecessarilyrangewidely。forweshouldneverforgetthattorangewidelyimpliesnotonlythepowerofcrossingbarriers，butthemoreimportantpowerofbeingvictoriousindistantlandsinthestruggleforlifewithforeignassociates。Butontheviewofallthespeciesofagenushavingdescendedfromasingleparent，thoughnowdistributedtothemostremotepointsoftheworld，weoughttofind，andIbelieveasageneralrulewedofind，thatsomeatleastofthespeciesrangeverywidely。foritisnecessarythattheunmodifiedparentshouldrangewidely，undergoingmodificationduringitsdiffusion，andshouldplaceitselfunderdiverseconditionsfavourablefortheconversionofitsoffspring，firstlyintonewvarietiesandultimatelyintonewspecies。

Inconsideringthewidedistributionofcertaingenera，weshouldbearinmindthatsomeareextremelyancient，andmusthavebranchedofffromacommonparentataremoteepoch。sothatinsuchcasestherewillhavebeenampletimeforgreatclimatalandgeographicalchangesandforaccidentsoftransport。andconsequentlyforthemigrationofsomeofthespeciesintoallquartersoftheworld，wheretheymayhavebecomeslightlymodifiedinrelationtotheirnewconditions。Thereis，also，somereasontobelievefromgeologicalevidencethatorganismslowinthescalewithineachgreatclass，generallychangeataslowerratethanthehigherforms。andconsequentlythelowerformswillhavehadabetterchanceofrangingwidelyandofstillretainingthesamespecificcharacter。Thisfact，togetherwiththeseedsandeggsofmanylowformsbeingveryminuteandbetterfittedfordistanttransportation，probablyaccountsforalawwhichhaslongbeenobserved，andwhichhaslatelybeenadmirablydiscussedbyAlph。deCandolleinregardtoplants，namely，thattheloweranygroupoforganismsis，themorewidelyitisapttorange。

Therelationsjustdiscussed，namely，lowandslowly-changingorganismsrangingmorewidelythanthehigh，someofthespeciesofwidely-ranginggenerathemselvesrangingwidely，suchfacts，asalpine，lacustrine，andmarshproductionsbeingrelated(withtheexceptionsbeforespecified)

tothoseonthesurroundinglowlandsanddrylands，thoughthesestationsaresodifferenttheverycloserelationofthedistinctspecieswhichinhabittheisletsofthesamearchipelago，andespeciallythestrikingrelationoftheinhabitantsofeachwholearchipelagoorislandtothoseofthenearestmainland，are，Ithink，utterlyinexplicableontheordinaryviewoftheindependentcreationofeachspecies，butareexplicableontheviewofcolonisationfromthenearestandreadiestsource，togetherwiththesubsequentmodificationandbetteradaptationofthecoloniststotheirnewhomes。SummaryoflastandpresentChaptersInthesechaptersIhaveendeavouredtoshow，thatifwemakedueallowanceforourignoranceofthefulleffectsofallthechangesofclimateandoftheleveloftheland，whichhavecertainlyoccurredwithintherecentperiod，andofothersimilarchangeswhichmayhaveoccurredwithinthesameperiod。ifwerememberhowprofoundlyignorantwearewithrespecttothemanyandcuriousmeansofoccasionaltransport，asubjectwhichhashardlyeverbeenproperlyexperimentisedon。ifwebearinmindhowoftenaspeciesmayhaverangedcontinuouslyoverawidearea，andthenhavebecomeextinctintheintermediatetracts，Ithinkthedifficultiesinbelievingthatalltheindividualsofthesamespecies，whereverlocated，havedescendedfromthesameparents，arenotinsuperable。Andweareledtothisconclusion，whichhasbeenarrivedatbymanynaturalistsunderthedesignationofsinglecentresofcreation，bysomegeneralconsiderations，moreespeciallyfromtheimportanceofbarriersandfromtheanalogicaldistributionofsub-genera，genera，andfamilies。

Withrespecttothedistinctspeciesofthesamegenus，whichonmytheorymusthavespreadfromoneparent-source。ifwemakethesameallowancesasbeforeforourignorance，andrememberthatsomeformsoflifechangemostslowly，enormousperiodsoftimebeingthusgrantedfortheirmigration，Idonotthinkthatthedifficultiesareinsuperable。thoughtheyoftenareinthiscase，andinthatoftheindividualsofthesamespecies，extremelygrave。

Asexemplifyingtheeffectsofclimatalchangesondistribution，IhaveattemptedtoshowhowimportanthasbeentheinfluenceofthemodernGlacialperiod，whichIamfullyconvincedsimultaneouslyaffectedthewholeworld，oratleastgreatmeridionalbelts。Asshowinghowdiversifiedarethemeansofoccasionaltransport，Ihavediscussedatsomelittlelengththemeansofdispersaloffresh-waterproductions。

Ifthedifficultiesbenotinsuperableinadmittingthatinthelongcourseoftimetheindividualsofthesamespecies，andlikewiseofalliedspecies，haveproceededfromsomeonesource。thenIthinkallthegrandleadingfactsofgeographicaldistributionareexplicableonthetheoryofmigration(generallyofthemoredominantformsoflife)，togetherwithsubsequentmodificationandthemultiplicationofnewforms。Wecanthusunderstandthehighimportanceofbarriers，whetheroflandorwater，whichseparateourseveralzoologicalandbotanicalprovinces。Wecanthusunderstandthelocalisationofsub-genera，genera，andfamilies。andhowitisthatunderdifferentlatitudes，forinstanceinSouthAmerica，theinhabitantsoftheplainsandmountains，oftheforests，marshes，anddeserts，areinsomysteriousamannerlinkedtogetherbyaffinity，andarelikewiselinkedtotheextinctbeingswhichformerlyinhabitedthesamecontinent。

Bearinginmindthatthemutualrelationsoforganismtoorganismareofthehighestimportance，wecanseewhytwoareashavingnearlythesamephysicalconditionsshouldoftenbeinhabitedbyverydifferentformsoflife。foraccordingtothelengthoftimewhichhaselapsedsincenewinhabitantsenteredoneregion。accordingtothenatureofthecommunicationwhichallowedcertainformsandnototherstoenter，eitheringreaterorlessernumbers。accordingornot，asthosewhichenteredhappenedtocomeinmoreorlessdirectcompetitionwitheachotherandwiththeaborigines。andaccordingastheimmigrantswerecapableofvaryingmoreorlessrapidly，therewouldensueindifferentregions，independentlyoftheirphysicalconditions，infinitelydiversifiedconditionsoflife，therewouldbeanalmostendlessamountoforganicactionandreaction，andweshouldfind，aswedofind，somegroupsofbeingsgreatly，andsomeonlyslightlymodified，somedevelopedingreatforce，someexistinginscantynumbersinthedifferentgreatgeographicalprovincesoftheworld。

Onthesesameprinciples，wecanunderstand，asIhaveendeavouredtoshow，whyoceanicislandsshouldhavefewinhabitants，butoftheseagreatnumbershouldbeendemicorpeculiar。andwhy，inrelationtothemeansofmigration，onegroupofbeings，evenwithinthesameclass，shouldhaveallitsspeciesendemic，andanothergroupshouldhaveallitsspeciescommontootherquartersoftheworld。Wecanseewhywholegroupsoforganisms，asbatrachiansandterrestrialmammals，shouldbeabsentfromoceanicislands，whilstthemostisolatedislandspossesstheirownpeculiarspeciesofaë。rialmammalsorbats。Wecanseewhythereshouldbesomerelationbetweenthepresenceofmammals，inamoreorlessmodifiedcondition，andthedepthoftheseabetweenanislandandthemainland。Wecanclearlyseewhyalltheinhabitantsofanarchipelago，thoughspecificallydistinctontheseveralislets，shouldbecloselyrelatedtoeachother，andlikewiseberelated，butlessclosely，tothoseofthenearestcontinentorothersourcewhenceimmigrantswereprobablyderived。Wecanseewhyintwoareas，howeverdistantfromeachother，thereshouldbeacorrelation，inthepresenceofidenticalspecies，ofvarieties，ofdoubtfulspecies，andofdistinctbutrepresentativespecies。

AsthelateEdwardForbesofteninsisted，thereisastrikingparallelisminthelawsoflifethroughouttimeandspace：thelawsgoverningthesuccessionofformsinpasttimesbeingnearlythesamewiththosegoverningatthepresenttimethedifferencesindifferentareas。Weseethisinmanyfacts。

Theenduranceofeachspeciesandgroupofspeciesiscontinuousintime。

fortheexceptionstotherulearesofew，thattheymayfairlybeattributedtoournothavingasyetdiscoveredinanintermediatedeposittheformswhicharethereinabsent，butwhichoccuraboveandbelow：soinspace，itcertainlyisthegeneralrulethattheareainhabitedbyasinglespecies，orbyagroupofspecies，iscontinuous。andtheexceptions，whicharenotrare，may，asIhaveattemptedtoshow，beaccountedforbymigrationatsomeformerperiodunderdifferentconditionsorbyoccasionalmeansoftransport，andbythespecieshavingbecomeextinctintheintermediatetracts。Bothintimeandspace，speciesandgroupsofspecieshavetheirpointsofmaximumdevelopment。Groupsofspecies，belongingeithertoacertainperiodoftime，ortoacertainarea，areoftencharacterisedbytriflingcharactersincommon，asofsculptureorcolour。Inlookingtothelongsuccessionofages，asinnowlookingtodistantprovincesthroughouttheworld，wefindthatsomeorganismsdifferlittle，whilstothersbelongingtoadifferentclass，ortoadifferentorder，orevenonlytoadifferentfamilyofthesameorder，differgreatly。Inbothtimeandspacethelowermembersofeachclassgenerallychangelessthanthehigher。butthereareinbothcasesmarkedexceptionstotherule。Onmytheorytheseseveralrelationsthroughouttimeandspaceareintelligible。forwhetherwelooktotheformsoflifewhichhavechangedduringsuccessiveageswithinthesamequarteroftheworld，ortothosewhichhavechangedafterhavingmigratedintodistantquarters，inbothcasestheformswithineachclasshavebeenconnectedbythesamebondofordinarygeneration。andthemorenearlyanytwoformsarerelatedinblood，thenearertheywillgenerallystandtoeachotherintimeandspace。inbothcasesthelawsofvariationhavebeenthesame，andmodificationshavebeenaccumulatedbythesamepowerofnaturalselection。PreviousChapterNextChapterTheOriginofSpecies-Chapter13TheOriginofSpeciesbyCharlesDarwinPreviousChapterNextChapterChapter13-MutualAffinitiesofOrganicBeings：Morphology：Embryology：

RudimentaryOrgansFromthefirstdawnoflife，allorganicbeingsarefoundtoresembleeachotherindescendingdegrees，sothattheycanbeclassedingroupsundergroups。Thisclassificationisevidentlynotarbitrarylikethegroupingofthestarsinconstellations。Theexistenceofgroupswouldhavebeenofsimplesignification，ifonegrouphadbeenexclusivelyfittedtoinhabittheland，andanotherthewater。onetofeedonflesh，anotheronvegetablematter，andsoon。butthecaseiswidelydifferentinnature。foritisnotorioushowcommonlymembersofeventhesamesubgrouphavedifferenthabits。Inoursecondandfourthchapters，onVariationandonNaturalSelection，Ihaveattemptedtoshowthatitisthewidelyranging，themuchdiffusedandcommon，thatisthedominantspeciesbelongingtothelargergenera，whichvarymost。Thevarieties，orincipientspecies，thusproducedultimatelybecomeconverted，asIbelieve，intonewanddistinctspecies。andthese，ontheprincipleofinheritance，tendtoproduceothernewanddominantspecies。Consequentlythegroupswhicharenowlarge，andwhichgenerallyincludemanydominantspecies，tendtogoonincreasingindefinitelyinsize。Ifurtherattemptedtoshowthatfromthevaryingdescendantsofeachspeciestryingtooccupyasmanyandasdifferentplacesaspossibleintheeconomyofnature，thereisaconstanttendencyintheircharacterstodiverge。Thisconclusionwassupportedbylookingatthegreatdiversityoftheformsoflifewhich，inanysmallarea，comeintotheclosestcompetition，andbylookingtocertainfactsinnaturalisation。

Iattemptedalsotoshowthatthereisaconstanttendencyintheformswhichareincreasinginnumberanddivergingincharacter，tosupplantandexterminatethelessdivergent，thelessimproved，andprecedingforms。

Irequestthereadertoturntothediagramillustratingtheaction，asformerlyexplained，oftheseseveralprinciples。andhewillseethattheinevitableresultisthatthemodifieddescendantsproceedingfromoneprogenitorbecomebrokenupintogroupssubordinatetogroups。Inthediagrameachletterontheuppermostlinemayrepresentagenusincludingseveralspecies。andallthegeneraonthislineformtogetheroneclass，forallhavedescendedfromoneancientbutunseenparent，and，consequently，haveinheritedsomethingincommon。Butthethreegeneraonthelefthandhave，onthissameprinciple，muchincommon，andformasub-family，distinctfromthatincludingthenexttwogeneraontherighthand，whichdivergedfromacommonparentatthefifthstageofdescent。Thesefivegenerahavealsomuch，thoughless，incommon。andtheyformafamilydistinctfromthatincludingthethreegenerastillfurthertotherighthand，whichdivergedatastillearlierperiod。Andallthesegenera，descendedfrom(A)，formanorderdistinctfromthegeneradescendedfrom(I)。Sothatweherehavemanyspeciesdescendedfromasingleprogenitorgroupedintogenera。andthegeneraareincludedin，orsubordinateto，sub-families，families，andorders，allunitedintooneclass。Thus，thegrandfactinnaturalhistoryofthesubordinationofgroupundergroup，which，fromitsfamiliarity，doesnotalwayssufficientlystrikeus，isinmyjudgementfullyexplained。

Naturaliststrytoarrangethespecies，genera，andfamiliesineachclass，onwhatiscalledtheNaturalSystem。Butwhatismeantbythissystem？Someauthorslookatitmerelyasaschemeforarrangingtogetherthoselivingobjectswhicharemostalike，andforseparatingthosewhicharemostunlike。orasanartificialmeansforenunciating，asbrieflyaspossible，generalpropositions，thatis，byonesentencetogivethecharacterscommon，forinstance，toallmammals，byanotherthosecommontoallcarnivora，byanotherthosecommontothedog-genus，andthenbyaddingasinglesentence，afulldescriptionisgivenofeachkindofdog。

Theingenuityandutilityofthissystemareindisputable。ButmanynaturaliststhinkthatsomethingmoreismeantbytheNaturalSystem。theybelievethatitrevealstheplanoftheCreator。butunlessitbespecifiedwhetherorderintimeorspace，orwhatelseismeantbytheplanoftheCreator，itseemstomethatnothingisthusaddedtoourknowledge。SuchexpressionsasthatfamousoneofLinnaeus，andwhichweoftenmeetwithinamoreorlessconcealedform，thatthecharactersdonotmakethegenus，butthatthegenusgivesthecharacters，seemtoimplythatsomethingmoreisincludedinourclassification，thanmereresemblance。Ibelievethatsomethingmoreisincluded。andthatpropinquityofdescent，theonlyknowncauseofthesimilarityoforganicbeings，isthebond，hiddenasitisbyvariousdegreesofmodification，whichispartiallyrevealedtousbyourclassifications。

Letusnowconsidertherulesfollowedinclassification，andthedifficultieswhichareencounteredontheviewthatclassificationeithergivessomeunknownplanofcreation，orissimplyaschemeforenunciatinggeneralpropositionsandofplacingtogethertheformsmostlikeeachother。Itmighthavebeenthought(andwasinancienttimesthought)thatthosepartsofthestructurewhichdeterminedthehabitsoflife，andthegeneralplaceofeachbeingintheeconomyofnature，wouldbeofveryhighimportanceinclassification。Nothingcanbemorefalse。Nooneregardstheexternalsimilarityofamousetoashrew，ofadugongtoawhale，ofawhaletoafish，asofanyimportance。Theseresemblances，thoughsointimatelyconnectedwiththewholelifeofthebeing，arerankedasmerely`adaptiveoranalogicalcharacters。’buttotheconsiderationoftheseresemblancesweshallhavetorecur。Itmayevenbegivenasageneralrule，thatthelessanypartoftheorganisationisconcernedwithspecialhabits，themoreimportantitbecomesforclassification。Asaninstance：Owen，inspeakingofthedugong，says，`Thegenerativeorgansbeingthosewhicharemostremotelyrelatedtothehabitsandfoodofananimal，Ihavealwaysregardedasaffordingveryclearindicationsofitstrueaffinities。Weareleastlikelyinthemodificationsoftheseorganstomistakeamerelyadaptiveforanessentialcharacter。’Sowithplants，howremarkableitisthattheorgansofvegetation，onwhichtheirwholelifedepends，areoflittlesignification，exceptinginthefirstmaindivisions。whereastheorgansofreproduction，withtheirproducttheseed，areofparamountimportance！

Wemustnot，therefore，inclassifying，trusttoresemblancesinpartsoftheorganisation，howeverimportanttheymaybeforthewelfareofthebeinginrelationtotheouterworld。Perhapsfromthiscauseithaspartlyarisen，thatalmostallnaturalistslaythegreateststressonresemblancesinorgansofhighvitalorphysiologicalimportance。Nodoubtthisviewoftheclassificatoryimportanceoforganswhichareimportantisgenerally，butbynomeansalways，true。Buttheirimportanceforclassification，Ibelieve，dependsontheirgreaterconstancythroughoutlargegroupsofspecies。andthisconstancydependsonsuchorganshavinggenerallybeensubjectedtolesschangeintheadaptationofthespeciestotheirconditionsoflife。Thatthemerephysiologicalimportanceofanorgandoesnotdeterminetheclassificatoryvalue，isalmostshownbytheonefact，thatinalliedgroups，inwhichthesameorgan，aswehaveeveryreasontosuppose，hasnearlythesamephysiologicalvalue，itsclassificatoryvalueiswidelydifferent。Nonaturalistcanhaveworkedatanygroupwithoutbeingstruckwiththisfact。andithasbeenmostfullyacknowledgedinthewritingsofalmosteveryauthor。Itwillsufficetoquotethehighestauthority，RobertBrown，whoinspeakingofcertainorgansintheProteaceae，saystheirgenericimportance，`likethatofalltheirparts，notonlyinthisbut，asIapprehend，ineverynaturalfamily，isveryunequal，andinsomecasesseemstobeentirelylost。’Againinanotherworkhesays，thegeneraoftheConnaraceae`differinhavingoneormoreovaria，intheexistenceorabsenceofalbumen，intheimbricateorvalvularaestivation。Anyoneofthesecharacterssinglyisfrequentlyofmorethangenericimportance，thoughhereevenwhenalltakentogethertheyappearinsufficienttoseparateCnestisfromConnarus。’Togiveanexampleamongstinsects，inonegreatdivisionoftheHymenoptera，theantennae，asWestwoodhasremarked，aremostconstantinstructure。inanotherdivisiontheydiffermuch，andthedifferencesareofquitesubordinatevalueinclassification。yetnooneprobablywillsaythattheantennaeinthesetwodivisionsofthesameorderareofunequalphysiologicalimportance。Anynumberofinstancescouldbegivenofthevaryingimportanceforclassificationofthesameimportantorganwithinthesamegroupofbeings。

Again，noonewillsaythatrudimentaryoratrophiedorgansareofhighphysiologicalorvitalimportance。yet，undoubtedly，organsinthisconditionareoftenofhighvalueinclassification。Noonewilldisputethattherudimentaryteethintheupperjawsofyoungruminants，andcertainrudimentarybonesoftheleg，arehighlyserviceableinexhibitingthecloseaffinitybetweenRuminantsandPachyderms。RobertBrownhasstronglyinsistedonthefactthattherudimentaryfloretsareofthehighestimportanceintheclassificationoftheGrasses。

Numerousinstancescouldbegivenofcharactersderivedfrompartswhichmustbeconsideredofverytriflingphysiologicalimportance，butwhichareuniversallyadmittedashighlyserviceableinthedefinitionofwholegroups。Forinstance，whetherornotthereisanopenpassagefromthenostrilstothemouth，theonlycharacter，accordingtoOwen，whichabsolutelydistinguishesfishesandreptilestheinflectionoftheangleofthejawsinMarsupials——themannerinwhichthewingsofinsectsarefoldedmerecolourincertainAlgaemerepubescenceonpartsofthefloweringrassesthenatureofthedermalcovering，ashairorfeathers，intheVertebrata。

IftheOrnithorhynchushadbeencoveredwithfeathersinsteadofhair，thisexternalandtriflingcharacterwould，Ithink，havebeenconsideredbynaturalistsasimportantanaidindeterminingthedegreeofaffinityofthisstrangecreaturetobirdsandreptiles，asanapproachinstructureinanyoneinternalandimportantorgan。

Theimportance，forclassification，oftriflingcharacters，mainlydependsontheirbeingcorrelatedwithseveralothercharactersofmoreorlessimportance。Thevalueindeedofanaggregateofcharactersisveryevidentinnaturalhistory。Hence，ashasoftenbeenremarked，aspeciesmaydepartfromitsalliesinseveralcharacters，bothofhighphysiologicalimportanceandofalmostuniversalprevalence，andyetleaveusinnodoubtwhereitshouldberanked。Hence，also，ithasbeenfound，thataclassificationfoundedonanysinglecharacter，howeverimportantthatmaybe，hasalwaysfailed。fornopartoftheorganisationisuniversallyconstant。Theimportanceofanaggregateofcharacters，evenwhennoneareimportant，aloneexplains，Ithink，thatsayingofLinnaeus，thatthecharactersdonotgivethegenus，butthegenusgivesthecharacters。forthissayingseemsfoundedonanappreciationofmanytriflingpointsofresemblance，tooslighttobedefined。

Certainplants，belongingtotheMalpighiaceae，bearperfectanddegradedflowers。inthelatter，asA。deJussieuhasremarked，`thegreaternumberofthecharacterspropertothespecies，tothegenus，tothefamily，totheclass，disappear，andthuslaughatourclassification。’ButwhenAspicarpaproducedinFrance，duringseveralyears，onlydegradedflowers，departingsowonderfullyinanumberofthemostimportantpointsofstructurefromthepropertypeoftheorder，yetM。Richardsagaciouslysaw，asJussieuobserves，thatthisgenusshouldstillberetainedamongsttheMalpighiaceae。

Thiscaseseemstomewelltoillustratethespiritwithwhichourclassificationsaresometimesnecessarilyfounded。

Practicallywhennaturalistsareatwork，theydonottroublethemselvesaboutthephysiologicalvalueofthecharacterswhichtheyuseindefiningagroup，orinallocatinganyparticularspecies。Iftheyfindacharacternearlyuniform，andcommontoagreatnumberofforms，andnotcommontoothers，theyuseitasoneofhighvalue。ifcommontosomelessernumber，theyuseitasofsubordinatevalue。Thisprinciplehasbeenbroadlyconfessedbysomenaturaliststobethetrueone。andbynonemoreclearlythanbythatexcellentbotanist，Aug。St。Hilaire。Ifcertaincharactersarealwaysfoundcorrelatedwithothers，thoughnoapparentbondofconnexioncanbediscoveredbetweenthem，especialvalueissetonthem。Asinmostgroupsofanimals，importantorgans，suchasthoseforpropellingtheblood，orforaë。ratingit，orthoseforpropagatingtherace，arefoundnearlyuniform，theyareconsideredashighlyserviceableinclassification。butinsomegroupsofanimalsallthese，themostimportantvitalorgans，arefoundtooffercharactersofquitesubordinatevalue。

Wecanseewhycharactersderivedfromtheembryoshouldbeofequalimportancewiththosederivedfromtheadult，forourclassificationsofcourseincludeallagesofeachspecies。Butitisbynomeansobvious，ontheordinaryview，whythestructureoftheembryoshouldbemoreimportantforthispurposethanthatoftheadult，whichaloneplaysitsfullpartintheeconomyofnature。Yetithasbeenstronglyurgedbythosegreatnaturalists，MilneEdwardsandAgassiz，thatembryoniccharactersarethemostimportantofanyintheclassificationofanimals。andthisdoctrinehasverygenerallybeenadmittedastrue。Thesamefactholdsgoodwithfloweringplants，ofwhichthetwomaindivisionshavebeenfoundedoncharactersderivedfromtheembryo，onthenumberandpositionoftheembryonicleavesorcotyledons，andonthemodeofdevelopmentoftheplumuleandradicle。Inourdiscussiononembryology，weshallseewhysuchcharactersaresovaluable，ontheviewofclassificationtacitlyincludingtheideaofdescent。

Ourclassificationsareoftenplainlyinfluencedbychainsofaffinities。

Nothingcanbeeasierthantodefineanumberofcharacterscommontoallbirds。butinthecaseofcrustaceans，suchdefinitionhashithertobeenfoundimpossible。Therearecrustaceansattheoppositeendsoftheseries，whichhavehardlyacharacterincommon。yetthespeciesatbothends，frombeingplainlyalliedtoothers，andthesetoothers，andsoonwards，canberecognisedasunequivocallybelongingtothis，andtonootherclassoftheArticulata。

Geographicaldistributionhasoftenbeenused，thoughperhapsnotquitelogically，inclassification，moreespeciallyinverylargegroupsofcloselyalliedforms。Temminckinsistsontheutilityorevennecessityofthispracticeincertaingroupsofbirds。andithasbeenfollowedbyseveralentomologistsandbotanists。

Finally，withrespecttothecomparativevalueofthevariousgroupsofspecies，suchasorders，sub-orders，families，sub-families，andgenera，theyseemtobe，atleastatpresent，almostarbitrary。Severalofthebestbotanists，suchasMrBenthamandothers，havestronglyinsistedontheirarbitraryvalue。Instancescouldbegivenamongstplantsandinsects，ofagroupofforms，firstrankedbypractisednaturalistsasonlyagenus，andthenraisedtotherankofasub-familyorfamily。andthishasbeendone，notbecausefurtherresearchhasdetectedimportantstructuraldifferences，atfirstoverlooked，butbecausenumerousalliedspecies，withslightlydifferentgradesofdifference，havebeensubsequentlydiscovered。

Alltheforegoingrulesandaidsanddifficultiesinclassificationareexplained，ifIdonotgreatlydeceivemyself，ontheviewthatthenaturalsystemisfoundedondescentwithmodification。thatthecharacterswhichnaturalistsconsiderasshowingtrueaffinitybetweenanytwoormorespecies，arethosewhichhavebeeninheritedfromacommonparent，and，insofar，alltrueclassificationisgenealogical。thatcommunityofdescentisthehiddenbondwhichnaturalistshavebeenunconsciouslyseeking，andnotsomeunknownplanofcreation，ortheenunciationofgeneralpropositions，andthemereputtingtogetherandseparatingobjectsmoreorlessalike。

ButImustexplainmymeaningmorefully。Ibelievethatthearrangementofthegroupswithineachclass，induesubordinationandrelationtotheothergroups，mustbestrictlygenealogicalinordertobenatural。butthattheamountofdifferenceintheseveralbranchesorgroups，thoughalliedinthesamedegreeinbloodtotheircommonprogenitor，maydiffergreatly，beingduetothedifferentdegreesofmodificationwhichtheyhaveundergone。andthisisexpressedbytheformsbeingrankedunderdifferentgenera，families，sections，ororders。Thereaderwillbestunderstandwhatismeant，ifhewilltakethetroubleofreferringtothediagraminthefourthchapter。WewillsupposethelettersAtoLtorepresentalliedgenera，whichlivedduringtheSilurianepoch，andthesehavedescendedfromaspecieswhichexistedatanunknownanteriorperiod。Speciesofthreeofthesegenera(A，F，andI)havetransmittedmodifieddescendantstothepresentday，representedbythefifteengenera(a14toz14)ontheuppermosthorizontalline。Nowallthesemodifieddescendantsfromasinglespecies，arerepresentedasrelatedinbloodordescenttothesamedegree。

theymaymetaphoricallybecalledcousinstothesamemillionthdegree。

yettheydifferwidelyandindifferentdegreesfromeachother。TheformsdescendedfromA，nowbrokenupintotwoorthreefamilies，constituteadistinctorderfromthosedescendedfromI，alsobrokenupintotwofamilies。

Norcantheexistingspecies，descendedfromA，berankedinthesamegenuswiththeparentA。orthosefromI，withtheparentI。ButtheexistinggenusF14maybesupposedtohavebeenbutslightlymodified。anditwillthenrankwiththeparent-genusF。justassomefewstilllivingorganicbeingsbelongtoSiluriangenera。Sothattheamountorvalueofthedifferencesbetweenorganicbeingsallrelatedtoeachotherinthesamedegreeinblood，hascometobewidelydifferent。Neverthelesstheirgenealogicalarrangementremainsstrictlytrue，notonlyatthepresenttime，butateachsuccessiveperiodofdescent。AllthemodifieddescendantsfromAwillhaveinheritedsomethingincommonfromtheircommonparent，aswillallthedescendantsfromI。sowillitbewitheachsubordinatebranchofdescendants，ateachsuccessiveperiod。If，however，wechoosetosupposethatanyofthedescendantsofAorofIhavebeensomuchmodifiedastohavemoreorlesscompletelylosttracesoftheirparentage，inthiscase，theirplacesinanaturalclassificationwillhavebeenmoreorlesscompletelylost，assometimesseemstohaveoccurredwithexistingorganisms。

AllthedescendantsofthegenusF，alongitswholelineofdescent，aresupposedtohavebeenbutlittlemodified，andtheyyetformasinglegenus。

Butthisgenus，thoughmuchisolated，willstilloccupyitsproperintermediateposition。forForiginallywasintermediateincharacterbetweenAandI，andtheseveralgeneradescendedfromthesetwogenerawillhaveinheritedtoacertainextenttheircharacters。Thisnaturalarrangementisshown，asfarasispossibleonpaper，inthediagram，butinmuchtoosimpleamanner。Ifabranchingdiagramhadnotbeenused，andonlythenamesofthegroupshadbeenwritteninalinearseries，itwouldhavebeenstilllesspossibletohavegivenanaturalarrangement。anditisnotoriouslynotpossibletorepresentinaseries，onaflatsurface，theaffinitieswhichwediscoverinnatureamongstthebeingsofthesamegroup。Thus，ontheviewwhichIhold，thenaturalsystemisgenealogicalinitsarrangement，likeapedigree。butthedegreesofmodificationwhichthedifferentgroupshaveundergone，havetobeexpressedbyrankingthemunderdifferentso-calledgenera，sub-families，families，sections，orders，andclasses。

Itmaybeworthwhiletoillustratethisviewofclassification，bytakingthecaseoflanguages。Ifwepossessedaperfectpedigreeofmankind，agenealogicalarrangementoftheracesofmanwouldaffordthebestclassificationofthevariouslanguagesnowspokenthroughouttheworld。andifallextinctlanguages，andallintermediateandslowlychangingdialects，hadtobeincluded，suchanarrangementwould，Ithink，betheonlypossibleone。

Yetitmightbethatsomeveryancientlanguagehadalteredlittle，andhadgivenrisetofewnewlanguages，whilstothers(owingtothespreadingandsubsequentisolationandstatesofcivilisationoftheseveralraces，descendedfromacommonrace)hadalteredmuch，andhadgivenrisetomanynewlanguagesanddialects。Thevariousdegreesofdifferenceinthelanguagesfromthesamestock，wouldhavetobeexpressedbygroupssubordinatetogroups。buttheproperorevenonlypossiblearrangementwouldstillbegenealogical。andthiswouldbestrictlynatural，asitwouldconnecttogetheralllanguages，extinctandmodern，bytheclosestaffinities，andwouldgivethefiliationandoriginofeachtongue。

Inconfirmationofthisview，letusglanceattheclassificationofvarieties，whicharebelievedorknowntohavedescendedfromonespecies。

Thesearegroupedunderspecies，withsub-varietiesundervarieties。andwithourdomesticproductions，severalothergradesofdifferencearerequisite，aswehaveseenwithpigeons。Theoriginoftheexistenceofgroupssubordinatetogroups，isthesamewithvarietiesaswithspecies，namely，closenessofdescentwithvariousdegreesofmodification。Nearlythesamerulesarefollowedinclassifyingvarieties，aswithspecies。Authorshaveinsistedonthenecessityofclassingvarietiesonanaturalinsteadofanartificialsystem。wearecautioned，forinstance，nottoclasstwovarietiesofthepine-appletogether，merelybecausetheirfruit，thoughthemostimportantpart，happenstobenearlyidentical。nooneputstheswedishandcommonturnipstogether，thoughtheesculentandthickenedstemsaresosimilar。

Whateverpartisfoundtobemostconstant，isusedinclassingvarieties：

thusthegreatagriculturistMarshallsaysthehornsareveryusefulforthispurposewithcattle，becausetheyarelessvariablethantheshapeorcolourofthebody，&。c……whereaswithsheepthehornsaremuchlessserviceable，becauselessconstant。Inclassingvarieties，Iapprehendifwehadarealpedigree，agenealogicalclassificationwouldbeuniversallypreferred。andithasbeenattemptedbysomeauthors。Forwemightfeelsure，whethertherehadbeenmoreorlessmodification，theprincipleofinheritancewouldkeeptheformstogetherwhichwerealliedinthegreatestnumberofpoints。Intumblerpigeons，thoughsomesub-varietiesdifferfromtheothersintheimportantcharacterofhavingalongerbeak，yetallarekepttogetherfromhavingthecommonhabitoftumbling。buttheshort-facedbreedhasnearlyorquitelostthishabit。nevertheless，withoutanyreasoningorthinkingonthesubject，thesetumblersarekeptinthesamegroup，becausealliedinbloodandalikeinsomeotherrespects。IfitcouldbeprovedthattheHottentothaddescendedfromtheNegro，IthinkhewouldbeclassedundertheNegrogroup，howevermuchhemightdifferincolourandotherimportantcharactersfromnegroes。

Withspeciesinastateofnature，everynaturalisthasinfactbroughtdescentintohisclassification。forheincludesinhislowestgrade，orthatofaspecies，thetwosexes。andhowenormouslythesesometimesdifferinthemostimportantcharacters，isknowntoeverynaturalist：scarcelyasinglefactcanbepredicatedincommonofthemalesandhermaphroditesofcertaincirripedes，whenadult，andyetnoonedreamsofseparatingthem。Thenaturalistincludesasonespeciestheseverallarvalstagesofthesameindividual，howevermuchtheymaydifferfromeachotherandfromtheadult。ashelikewiseincludestheso-calledalternategenerationsofSteenstrup，whichcanonlyinatechnicalsensebeconsideredasthesameindividual。Heincludesmonsters。heincludesvarieties，notsolelybecausetheycloselyresembletheparent-form，butbecausetheyaredescendedfromit。Hewhobelievesthatthecowslipisdescendedfromtheprimrose，orconversely，ranksthemtogetherasasinglespecies，andgivesasingledefinition。AssoonasthreeOrchideanforms(Monochanthus，Myanthus，andCatasetum)，whichhadpreviouslybeenrankedasthreedistinctgenera，wereknowntobesometimesproducedonthesamespike，theywereimmediatelyincludedasasinglespecies。Butitmaybeasked，whatoughtwetodo，ifitcouldbeprovedthatonespeciesofkangaroohadbeenproduced，byalongcourseofmodification，fromabear？Oughtwetorankthisonespecieswithbears，andwhatshouldwedowiththeotherspecies？Thesuppositionisofcoursepreposterous。andImightanswerbytheargumentumadhominem，andaskwhatshouldbedoneifaperfectkangaroowereseentocomeoutofthewombofabear？Accordingtoallanalogy，itwouldberankedwithbears。butthenassuredlyalltheotherspeciesofthekangaroofamilywouldhavetobeclassedunderthebeargenus。Thewholecaseispreposterous。

forwheretherehasbeenclosedescentincommon，therewillcertainlybecloseresemblanceoraffinity。

Asdescenthasuniversallybeenusedinclassingtogethertheindividualsofthesamespecies，thoughthemalesandfemalesandlarvaearesometimesextremelydifferent。andasithasbeenusedinclassingvarietieswhichhaveundergoneacertain，andsometimesaconsiderableamountofmodification，maynotthissameelementofdescenthavebeenunconsciouslyusedingroupingspeciesundergenera，andgeneraunderhighergroups，thoughinthesecasesthemodificationhasbeengreaterindegree，andhastakenalongertimetocomplete？Ibelieveithasthusbeenunconsciouslyused。andonlythuscanIunderstandtheseveralrulesandguideswhichhavebeenfollowedbyourbestsystematists。Wehavenowrittenpedigrees。wehavetomakeoutcommunityofdescentbyresemblancesofanykind。Thereforewechoosethosecharacterswhich，asfaraswecanjudge，aretheleastlikelytohavebeenmodifiedinrelationtotheconditionsoflifetowhicheachspecieshasbeenrecentlyexposed。Rudimentarystructuresonthisviewareasgoodas，orevensometimesbetterthan，otherpartsoftheorganisation。

Wecarenothowtriflingacharactermaybeletitbethemereinflectionoftheangleofthejaw，themannerinwhichaninsect’swingisfolded，whethertheskinbecoveredbyhairorfeathersifitprevailthroughoutmanyanddifferentspecies，especiallythosehavingverydifferenthabitsoflife，itassumeshighvalue。forwecanaccountforitspresenceinsomanyformswithsuchdifferenthabits，onlybyitsinheritancefromacommonparent。Wemayerrinthisrespectinregardtosinglepointsofstructure，butwhenseveralcharacters，letthembeeversotrifling，occurtogetherthroughoutalargegroupofbeingshavingdifferenthabits，wemayfeelalmostsure，onthetheoryofdescent，thatthesecharactershavebeeninheritedfromacommonancestor。Andweknowthatsuchcorrelatedoraggregatedcharactershaveespecialvalueinclassification。

Wecanunderstandwhyaspeciesoragroupofspeciesmaydepart，inseveralofitsmostimportantcharacteristics，fromitsallies，andyetbesafelyclassedwiththem。Thismaybesafelydone，andisoftendone，aslongasasufficientnumberofcharacters，letthembeeversounimportant，betraysthehiddenbondofcommunityofdescent。Lettwoformshavenotasinglecharacterincommon，yetiftheseextremeformsareconnectedtogetherbyachainofintermediategroups，wemayatonceinfertheircommunityofdescent，andweputthemallintothesameclass。Aswefindorgansofhighphysiologicalimportancethosewhichservetopreservelifeunderthemostdiverseconditionsofexistencearegenerallythemostconstant，weattachespecialvaluetothem。butifthesesameorgans，inanothergrouporsectionofagroup，arefoundtodiffermuch，weatoncevaluethemlessinourclassification。Weshallhereafter，Ithink，clearlyseewhyembryologicalcharactersareofsuchhighclassificatoryimportance。

Geographicaldistributionmaysometimesbebroughtusefullyintoplayinclassinglargeandwidely-distributedgenera，becauseallthespeciesofthesamegenus，inhabitinganydistinctandisolatedregion，haveinallprobabilitydescendedfromthesameparents。