Agricult。Soc。’volume4page43。）Althoughfreecrossingisadangerontheonesidewhicheveryonecansee，toocloseinterbreedingisahiddendangerontheotherside。Unfavourableconditionsoflifeoverrulethepowerofselection。Ourimprovedheavybreedsofcattleandsheepcouldnothavebeenformedonmountainouspastures;norcoulddray—horseshavebeenraisedonabarrenandinhospitableland，suchastheFalklandIslands，whereeventhelighthorsesofLaPlatarapidlydecreaseinsize。ItseemsimpossibletopreserveseveralEnglishbreedsofsheepinFrance;forassoonasthelambsareweanedtheirvigourdecaysastheheatofthesummerincreases（21/42。
Malingie—Nouel’JournalR。Agricult。Soc。’volume141853pages215，217。）:itwouldbeimpossibletogivegreatlengthofwooltosheepwithinthetropics;
yetselectionhaskepttheMerinobreednearlytrueunderdiversifiedandunfavourableconditions。Thepowerofselectionissogreat，thatbreedsofthedog，sheep，andpoultry，ofthelargestandsmallestsize，longandshortbeakedpigeons，andotherbreedswithoppositecharacters，havehadtheircharacteristicqualitiesaugmented，thoughtreatedineverywayalike，beingexposedtothesameclimateandfedonthesamefood。Selection，however，iseithercheckedorfavouredbytheeffectsofuseorhabit。Ourwonderfully—
improvedpigscouldneverhavebeenformediftheyhadbeenforcedtosearchfortheirownfood;theEnglishracehorseandgreyhoundcouldnothavebeenimproveduptotheirpresenthighstandardofexcellencewithoutconstanttraining。
Asconspicuousdeviationsofstructureoccurrarely，theimprovementofeachbreedisgenerallytheresultoftheselectionofslightindividualdifferences。Hencetheclosestattention，thesharpestpowersofobservation，andindomitableperseverance，areindispensable。Itis，also，highlyimportantthatmanyindividualsofthebreedwhichistobeimprovedshouldberaised;
forthustherewillbeabetterchanceoftheappearanceofvariationsintherightdirection，andindividualsvaryinginanunfavourablemannermaybefreelyrejectedordestroyed。Butthatalargenumberofindividualsshouldberaised，itisnecessarythattheconditionsoflifeshouldfavourthepropagationofthespecies。Hadthepeacockbeenrearedaseasilyasthefowl，weshouldprobablyerethishavehadmanydistinctraces。Weseetheimportanceofalargenumberofplants，fromthefactofnurserygardenersalmostalwaysbeatingamateursintheexhibitionofnewvarieties。In1845itwasestimated（21/43。Gardener’sChronicle’1845page273。）thatbetween4000
and5000pelargoniumswereannuallyraisedfromseedinEngland，yetadecidedlyimprovedvarietyisrarelyobtained。AtMessrs。Carter’sgrounds，inEssex，wheresuchflowersastheLobelia，Nemophila，Mignonette，etc。，aregrownbytheacreforseed，\"scarcelyaseasonpasseswithoutsomenewkindsbeingraised，orsomeimprovementeffectedonoldkinds。\"（21/44。’JournalofHorticulture’1862page157。）AtKew，asMr。Beatonremarks，wheremanyseedlingsofcommonplantsareraised，\"youseenewformsofLaburnums，Spiraeas，andothershrubs。\"（21/45。’CottageGardener’1860page368。）Sowithanimals:Marshall（21/46。’AReviewofReports’1808page406。），inspeakingofthesheepinonepartofYorkshire，remarks，\"astheybelongtopoorpeople，andaremostlyinsmalllots，theynevercanbeimproved。\"LordRivers，whenaskedhowhesucceededinalwayshavingfirst—rategreyhounds，answered，\"Ibreedmany，andhangmany。\"This，asanothermanremarks，\"wasthesecretofhissuccess;andthesamewillbefoundinexhibitingfowls，——
successfulcompetitorsbreedlargely，andkeepthebest。\"（21/47。’Gardener’sChronicle’1853page45。）
Itfollowsfromthisthatthecapacityofbreedingatanearlyageandatshortintervals，aswithpigeons，rabbits，etc。，facilitatesselection;fortheresultisthussoonmadevisible，andperseveranceintheworkencouraged。
Itcanhardlybeanaccidentthatthegreatmajorityoftheculinaryandagriculturalplantswhichhaveyieldednumerousracesareannualsorbiennials，whichthereforearecapableofrapidpropagation，andthusofimprovement。Sea—kale，asparagus，commonandJerusalemartichokes，potatoes，andonions，mustbeexcepted，astheyareperennials:butonionsarepropagatedlikeannuals，andoftheotherplantsjustspecified，none，withtheexceptionofthepotato，haveyieldedinthiscountrymorethanoneortwovarieties。IntheMediterraneanregion，whereartichokesareoftenraisedfromseed，thereareseveralkinds，asIhearfromMr。Bentham。Nodoubtfruit—
trees，whichcannotbepropagatedquicklybyseed，haveyieldedahostofvarieties，thoughnotpermanentraces;butthese，judgingfromprehistoricremains，havebeenproducedatacomparativelylateperiod。
Aspeciesmaybehighlyvariable，butdistinctraceswillnotbeformed，iffromanycauseselectionbenotapplied。Itwouldbedifficulttoselectslightvariationsinfishesfromtheirplaceofhabitation;andthoughthecarpisextremelyvariableandismuchattendedtoinGermany，onlyonewell—
markedracehasbeenformed，asIhearfromLordA。Russell，namelythespiegel—carpe;andthisiscarefullysecludedfromthecommonscalykind。Ontheotherhand，acloselyalliedspecies，thegold—fish，frombeingrearedinsmallvessels，andfromhavingbeencarefullyattendedtobytheChinese，hasyieldedmanyraces。Neitherthebee，whichhasbeensemi—domesticatedfromanextremelyremoteperiod，northecochinealinsect，whichwascultivatedbytheaboriginalMexicans（21/48。IsidoreGeoffroySaint—Hilaire’Hist。Nat。Gen。’
tome3page49。’OntheCochinealInsect’page46。），hasyieldedraces;anditwouldbeimpossibletomatchthequeen—beewithanyparticulardrone，andmostdifficulttomatchcochinealinsects。Silk—moths，ontheotherhand，havebeensubjectedtorigorousselection，andhaveproducedahostofraces。Cats，whichfromtheirnocturnalhabitscannotbeselectedforbreeding，donot，asformerlyremarked，yielddistinctraceswithinthesamecountry。DogsareheldinabominationintheEast，andtheirbreedingisneglected;consequently，asProf。MoritzWagner（21/49。’DieDarwin’scheTheorieunddasMigrationsgesetzderOrganismen’1868page19。）remarks，onekindaloneexiststhere。TheassinEnglandvariesmuchincolourandsize;butasitisananimaloflittlevalueandbredbypoorpeople，therehasbeennoselection，anddistinctraceshavenotbeenformed。Wemustnotattributetheinferiorityofourassestoclimate，forinIndiatheyareofevensmallersizethaninEurope。Butwhenselectionisbroughttobearontheass，allischanged。NearCordova，asIaminformed（February1860）byMr。W。E。Webb，C。E。，theyarecarefullybred，asmuchas200poundshavingbeenpaidforastallionass，andtheyhavebeenimmenselyimproved。InKentucky，asseshavebeenimported（forbreedingmules）
fromSpain，Malta，andFrance;these\"seldomaveragedmorethanfourteenhandshigh:buttheKentuckians，bygreatcare，haveraisedthemuptofifteenhands，andsometimeseventosixteen。Thepricespaidforthesesplendidanimals，forsuchtheyreallyare，willprovehowmuchtheyareinrequest。
Onemale，ofgreatcelebrity，wassoldforupwardsofonethousandpoundssterling。\"Thesechoiceassesaresenttocattle—shows，adaybeinggivenfortheirexhibition。（21/50。Capt。MarryatquotedbyBlythin’Journ。AsiaticSoc。ofBengal’volume28page229。）
Analogousfactshavebeenobservedwithplants:thenutmeg—treeintheMalayarchipelagoishighlyvariable，buttherehasbeennoselection，andtherearenodistinctraces。（21/51。Mr。Oxley’JournaloftheIndianArchipelago’
volume21848page645。）Thecommonmignonette（Resedaodorata），frombearinginconspicuousflowers，valuedsolelyfortheirfragrance，\"remainsinthesameunimprovedconditionaswhenfirstintroduced。\"（21/52。Mr。Abbey’JournalofHorticulture’December1，1863page430。）Ourcommonforest—treesareveryvariable，asmaybeseenineveryextensivenursery—ground;butastheyarenotvaluedlikefruit—trees，andastheyseedlateinlife，noselectionhasbeenappliedtothem;consequently，asMr。PatrickMatthewsremarks（21/53。
’OnNavalTimber’1831page107。），theyhavenotyieldeddistinctraces，leafingatdifferentperiods，growingtodifferentsizes，andproducingtimberfitfordifferentpurposes。Wehavegainedonlysomefancifulandsemi—
monstrousvarieties，whichnodoubtappearedsuddenlyaswenowseethem。
Somebotanistshavearguedthatplantscannothavesostrongatendencytovaryasisgenerallysupposed，becausemanyspecieslonggrowninbotanicgardens，orunintentionallycultivatedyearafteryearmingledwithourcorncrops，havenotproduceddistinctraces;butthisisaccountedforbyslightvariationsnothavingbeenselectedandpropagated。Letaplantwhichisnowgrowninabotanicgarden，oranycommonweed，becultivatedonalargescale，andletasharp—sightedgardenerlookoutforeachslightvarietyandsowtheseed，andthen，ifdistinctracesarenotproduced，theargumentwillbevalid。
Theimportanceofselectionislikewiseshownbyconsideringspecialcharacters。Forinstance，withmostbreedsoffowlstheformofthecombandthecolouroftheplumagehavebeenattendedto，andareeminentlycharacteristicofeachrace;butinDorkingsfashionhasneverdemandeduniformityofcomborcolour;andtheutmostdiversityintheserespectsprevails。Rose—combs，double—combs，cup—combs，etc。，andcoloursofallkinds，maybeseeninpurelybredandcloselyrelatedDorkingfowls，whilstotherpoints，suchasthegeneralformofbody，andthepresenceofanadditionaltoe，havebeenattendedto，andareinvariablypresent。Ithasalsobeenascertainedthatcolourcanbefixedinthisbreed，aswellasinanyother。
（21/54。Mr。Bailyin’ThePoultryChronicle’volume21854page150。Alsovolume1page342;volume3page245。）
Duringtheformationorimprovementofabreed，itsmemberswillalwaysbefoundtovarymuchinthosecharacterstowhichespecialattentionisdirected，andofwhicheachslightimprovementiseagerlysoughtandselected。
Thus，withshort—facedtumbler—pigeons，theshortnessofthebeak，shapeofheadandplumage，——withcarriers，thelengthofthebeakandwattle，——withfantails，thetailandcarriage，——withSpanishfowls，thewhitefaceandcomb，——withlong—earedrabbits，thelengthofear，areallpointswhichareeminentlyvariable。Soitisineverycase;andthelargepricepaidforfirst—rateanimalsprovesthedifficultyofbreedingthemuptothehigheststandardofexcellence。Thissubjecthasbeendiscussedbyfanciers（21/55。
’CottageGardener’1855Decemberpage171;1856Januarypages248，323。），andthegreaterprizesgivenforhighlyimprovedbreeds，incomparisonwiththosegivenforoldbreedswhicharenotnowundergoingrapidimprovement，havebeenfullyjustified。Nathusiusmakes（21/56。’UeberShorthornRindvieh’1857s。
51。）asimilarremarkwhendiscussingthelessuniformcharacterofimprovedShorthorncattleandoftheEnglishhorse，incomparison，forexample，withtheunennobledcattleofHungary，orwiththehorsesoftheAsiaticsteppes。
Thiswantofuniformityinthepartswhichatthetimeareundergoingselectionchieflydependsonthestrengthoftheprincipleofreversion;butitlikewisedependstoacertainextentonthecontinuedvariabilityofthepartswhichhaverecentlyvaried。Thatthesamepartsdocontinuevaryinginthesamemannerwemustadmit，forifitwerenotso，therecouldbenoimprovementbeyondanearlystandardofexcellence，andweknowthatsuchimprovementisnotonlypossible，butisofgeneraloccurrence。
Asaconsequenceofcontinuedvariability，andmoreespeciallyofreversion，allhighlyimprovedraces，ifneglectedornotsubjectedtoincessantselection，soondegenerate。YouattgivesacuriousinstanceofthisinsomecattleformerlykeptinGlamorganshire;butinthiscasethecattlewerenotfedwithsufficientcare。Mr。Baker，inhismemoirontheHorse，sumsup:\"Itmusthavebeenobservedintheprecedingpagesthat，whenevertherehasbeenneglect，thebreedhasproportionallydeteriorated。\"（21/57。’TheVeterinary’
volume13page720。FortheGlamorganshirecattleseeYouatton’Cattle’page51。）Ifaconsiderablenumberofimprovedcattle，sheep，orotheranimalsofthesamerace，wereallowedtobreedfreelytogether，withnoselection，butwithnochangeintheirconditionoflife，therecanbenodoubtthatafterascoreorhundredgenerationstheywouldbeveryfarfromexcellentoftheirkind;but，fromwhatweseeofthemanycommonracesofdogs，cattle，fowls，pigeons，etc。，whichwithoutanyparticularcarehavelongretainednearlythesamecharacter，wehavenogroundsforbelievingthattheywouldaltogetherdepartfromtheirtype。
Itisageneralbeliefamongstbreedersthatcharactersofallkindsbecomefixedbylong—continuedinheritance。ButIhaveattemptedtoshowinthefourteenthchapterthatthisbeliefapparentlyresolvesitselfintothefollowingproposition，namely，thatallcharacterswhatever，whetherrecentlyacquiredorancient，tendtobetransmitted，butthatthosewhichhavealreadylongwithstoodallcounteractinginfluences，will，asageneralrule，continuetowithstandthem，andconsequentlybefaithfullytransmitted。
TENDENCYINMANTOCARRYTHEPRACTICEOFSELECTIONTOANEXTREMEPOINT。
Itisanimportantprinciplethatintheprocessofselectionmanalmostinvariablywishestogotoanextremepoint。Thus，thereisnolimittohisdesiretobreedcertainkindsofhorsesanddogsasfleetaspossible，andothersasstrongaspossible;certainkindsofsheepforextremefineness，andothersforextremelengthofwool;andhewishestoproducefruit，grain，tubers，andotherusefulpartsofplants，aslargeandexcellentaspossible。
Withanimalsbredforamusement，thesameprincipleisevenmorepowerful;forfashion，asweseeinourdress，alwaysrunstoextremes。Thisviewhasbeenexpresslyadmittedbyfanciers。Instancesweregiveninthechaptersonthepigeon，buthereisanother:Mr。Eaton，afterdescribingacomparativelynewvariety，namely，theArchangel，remarks，\"WhatfanciersintenddoingwiththisbirdIamatalosstoknow，whethertheyintendtobreeditdowntothetumbler’sheadandbeak，orcarryitouttothecarrier’sheadandbeak;
leavingitastheyfoundit，isnotprogressing。\"Ferguson，speakingoffowls，says，\"theirpeculiarities，whatevertheymaybe，mustnecessarilybefullydeveloped:alittlepeculiarityformsnoughtbutugliness，seeingitviolatestheexistinglawsofsymmetry。\"SoMr。Brent，indiscussingthemeritsofthesub—varietiesoftheBelgiancanary—bird，remarks，\"Fanciersalwaysgotoextremes;theydonotadmireindefiniteproperties。\"（21/58。J。M。Eaton’A
TreatiseonFancyPigeons’page82;Fergusonon’RareandPrizePoultry’page162;Mr。Brentin’CottageGardener’October1860page13。）
Thisprinciple，whichnecessarilyleadstodivergenceofcharacter，explainsthepresentstateofvariousdomesticraces。Wecanthusseehowitisthatracehorsesanddray—horses，greyhoundsandmastiffs，whichareopposedtoeachotherineverycharacter，——howvarietiessodistinctasCochin—chinafowlsandbantams，orcarrier—pigeonswithverylongbeaks，andtumblerswithexcessivelyshortbeaks，havebeenderivedfromthesamestock。Aseachbreedisslowlyimproved，theinferiorvarietiesarefirstneglectedandfinallylost。Inafewcases，bytheaidofoldrecords，orfromintermediatevarietiesstillexistingincountrieswhereotherfashionshaveprevailed，weareenabledpartiallytotracethegraduatedchangesthroughwhichcertainbreedshavepassed。Selection，whethermethodicalorunconscious，alwaystendingtowardsanextremepoint，togetherwiththeneglectandslowextinctionoftheintermediateandless—valuedforms，isthekeywhichunlocksthemysteryofhowmanhasproducedsuchwonderfulresults。
Inafewinstancesselection，guidedbyutilityforasinglepurpose，hasledtoconvergenceofcharacter。Alltheimprovedanddifferentracesofthepig，asNathusiushaswellshown（21/59。’DieRacendesSchweines’1860s。48。），closelyapproacheachotherincharacter，intheirshortenedlegsandmuzzles，theiralmosthairless，large，roundedbodies，andsmalltusks。Weseesomedegreeofconvergenceinthesimilaroutlineofthebodyinwell—bredcattlebelongingtodistinctraces。（21/60。SeesomegoodremarksonthisheadbyM。
deQuatrefages’Unitedel’EspeceHumaine’1861page119。）Iknowofnoothersuchcases。
Continueddivergenceofcharacterdependson，andisindeedaclearproof，aspreviouslyremarked，ofthesamepartscontinuingtovaryinthesamedirection。Thetendencytomeregeneralvariabilityorplasticityoforganisationcancertainlybeinherited，evenfromoneparent，ashasbeenshownbyGartnerandKolreuter，intheproductionofvaryinghybridsfromtwospecies，ofwhichonealonewasvariable。Itisinitselfprobablethat，whenanorganhasvariedinanymanner，itwillagainvaryinthesamemanner，iftheconditionswhichfirstcausedthebeingtovaryremain，asfarascanbejudged，thesame。Thisiseithertacitlyorexpresslyadmittedbyallhorticulturists:ifagardenerobservesoneortwoadditionalpetalsinaflower，hefeelsconfidentthatinafewgenerationshewillbeabletoraiseadoubleflower，crowdedwithpetals。SomeoftheseedlingsfromtheweepingMoccasoakweresoprostratethattheyonlycrawledalongtheground。A
seedlingfromthefastigiateoruprightIrishyewisdescribedasdifferinggreatlyfromtheparent—form\"bytheexaggerationofthefastigiatehabitofitsbranches。\"（21/61。Verlot’DesVarietes’1865page94。）Mr。Shirreff，whohasbeenhighlysuccessfulinraisingnewkindsofwheat，remarks，\"Agoodvarietymaysafelyberegardedastheforerunnerofabetterone。\"（21/62。Mr。
PatrickShirreff’Gardener’sChronicle’1858page771。）Agreatrose—grower，Mr。Rivers，hasmadethesameremarkwithrespecttoroses。Sageret（21/63。
’PomologiePhysiolog。’1830page106。），whohadlargeexperience，inspeakingofthefutureprogressoffruit—trees，observesthatthemostimportantprincipleis\"thatthemoreplantshavedepartedfromtheiroriginaltype，themoretheytendtodepartfromit。\"Thereisapparentlymuchtruthinthisremark;forwecaninnootherwayunderstandthesurprisingamountofdifferencebetweenvarietiesinthepartsorqualitieswhicharevalued，whilstotherpartsretainnearlytheiroriginalcharacter。
Theforegoingdiscussionnaturallyleadstothequestion，whatisthelimittothepossibleamountofvariationinanypartorquality，and，consequently，isthereanylimittowhatselectioncaneffect？WillaracehorseeverberearedfleeterthanEclipse？Canourprize—cattleandsheepbestillfurtherimproved？Willagooseberryeverweighmorethanthatproducedby\"London\"in1852？Willthebeet—rootinFranceyieldagreaterpercentageofsugar？Willfuturevarietiesofwheatandothergrainproduceheaviercropsthanourpresentvarieties？Thesequestionscannotbepositivelyanswered;butitiscertainthatweoughttobecautiousinansweringthembyanegative。Insomelinesofvariationthelimithasprobablybeenreached。Youattbelievesthatthereductionofboneinsomeofoursheephasalreadybeencarriedsofarthatitentailsgreatdelicacyofconstitution。（21/64。Youatton’Sheep’page521。）Butseeingthegreatimprovementwithinrecenttimesinourcattleandsheep，andespeciallyinourpigs;seeingthewonderfulincreaseinweightinourpoultryofallkindsduringthelastfewyears;hewouldbeaboldmanwhowouldassertthatperfectionhasbeenreached。IthasoftenbeensaidthatEclipseneverwas，andneverwillbe，beateninspeedbyanyotherhorse;butonmakinginquiriesIfindthatthebestjudgesbelievethatourpresentracehorsesarefleeter。（21/65。SeealsoStonehenge’BritishRuralSports’
editionof1871page384。）Theattempttoraiseanewvarietyofwheatmoreproductivethanthemanyoldkinds，mighthavebeenthoughtuntillatelyquitehopeless;butthishasbeeneffectedbyMajorHallett，bycarefulselection。
Withrespecttoalmostallouranimalsandplants，thosewhoarebestqualifiedtojudgedonotbelievethattheextremepointofperfectionhasyetbeenreachedeveninthecharacterswhichhavealreadybeencarriedtoahighstandard。Forinstance，theshort—facedtumbler—pigeonhasbeengreatlymodified;nevertheless，accordingtoMr。Eaton（21/66。’ATreatiseontheAlmondTumbler’page1。）\"thefieldisstillasopenforfreshcompetitorsasitwasonehundredyearsago。\"Overandoveragainithasbeensaidthatperfectionhadbeenattainedwithourflowers，butahigherstandardhassoonbeenreached。Hardlyanyfruithasbeenmoreimprovedthanthestrawberry，yetagreatauthorityremarks（21/67。M。J。deJonghein’Gardener’sChronicle’
1858page173。），\"itmustnotbeconcealedthatwearefarfromtheextremelimitsatwhichwemayarrive。\"
Nodoubtthereisalimitbeyondwhichtheorganisationcannotbemodifiedcompatiblywithhealthorlife。Theextremedegreeoffleetness，forinstance，ofwhichaterrestrialanimaliscapable，mayhavebeenacquiredbyourpresentracehorses;butasMr。Wallacehaswellremarked（21/68。
’ContributionstotheTheoryofNaturalSelection’2ndedition1871page292。），thequestionthatinterestsus，\"isnotwhetherindefiniteandunlimitedchangeinanyoralldirectionsispossible，butwhethersuchdifferencesasdooccurinnaturecouldhavebeenproducedbytheaccumulationofvarietiesbyselection。\"Andinthecaseofourdomesticproductions，therecanbenodoubtthatmanypartsoftheorganisation，towhichmanhasattended，havebeenthusmodifiedtoagreaterdegreethanthecorrespondingpartsinthenaturalspeciesofthesamegeneraorevenfamilies。Weseethisintheformandsizeofourlightandheavydogsorhorses，——inthebeakandmanyothercharactersofourpigeons，——inthesizeandqualityofmanyfruits，——incomparisonwiththespeciesbelongingtothesamenaturalgroups。
Timeisanimportantelementintheformationofourdomesticraces，asitpermitsinnumerableindividualstobeborn，andthesewhenexposedtodiversifiedconditionsarerenderedvariable。Methodicalselectionhasbeenoccasionallypractisedfromanancientperiodtothepresentday，evenbysemi—civilisedpeople，andduringformertimeswillhaveproducedsomeeffect。
Unconsciousselectionwillhavebeenstillmoreeffective;forduringalengthenedperiodthemorevaluableindividualanimalswilloccasionallyhavebeensaved，andthelessvaluableneglected。Inthecourseoftime，differentvarieties，especiallyinthelesscivilisedcountries，willalsohavebeenmoreorlessmodifiedthroughnaturalselection。Itisgenerallybelieved，thoughonthisheadwehavelittleornoevidence，thatnewcharactersintimebecomefixed;andafterhavinglongremainedfixeditseemspossiblethatundernewconditionstheymightagainberenderedvariable。
Howgreatthelapseoftimehasbeensincemanfirstdomesticatedanimalsandcultivatedplants，webegindimlytosee。Whenthelake—dwellingsofSwitzerlandwereinhabitedduringtheNeolithicperiod，severalanimalswerealreadydomesticatedandvariousplantscultivated。Thescienceoflanguagetellsusthattheartofploughingandsowingthelandwasfollowed，andthechiefanimalshadbeenalreadydomesticated，atanepochsoimmenselyremote，thattheSanskrit，Greek，Latin，Gothic，Celtic，andSclavoniclanguageshadnotasyetdivergedfromtheircommonparent—tongue。（21/69。MaxMuller’ScienceofLanguage’1861page223。）
Itisscarcelypossibletooverratetheeffectsofselectionoccasionallycarriedoninvariouswaysandplacesduringthousandsofgenerations。Allthatweknow，and，inastillstrongerdegree，allthatwedonotknow（21/70。
’YouattonCattle’pages116，128。），ofthehistoryofthegreatmajorityofourbreeds，evenofourmoremodernbreeds，agreeswiththeviewthattheirproduction，throughtheactionofunconsciousandmethodicalselection，hasbeenalmostinsensiblyslow。Whenamanattendsrathermorecloselythanisusualtothebreedingofhisanimals，heisalmostsuretoimprovethemtoaslightextent。Theyareinconsequencevaluedinhisimmediateneighbourhood，andarebredbyothers;andtheircharacteristicfeatures，whateverthesemaybe，willthenslowlybutsteadilybeincreased，sometimesbymethodicalandalmostalwaysbyunconsciousselection。Atlastastrain，deservingtobecalledasub—variety，becomesalittlemorewidelyknown，receivesalocalname，andspreads。Thespreadingwillhavebeenextremelyslowduringancientandlesscivilisedtimes，butnowisrapid。Bythetimethatthenewbreedhadassumedasomewhatdistinctcharacter，itshistory，hardlynoticedatthetime，willhavebeencompletelyforgotten;for，asLowremarks（21/71。
’DomesticatedAnimals’page188。），\"weknowhowquicklythememoryofsucheventsiseffaced。\"
Assoonasanewbreedisthusformed，itisliablethroughthesameprocesstobreakupintonewstrainsandsub—varieties。Fordifferentvarietiesaresuitedfor，andarevaluedunder，differentcircumstances。Fashionchanges，but，shouldafashionlastforevenamoderatelengthoftime，sostrongistheprincipleofinheritance，thatsomeeffectwillprobablybeimpressedonthebreed。Thusvarietiesgoonincreasinginnumber，andhistoryshowsushowwonderfullytheyhaveincreasedsincetheearliestrecords。（21/72。Volz’BeitragezurKulturgeschichte’1852s。99etpassim。）Aseachnewvarietyisproduced，theearlier，intermediate，andlessvaluableformswillbeneglected，andperish。Whenabreed，fromnotbeingvalued，iskeptinsmallnumbers，itsextinctionalmostinevitablyfollowssoonerorlater，eitherfromaccidentalcausesofdestructionorfromcloseinterbreeding;andthisisaneventwhich，inthecaseofwell—markedbreeds，excitesattention。Thebirthorproductionofanewdomesticraceissoslowaprocessthatitescapesnotice;itsdeathordestructioniscomparativelysudden，isoftenrecorded，andwhentoolatesometimesregretted。
Severalauthorshavedrawnawidedistinctionbetweenartificialandnaturalraces。Thelatteraremoreuniformincharacter，possessinginahighdegreetheappearanceofnaturalspecies，andareofancientorigin。Theyaregenerallyfoundinlesscivilisedcountries，andhaveprobablybeenlargelymodifiedbynaturalselection，andonlytoasmallextentbyman’sunconsciousandmethodicalselection。Theyhave，also，duringalongperiod，beendirectlyactedonbythephysicalconditionsofthecountrieswhichtheyinhabit。Theso—calledartificialraces，ontheotherhand，arenotsouniformincharacter;somehaveasemi—monstrouscharacter，suchas\"thewry—leggedterrierssousefulinrabbit—shooting\"（21/73。Blaine’Encyclop。ofRuralSports’page213。），turnspitdogs，anconsheep，niataoxen，Polishfowls，fantail—pigeons，etc。;theircharacteristicfeatureshavegenerallybeenacquiredsuddenly，thoughsubsequentlyincreasedbycarefulselectionsinmanycases。Otherraces，whichcertainlymustbecalledartificial，fortheyhavebeenlargelymodifiedbymethodicalselectionandbycrossing，astheEnglishracehorse，terrier—dogs，theEnglishgame—cock，Antwerpcarrier—pigeons，etc。，neverthelesscannotbesaidtohaveanunnaturalappearance;andnodistinctline，asitseemstome，canbedrawnbetweennaturalandartificialraces。
Itisnotsurprisingthatdomesticracesshouldgenerallypresentadifferentaspectfromnaturalspecies。Manselectsandpropagatesmodificationssolelyforhisownuseorfancy，andnotforthecreature’sowngood。Hisattentionisstruckbystronglymarkedmodifications，whichhaveappearedsuddenly，duetosomegreatdisturbingcauseintheorganisation。Heattendsalmostexclusivelytoexternalcharacters;andwhenhesucceedsinmodifyinginternalorgans，——whenforinstancehereducesthebonesandoffal，orloadstheviscerawithfat，orgivesearlymaturity，etc。—thechancesarestrongthathewillatthesametimeweakentheconstitution。Ontheotherhand，whenananimalhastostrugglethroughoutitslifewithmanycompetitorsandenemies，undercircumstancesinconceivablycomplexandliabletochange，modificationsofthemostvariednatureintheinternalorgansaswellasinexternalcharacters，inthefunctionsandmutualrelationsofparts，willberigorouslytested，preserved，orrejected。Naturalselectionoftenchecksman’scomparativelyfeebleandcapriciousattemptsatimprovement;andifitwerenotso，theresultofhiswork，andofnature’swork，wouldbeevenstillmoredifferent。Nevertheless，wemustnotoverratetheamountofdifferencebetweennaturalspeciesanddomesticraces;themostexperiencednaturalistshaveoftendisputedwhetherthelatteraredescendedfromoneorfromseveralaboriginalstocks，andthisclearlyshowsthatthereisnopalpabledifferencebetweenspeciesandraces。
Domesticracespropagatetheirkindfarmoretruly，andendureformunchlongerperiods，thanmostnaturalistsarewillingtoadmit。Breedersfeelnodoubtonthishead:askamanwhohaslongrearedShorthornorHerefordcattle，LeicesterorSouthdownsheep，SpanishorGamepoultry，tumblerorcarrier—pigeons，whethertheseracesmaynothavebeenderivedfromcommonprogenitors，andhewillprobablylaughyoutoscorn。Thebreederadmitsthathemayhopetoproducesheepwithfinerorlongerwoolandwithbettercarcases，orhandsomerfowls，orcarrier—pigeonswithbeaksjustperceptiblylongertothepractisedeye，andthusbesuccessfulatanexhibition。Thusfarhewillgo，butnofarther。Hedoesnotreflectonwhatfollowsfromaddingupduringalongcourseoftimemanyslight，successivemodifications;nordoeshereflectontheformerexistenceofnumerousvarieties，connectingthelinksineachdivergentlineofdescent。Heconcludes，aswasshownintheearlierchapters，thatallthechiefbreedstowhichhehaslongattendedareaboriginalproductions。Thesystematicnaturalist，ontheotherhand，whogenerallyknowsnothingoftheartofbreeding，whodoesnotpretendtoknowhowandwhentheseveraldomesticraceswereformed，whocannothaveseentheintermediategradations，fortheydonotnowexist，neverthelessfeelsnodoubtthattheseracesaresprungfromasinglesource。Butaskhimwhetherthecloselyalliednaturalspecieswhichhehasstudiedmaynothavedescendedfromacommonprogenitor，andheinhisturnwillperhapsrejectthenotionwithscorn。Thusthenaturalistandbreedermaymutuallylearnausefullessonfromeachother。
SUMMARYONSELECTIONBYMAN。
Therecanbenodoubtthatmethodicalselectionhaseffectedandwilleffectwonderfulresults。Itwasoccasionallypractisedinancienttimes，andisstillpractisedbysemi—civilisedpeople。Charactersofthehighestimportance，andothersoftriflingvalue，havebeenattendedto，andmodified。
Ineednothererepeatwhathasbeensooftensaidonthepartwhichunconsciousselectionhasplayed:weseeitspowerinthedifferencebetweenflockswhichhavebeenseparatelybred，andintheslowchanges，ascircumstanceshaveslowlychanged，whichmanyanimalshaveundergoneinthesamecountry，orwhentransportedintoaforeignland。Weseethecombinedeffectsofmethodicalandunconsciousselection，inthegreatamountofdifferenceinthosepartsorqualitieswhicharevaluedbymanincomparisonwiththepartswhicharenotvalued，andconsequentlyhavenotbeenattendedto。Naturalselectionoftendeterminesman’spowerofselection。Wesometimeserrinimaginingthatcharacters，whichareconsideredasunimportantbythesystematicnaturalist，couldnotbeaffectedbythestruggleforexistence，andcouldnotbeactedonbynaturalselection;butstrikingcaseshavebeengiven，showinghowgreatanerrorthisis。
Thepossibilityofselectioncomingintoactionrestsonvariability;andthisismainlycaused，asweshallhereaftersee，bychangesintheconditionsoflife。Selectionissometimesrendereddifficult，orevenimpossible，bytheconditionsbeingopposedtothedesiredcharacterorquality。Itissometimescheckedbythelessenedfertilityandweakenedconstitutionwhichfollowfromlong—continuedcloseinterbreeding。Thatmethodicalselectionmaybesuccessful，theclosestattentionanddiscernment，combinedwithunweariedpatience，areabsolutelynecessary;andthesesamequalities，thoughnotindispensable，arehighlyserviceableinthecaseofunconsciousselection。Itisalmostnecessarythatalargenumberofindividualsshouldbereared;forthustherewillbeafairchanceofvariationsofthedesirednaturearising，andofeveryindividualwiththeslightestblemishorinanydegreeinferiorbeingfreelyrejected。Hencelengthoftimeisanimportantelementofsuccess。Thus，also，reproductionatanearlyageandatshortintervalsfavoursthework。Facilityinpairinganimals，ortheirinhabitingaconfinedarea，isadvantageousasachecktofreecrossing。Wheneverandwhereverselectionisnotpractised，distinctracesarenotformedwithinthesamecountry。Whenanyonepartofthebodyoronequalityisnotattendedto，itremainseitherunchangedorvariesinafluctuatingmanner，whilstatthesametimeotherpartsandotherqualitiesmaybecomepermanentlyandgreatlymodified。Butfromthetendencytoreversionandtocontinuedvariability，thosepartsororganswhicharenowundergoingrapidimprovementthroughselection，arelikewisefoundtovarymuch。Consequentlyhighly—bredanimalswhenneglectedsoondegenerate;butwehavenoreasontobelievethattheeffectsoflong—continuedselectionwould，iftheconditionsofliferemainedthesame，besoonandcompletelylost。
Manalwaystendstogotoanextremepointintheselection，whethermethodicalorunconscious，ofallusefulandpleasingqualities。Thisisanimportantprinciple，asitleadstocontinueddivergence，andinsomerarecasestoconvergenceofcharacter。Thepossibilityofcontinueddivergencerestsonthetendencyineachpartororgantogoonvaryinginthesamemannerinwhichithasalreadyvaried;andthatthisoccurs，isprovedbythesteadyandgradualimprovementofmanyanimalsandplantsduringlengthenedperiods。Theprincipleofdivergenceofcharacter，combinedwiththeneglectandfinalextinctionofallprevious，less—valued，andintermediatevarieties，explainstheamountofdifferenceandthedistinctnessofourseveralraces。
Althoughwemayhavereachedtheutmostlimittowhichcertaincharacterscanbemodified，yetwearefarfromhavingreached，aswehavegoodreasontobelieve，thelimitinthemajorityofcases。Finally，fromthedifferencebetweenselectionascarriedonbymanandbynature，wecanunderstandhowitisthatdomesticracesoften，thoughbynomeansalways，differingeneralaspectfromcloselyalliednaturalspecies。
ThroughoutthischapterandelsewhereIhavespokenofselectionastheparamountpower，yetitsactionabsolutelydependsonwhatweinourignorancecallspontaneousoraccidentalvariability。Letanarchitectbecompelledtobuildanedificewithuncutstones，fallenfromaprecipice。Theshapeofeachfragmentmaybecalledaccidental;yettheshapeofeachhasbeendeterminedbytheforceofgravity，thenatureoftherock，andtheslopeoftheprecipice，——eventsandcircumstances，allofwhichdependonnaturallaws;butthereisnorelationbetweentheselawsandthepurposeforwhicheachfragmentisusedbythebuilder。Inthesamemannerthevariationsofeachcreaturearedeterminedbyfixedandimmutablelaws;butthesebearnorelationtothelivingstructurewhichisslowlybuiltupthroughthepowerofselection，whetherthisbenaturalorartificialselection。
Ifourarchitectsucceededinrearinganobleedifice，usingtheroughwedge—
shapedfragmentsforthearches，thelongerstonesforthelintels，andsoforth，weshouldadmirehisskilleveninahigherdegreethanifhehadusedstonesshapedforthepurpose。Soitiswithselection，whetherappliedbymanorbynature;foralthoughvariabilityisindispensablynecessary，yet，whenwelookatsomehighlycomplexandexcellentlyadaptedorganism，variabilitysinkstoaquitesubordinatepositioninimportanceincomparisonwithselection，inthesamemannerastheshapeofeachfragmentusedbyoursupposedarchitectisunimportantincomparisonwithhisskill。
CHAPTER2。XXII。
CAUSESOFVARIABILITY。
VARIABILITYDOESNOTNECESSARILYACCOMPANYREPRODUCTION。
CAUSESASSIGNEDBYVARIOUSAUTHORS。
INDIVIDUALDIFFERENCES。
VARIABILITYOFEVERYKINDDUETOCHANGEDCONDITIONSOFLIFE。
ONTHENATUREOFSUCHCHANGES。
CLIMATE，FOOD，EXCESSOFNUTRIMENT。
SLIGHTCHANGESSUFFICIENT。
EFFECTSOFGRAFTINGONTHEVARIABILITYOFSEEDLING—TREES。
DOMESTICPRODUCTIONSBECOMEHABITUATEDTOCHANGEDCONDITIONS。
ONTHEACCUMULATIVEACTIONOFCHANGEDCONDITIONS。
CLOSEINTERBREEDINGANDTHEIMAGINATIONOFTHEMOTHERSUPPOSEDTOCAUSE
VARIABILITY。
CROSSINGASACAUSEOFTHEAPPEARANCEOFNEWCHARACTERS。
VARIABILITYFROMTHECOMMINGLINGOFCHARACTERSANDFROMREVERSION。
ONTHEMANNERANDPERIODOFACTIONOFTHECAUSESWHICHEITHERDIRECTLY，OR
INDIRECTLYTHROUGHTHEREPRODUCTIVESYSTEM，INDUCEVARIABILITY。
Wewillnowconsider，asfaraswecan，thecausesofthealmostuniversalvariabilityofourdomesticatedproductions。Thesubjectisanobscureone;
butitmaybeusefultoprobeourignorance。Someauthors，forinstanceDr。
ProsperLucas，lookatvariabilityasanecessarycontingentonreproduction，andasmuchanaboriginallawasgrowthorinheritance。Othershaveoflateencouraged，perhapsunintentionally，thisviewbyspeakingofinheritanceandvariabilityasequalandantagonisticprinciples。Pallasmaintained，andhehashadsomefollowers，thatvariabilitydependsexclusivelyonthecrossingofprimordiallydistinctforms。Otherauthorsattributevariabilitytoanexcessoffood，andwithanimalstoanexcessrelativelytotheamountofexercisetaken，oragaintotheeffectsofamoregenialclimate。Thatthesecausesarealleffectiveishighlyprobable。Butwemust，Ithink，takeabroaderview，andconcludethatorganicbeings，whensubjectedduringseveralgenerationstoanychangewhateverintheirconditions，tendtovary;thekindofvariationwhichensuesdependinginmostcasesinafarhigherdegreeonthenatureorconstitutionofthebeing，thanonthenatureofthechangedconditions。
Thoseauthorswhobelievethatitisalawofnaturethateachindividualshoulddifferinsomeslightdegreefromeveryother，maymaintain，apparentlywithtruth，thatthisisthefact，notonlywithalldomesticatedanimalsandcultivatedplants，butlikewisewithallorganicbeingsinastateofnature。
TheLaplanderbylongpracticeknowsandgivesanametoeachreindeer，though，asLinnaeusremarks，\"todistinguishonefromanotheramongsuchmultitudeswasbeyondmycomprehension，fortheywerelikeantsonananthill。\"InGermanyshepherdshavewonwagersbyrecognisingeachsheepinaflockofahundred，whichtheyhadneverseenuntilthepreviousfortnight。
Thispowerofdiscrimination，however，isasnothingcomparedtothatwhichsomefloristshaveacquired。Verlotmentionsagardenerwhocoulddistinguish150kindsofcamellia，whennotinflower;andithasbeenpositivelyassertedthatthefamousoldDutchfloristVoorhelm，whokeptabove1200varietiesofthehyacinth，washardlyeverdeceivedinknowingeachvarietybythebulbalone。Hencewemustconcludethatthebulbsofthehyacinthandthebranchesandleavesofthecamellia，thoughappearingtoanunpractisedeyeabsolutelyundistinguishable，yetreallydiffer。（22/1。’DesJacinthes’etc。Amsterdam1768page43;Verlot’DesVarietes’etc。page86。OnthereindeerseeLinnaeus’TourinLapland’translatedbySirJ。E。Smithvolume1page314。ThestatementinregardtoGermanshepherdsisgivenontheauthorityofDr。
Weinland。）
AsLinnaeushascomparedthereindeerinnumbertoants，Imayaddthateachantknowsitsfellowofthesamecommunity。SeveraltimesIcarriedantsofthesamespecies（Formicarufa）fromoneant—hilltoanother，inhabitedapparentlybytensofthousandsofants;butthestrangerswereinstantlydetectedandkilled。Ithenputsomeantstakenfromaverylargenestintoabottlestronglyperfumedwithassafoetida，andafteranintervaloftwenty—
fourhoursreturnedthemtotheirhome;theywereatfirstthreatenedbytheirfellows，butweresoonrecognisedandallowedtopass。Henceeachantcertainlyrecognised，independentlyofodour，itsfellow;andifalltheantsofthesamecommunityhavenotsomecountersignorwatchword，theymustpresenttoeachother’ssensessomedistinguishablecharacter。
Thedissimilarityofbrothersorsistersofthesamefamily，andofseedlingsfromthesamecapsule，maybeinpartaccountedforbytheunequalblendingofthecharactersofthetwoparents，andbythemoreorlesscompleterecoverythroughreversionofancestralcharactersoneitherside;butwethusonlypushthedifficultyfurtherbackintime，forwhatmadetheparentsortheirprogenitorsdifferent？Hencethebelief（22/2。Muller’Physiology’Englishtranslation，volume2page1662。Withrespecttothesimilarityoftwinsinconstitution，Dr。WilliamOglehasgivenmethefollowingextractfromProfessorTrousseau’sLectures’CliniqueMedicale’tome1page523，inwhichacuriouscaseisrecorded:——\"J’aidonnemessoinsadeuxfreresjumeaux，tousdeuxsiextraordinairementressemblantsqu’ilm’etaitimpossibledelesreconnaitre，amoindelesvoirl’unacotedel’autre。Cetteressemblancephysiques’etendaitplusloin:ilsavaient，permettez—moil’expression，unesimilitudepathologiqueplusremarquableencore。Ainsil’und’euxquejevoyaisauxneothermesaParismaladed’uneophthalmierhumatismalemedisait，’Encemomentmonfreredoitavoiruneophthalmiecommelamienne;’etcommejem’etaisrecrie，ilmemontraitquelquesjoursapresunelettrequ’ilvenaitderecevoirdecefrerealorsaVienne，etquiluiecrivaiteneffet——’J’aimonophthalmie，tudoisavoirlatienne。’Quelquesingulierquececipuisseparaitre，lefaitn’enestpasmoinsexact:onnemel’apasraconte，jel’aivu，etj’enaivud’autresanaloguesdansmapratique。Cesdeuxjumeauxetaientaussitousdeuxasthmatiques，etasthmatiquesauneffroyabledegre。
OriginairesdeMarseille，ilsn’ontjamaispudemeurerdanscetteville，ouleursinteretslesappelaientsouvent，sansetreprisdeleursacces;jamaisilsn’eneprouvaientaParis。Bienmieux，illeursuffisaitdegagnerToulonpouretreguerisdeleursattaquesdeMarseille。Voyageantsanscesseetdanstouspayspourleursaffaires，ilsavaientremarquequecertaineslocalitesleuretaientfunestes，quedansd’autresilsetaientexemptsdetoutphenomened’oppression。\"）thataninnatetendencytovaryexists，independentlyofexternaldifferences，seemsatfirstsightprobable。Buteventheseedsnurturedinthesamecapsulearenotsubjectedtoabsolutelyuniformconditions，astheydrawtheirnourishmentfromdifferentpoints;andweshallseeinafuturechapterthatthisdifferencesometimessufficestoaffectthecharacterofthefutureplant。Thegreaterdissimilarityofthesuccessivechildrenofthesamefamilyincomparisonwithtwins，whichoftenresembleeachotherinexternalappearance，mentaldisposition，andconstitution，insoextraordinaryamanner，apparentlyprovesthatthestateoftheparentsattheexactperiodofconception，orthenatureofthesubsequentembryonicdevelopment，hasadirectandpowerfulinfluenceonthecharacteroftheoffspring。Nevertheless，whenwereflectontheindividualdifferencesbetweenorganicbeingsinastateofnature，asshownbyeverywildanimalknowingitsmate;andwhenwereflectontheinfinitediversityofthemanyvarietiesofourdomesticatedproductions，wemaywellbeinclinedtoexclaim，thoughfalselyasIbelieve，thatVariabilitymustbelookedatasanultimatefact，necessarilycontingentonreproduction。
Thoseauthorswhoadoptthislatterviewwouldprobablydenythateachseparatevariationhasitsownproperexcitingcause。Althoughwecanseldomtracethepreciserelationbetweencauseandeffect，yettheconsiderationspresentlytobegivenleadtotheconclusionthateachmodificationmusthaveitsowndistinctcause，andisnottheresultofwhatweblindlycallaccident。ThefollowingstrikingcasehasbeencommunicatedtomebyDr。
WilliamOgle。Twogirls，bornastwins，andinallrespectsextremelyalike，hadtheirlittlefingersonbothhandscrooked;andinbothchildrenthesecondbicuspidtoothoftheseconddentitionontherightsideintheupperjawwasmisplaced;for，insteadofstandinginalinewiththeothers，itgrewfromtheroofofthemouthbehindthefirstbicuspid。Neithertheparentsnoranyothermembersofthefamilywereknowntohaveexhibitedanysimilarpeculiarity;butasonofoneofthesegirlshadthesametoothsimilarlymisplaced。Now，asboththegirlswereaffectedinexactlythesamemanner，theideaofaccidentisatonceexcluded:andwearecompelledtoadmitthattheremusthaveexistedsomepreciseandsufficientcausewhich，ifithadoccurredahundredtimes，wouldhavegivencrookedfingersandmisplacedbicuspidteethtoahundredchildren。Itisofcoursepossiblethatthiscasemayhavebeenduetoreversiontosomelong—forgottenprogenitor，andthiswouldmuchweakenthevalueoftheargument。Ihavebeenledtothinkoftheprobabilityofreversion，fromhavingbeentoldbyMr。Galtonofanothercaseoftwingirlsbornwiththeirlittlefingersslightlycrooked，whichtheyinheritedfromtheirmaternalgrandmother。
Wewillnowconsiderthegeneralarguments，whichappeartometohavegreatweight，infavouroftheviewthatvariationsofallkindsanddegreesaredirectlyorindirectlycausedbytheconditionsoflifetowhicheachbeing，andmoreespeciallyitsancestors，havebeenexposed。
Noonedoubtsthatdomesticatedproductionsaremorevariablethanorganicbeingswhichhaveneverbeenremovedfromtheirnaturalconditions。
Monstrositiesgraduatesoinsensiblyintomerevariationsthatitisimpossibletoseparatethem;andallthosewhohavestudiedmonstrositiesbelievethattheyarefarcommonerwithdomesticatedthanwithwildanimalsandplants（22/3。Isid。GeoffroySt。—Hilaire’Hist。desAnomalies’tome3page352;Moquin—Tandon’TeratologieVegetale’1841page115。）;andinthecaseofplants，monstrositieswouldbeequallynoticeableinthenaturalasinthecultivatedstate。Undernature，theindividualsofthesamespeciesareexposedtonearlyuniformconditions，fortheyarerigorouslykepttotheirproperplacesbyahostofcompetinganimalsandplants;