theyhave，also，longbeenhabituatedtotheirconditionsoflife;butitcannotbesaidthattheyaresubjecttoquiteuniformconditions，andtheyareliabletoacertainamountofvariation。Thecircumstancesunderwhichourdomesticproductionsarerearedarewidelydifferent:theyareprotectedfromcompetition;theyhavenotonlybeenremovedfromtheirnaturalconditionsandoftenfromtheirnativeland，buttheyarefrequentlycarriedfromdistricttodistrict，wheretheyaretreateddifferently，sothattheyrarelyremainduringanyconsiderablelengthoftimeexposedtocloselysimilarconditions。Inconformitywiththis，allourdomesticatedproductions，withtherarestexceptions，varyfarmorethannaturalspecies。Thehive—bee，whichfeedsitselfandfollowsinmostrespectsitsnaturalhabitsoflife，istheleastvariableofalldomesticatedanimals，andprobablythegooseisthenextleastvariable;buteventhegoosevariesmorethanalmostanywildbird，sothatitcannotbeaffiliatedwithperfectcertaintytoanynaturalspecies。Hardlyasingleplantcanbenamed，whichhaslongbeencultivatedandpropagatedbyseed，thatisnothighlyvariable;commonrye（Secalecereale）hasaffordedfewerandlessmarkedvarietiesthanalmostanyothercultivatedplant（22/4。
Metzger’DieGetreidarten’1841s。39。）;butitmaybedoubtedwhetherthevariationsofthis，theleastvaluableofallourcereals，havebeencloselyobserved。
Bud—variation，whichwasfullydiscussedinaformerchapter，showsusthatvariabilitymaybequiteindependentofseminalreproduction，andlikewiseofreversiontolong—lostancestralcharacters。Noonewillmaintainthatthesuddenappearanceofamoss—roseonaProvence—roseisareturntoaformerstate，formossinessofthecalyxhasbeenobservedinnonaturalspecies;thesameargumentisapplicabletovariegatedandlaciniatedleaves;norcantheappearanceofnectarinesonpeach—treesbeaccountedforontheprincipleofreversion。Butbud—variationsmoreimmediatelyconcernus，astheyoccurfarmorefrequentlyonplantswhichhavebeenhighlycultivatedduringalengthoftime，thanonotherandlesshighlycultivatedplants;andveryfewwell—
markedinstanceshavebeenobservedwithplantsgrowingunderstrictlynaturalconditions。Ihavegivenoneinstanceofanash—treegrowinginagentleman’spleasure—grounds;andoccasionallytheremaybeseen，onbeechandothertrees，twigsleafingatadifferentperiodfromtheotherbranches。ButourforesttreesinEnglandcanhardlybeconsideredaslivingunderstrictlynaturalconditions;theseedlingsareraisedandprotectedinnursery—grounds，andmustoftenbetransplantedintoplaceswherewildtreesofthekindwouldnotnaturallygrow。Itwouldbeesteemedaprodigyifadog—rosegrowinginahedgeproducedbybud—variationamoss—rose，orawildbullaceorwildcherry—
treeyieldedabranchbearingfruitofadifferentshapeandcolourfromtheordinaryfruit。Theprodigywouldbeenhancedifthesevaryingbrancheswerefoundcapableofpropagation，notonlybygrafts，butsometimesbyseed;yetanalogouscaseshaveoccurredwithmanyofourhighlycultivatedtreesandherbs。
Theseseveralconsiderationsalonerenderitprobablethatvariabilityofeverykindisdirectlyorindirectlycausedbychangedconditionsoflife。Or，toputthecaseunderanotherpointofview，ifitwerepossibletoexposealltheindividualsofaspeciesduringmanygenerationstoabsolutelyuniformconditionsoflife，therewouldbenovariability。
ONTHENATUREOFTHECHANGESINTHECONDITIONSOFLIFEWHICHINDUCE
VARIABILITY。
Fromaremoteperiodtothepresentday，underclimatesandcircumstancesasdifferentasitispossibletoconceive，organicbeingsofallkinds，whendomesticatedorcultivated，havevaried。Weseethiswiththemanydomesticracesofquadrupedsandbirdsbelongingtodifferentorders，withgoldfishandsilkworms，withplantsofmanykinds，raisedinvariousquartersoftheworld。
InthedesertsofnorthernAfricathedate—palmhasyieldedthirty—eightvarieties;inthefertileplainsofIndiaitisnotorioushowmanyvarietiesofriceandofahostofotherplantsexist;inasinglePolynesianisland，twenty—fourvarietiesofthebread—fruit，thesamenumberofthebanana，andtwenty—twovarietiesofthearum，arecultivatedbythenatives;themulberry—
treeinIndiaandEuropehasyieldedmanyvarietiesservingasfoodforthesilkworm;andinChinasixty—threevarietiesofthebambooareusedforvariousdomesticpurposes。（22/5。Onthedate—palmseeVogel’AnnalsandMag。
ofNat。Hist。’1854page460。OnIndianvarietiesDr。F。Hamilton’Transact。
Linn。Soc。’volume14page296。OnthevarietiescultivatedinTahitiseeDr。
BennettinLoudon’s’Mag。ofN。Hist。’volume51832page484。AlsoEllis’PolynesianResearches’volume1pages370，375。OntwentyvarietiesofthePandanusandothertreesintheMarianneIslandsee’Hooker’sMiscellany’
volume1page308。OnthebambooinChinaseeHuc’ChineseEmpire’volume2
page307。）Thesefacts，andinnumerableotherswhichcouldbeadded，indicatethatachangeofalmostanykindintheconditionsoflifesufficestocausevariability——differentchangesactingondifferentorganisms。
AndrewKnight（22/6。’TreatiseontheCultureoftheApple’etc。page3。）
attributedthevariationofbothanimalsandplantstoamoreabundantsupplyofnourishment，ortoamorefavourableclimate，thanthatnaturaltothespecies。Amoregenialclimate，however，isfarfromnecessary;thekidney—
bean，whichisofteninjuredbyourspringfrosts，andpeaches，whichrequiretheprotectionofawall，havevariedmuchinEngland，ashastheorange—treeinnorthernItaly，whereitisbarelyabletoexist。（22/7。Gallesio’TeoriadellaRiproduzioneVeg。’page125。）Norcanweoverlookthefact，thoughnotimmediatelyconnectedwithourpresentsubject，thattheplantsandshellsoftheArcticregionsareeminentlyvariable。（22/8。SeeDr。Hooker’sMemoironArcticPlantsin’Linn。Transact。’volume23part2。Mr。Woodward，andahigherauthoritycannotbequoted，speaksoftheArcticmolluscainhis’RudimentaryTreatise’1856page355asremarkablysubjecttovariation。）
Moreover，itdoesnotappearthatachangeofclimate，whethermoreorlessgenial，isoneofthemostpotentcausesofvariability;forinregardtoplantsAlph。DeCandolle，inhis’GeographieBotanique’repeatedlyshowsthatthenativecountryofaplant，whereinmostcasesithasbeenlongestcultivated，isthatwhereithasyieldedthegreatestnumberofvarieties。
Itisdoubtfulwhetherachangeinthenatureofthefoodisapotentcauseofvariability。Scarcelyanydomesticatedanimalhasvariedmorethanthepigeonorthefowl，buttheirfood，especiallythatofhighly—bredpigeons，isgenerallythesame。Norcanourcattleandsheephavebeensubjectedtoanygreatchangeinthisrespect。Butinallthesecasesthefoodprobablyismuchlessvariedinkindthanthatwhichwasconsumedbythespeciesinitsnaturalstate。（22/9。Bechsteininhis’NaturgeschichtederStubenvogel’1840s。238，hassomegoodremarksonthissubject。Hestatesthathiscanary—birdsvariedincolour，thoughkeptonuniformfood。）
Ofallthecauseswhichinducevariability，excessoffood，whetherornotchangedinnature，isprobablythemostpowerful。ThisviewwasheldwithregardtoplantsbyAndrewKnight，andisnowheldbySchleiden，moreespeciallyinreferencetotheinorganicelementsofthefood。（22/10。’ThePlant’bySchleidentranslatedbyHenfrey1848page169。SeealsoAlex。Braunin’Bot。Memoirs’RaySoc。1853page313。）Inordertogiveaplantmorefooditsufficesinmostcasestogrowitseparately，andthuspreventotherplantsrobbingitsroots。Itissurprising，asIhaveoftenseen，howvigorouslyourcommonwildspeciesflourishwhenplantedbythemselves，thoughnotinhighlymanuredland;separategrowthis，infact，thefirststepincultivation。Weseetheconverseofthebeliefthatexcessoffoodinducesvariabilityinthefollowingstatementbyagreatraiserofseedsofallkinds（22/11。Messrs。
HardyandSonofMaldonin’Gardener’sChronicle’1856page458。Carriere’ProductionetFixationdesVarietes’1865page31。）:\"Itisaruleinvariablywithus，whenwedesiretokeepatruestockofanyonekindofseed，togrowitonpoorlandwithoutdung;butwhenwegrowforquantity，weactcontrary，andsometimeshavedearlytorepentofit。\"AccordingalsotoCarriere，whohashadgreatexperiencewithflower—gardenseeds，\"Onremarqueengenerallesplantesdevigeurmoyennesontcellesquiconserventlemieuxleurscaracteres。\"
Inthecaseofanimalsthewantofaproperamountofexercise，asBechsteinremarked，hasperhapsplayed，independentlyofthedirecteffectsofthedisuseofanyparticularorgan，animportantpartincausingvariability。Wecanseeinavaguemannerthat，whentheorganisedandnutrientfluidsofthebodyarenotusedduringgrowth，orbythewearandtearofthetissues，theywillbeinexcess;andasgrowth，nutrition，andreproductionareintimatelyalliedprocesses，thissuperfluitymightdisturbthedueandproperactionofthereproductiveorgans，andconsequentlyaffectthecharacterofthefutureoffspring。Butitmaybearguedthatneitheranexcessoffoodnorasuperfluityintheorganisedfluidsofthebodynecessarilyinducesvariability。Thegooseandtheturkeyhavebeenwellfedformanygenerations，yethavevariedverylittle。Ourfruit—treesandculinaryplants，whicharesovariable，havebeencultivatedfromanancientperiod，and，thoughtheyprobablystillreceivemorenutrimentthanintheirnaturalstate，yettheymusthavereceivedduringmanygenerationsnearlythesameamount;anditmightbethoughtthattheywouldhavebecomehabituatedtotheexcess。
Nevertheless，onthewhole，Knight’sview，thatexcessoffoodisoneofthemostpotentcausesofvariability，appears，asfarasIcanjudge，probable。
Whetherornotourvariouscultivatedplantshavereceivednutrimentinexcess，allhavebeenexposedtochangesofvariouskinds。Fruit—treesaregraftedondifferentstocks，andgrowninvarioussoils。Theseedsofculinaryandagriculturalplantsarecarriedfromplacetoplace;andduringthelastcenturytherotationofourcropsandthemanuresusedhavebeengreatlychanged。
Slightchangesoftreatmentoftensufficetoinducevariability。Thesimplefactofalmostallourcultivatedplantsanddomesticatedanimalshavingvariedinallplacesandatalltimes，leadstothisconclusion。SeedstakenfromcommonEnglishforest—trees，grownundertheirnativeclimate，nothighlymanuredorotherwiseartificiallytreated，yieldseedlingswhichvarymuch，asmaybeseenineveryextensiveseed—bed。Ihaveshowninaformerchapterwhatanumberofwell—markedandsingularvarietiesthethorn（Crataegusoxycantha）
hasproduced:yetthistreehasbeensubjectedtohardlyanycultivation。InStaffordshireIcarefullyexaminedalargenumberoftwoBritishplants，namelyGeraniumphaeumandpyrenaicum，whichhaveneverbeenhighlycultivated。Theseplantshadspreadspontaneouslybyseedfromacommongardenintoanopenplantation;andtheseedlingsvariedinalmosteverysinglecharacter，bothintheirflowerandfoliage，toadegreewhichIhaveneverseenexceeded;yettheycouldnothavebeenexposedtoanygreatchangeintheirconditions。
Withrespecttoanimals，Azarahasremarkedwithmuchsurprise（22/12。
’QuadrupedesduParaguay’1801tome2page319。）that，whilsttheferalhorsesonthePampasarealwaysofoneofthreecolours，andthecattlealwaysofauniformcolour，yettheseanimals，whenbredontheunenclosedestancias，thoughkeptinastatewhichcanhardlybecalleddomesticated，andapparentlyexposedtoalmostidenticallythesameconditionsaswhentheyareferal，neverthelessdisplayagreatdiversityofcolour。SoagaininIndiaseveralspeciesoffresh—waterfishareonlysofartreatedartificially，thattheyarerearedingreattanks;butthissmallchangeissufficienttoinducemuchvariability。（22/13。M’ClellandonIndianCyprinidae’AsiaticResearches’
volume19part21839pages266，268，313。）
Somefactsontheeffectsofgrafting，inregardtothevariabilityoftrees，deserveattention。Cabanisassertsthatwhencertainpearsaregraftedonthequince，theirseedsyieldagreaternumberofvarietiesthandotheseedsofthesamevarietyofpearwhengraftedonthewildpear。（22/14。QuotedbySageret’Pom。Phys。’1830page43。Thisstatement，however，isnotbelievedbyDecaisne。）Butasthepearandquincearedistinctspecies，thoughsocloselyrelatedthattheonecanbereadilygraftedandsucceedsadmirablyontheother，thefactofvariabilitybeingthuscausedisnotsurprising;aswearehereenabledtoseethecause，namely，theverydifferentnatureofthestockandgraft。SeveralNorthAmericanvarietiesoftheplumandpeacharewellknowntoreproducethemselvestrulybyseed;butDowningasserts（22/15。’TheFruitsofAmerica’1845page5。），\"thatwhenagraftistakenfromoneofthesetreesandplaceduponanotherstock，thisgraftedtreeisfoundtoloseitssingularpropertyofproducingthesamevarietybyseed，andbecomeslikeallotherworkedtrees;\"——thatis，itsseedlingsbecomehighlyvariable。
Anothercaseisworthgiving:theLalandevarietyofthewalnut—treeleafsbetweenApril20thandMay15th，anditsseedlingsinvariablyinheritthesamehabit;whilstseveralothervarietiesofthewalnutleafinJune。Now，ifseedlingsareraisedfromtheMay—leafingLalandevariety，graftedonanotherMay—leafingvariety，thoughbothstockandgrafthavethesameearlyhabitofleafing，yettheseedlingsleafatvarioustimes，evenaslateasthe5thofJune。（22/16。M。Cardanin’ComptesRendus’December1848quotedin’Gardener’sChronicle’1849page101。）Suchfactsasthesearewellfittedtoshowonwhatobscureandslightcausesvariabilitydepends。
[Imayherejustalludetotheappearanceofnewandvaluablevarietiesoffruit—treesandofwheatinwoodsandwasteplaces，whichatfirstsightseemsamostanomalouscircumstance。InFranceaconsiderablenumberofthebestpearshavebeendiscoveredinwoods;andthishasoccurredsofrequently，thatPoiteauassertsthat\"improvedvarietiesofourcultivatedfruitsrarelyoriginatewithnurserymen。\"（22/17。M。AlexisJordanmentionsfourexcellentpearsfoundinwoodsinFrance，andalludestoothers（’Mem。Acad。deLyon’
tome21852page159）。Poiteau’sremarkisquotedin’Gardener’sMag。’volume41828page385。See’Gardener’sChronicle’1862page335，foranothercaseofanewvarietyofthepearfoundinahedgeinFrance。Alsoforanothercase，seeLoudon’s’Encyclop。ofGardening’page901。Mr。Rivershasgivenmesimilarinformation。）InEngland，ontheotherhand，noinstanceofagoodpearhavingbeenfoundwildhasbeenrecorded;andMr。Riversinformsmethatheknowsofonlyoneinstancewithapples，namely，theBessPoole，whichwasdiscoveredinawoodinNottinghamshire。ThisdifferencebetweenthetwocountriesmaybeinpartaccountedforbythemorefavourableclimateofFrance，butchieflyfromthegreatnumberofseedlingswhichspringupthereinthewoods。IinferthatthisisthecasefromaremarkmadebyaFrenchgardener（22/18。Duval’Hist。duPoirier’1849page2。），whoregardsitasanationalcalamitythatsuchanumberofpear—treesareperiodicallycutdownforfirewood，beforetheyhavebornefruit。Thenewvarietieswhichthusspringupinthewoods，thoughtheycannothavereceivedanyexcessofnutriment，willhavebeenexposedtoabruptlychangedconditions，butwhetherthisisthecauseoftheirproductionisverydoubtful。Thesevarieties，however，areprobablyalldescended（22/19。IinferthatthisisthefactfromVanMons’statement（’ArbresFruitiers’1835tome1page446）thathefindsinthewoodsseedlingsresemblingallthechiefcultivatedracesofboththepearandapple。VanMons，however，lookedatthesewildvarietiesasaboriginalspecies。）fromoldcultivatedkindsgrowinginadjoiningorchards——
acircumstancewhichwillaccountfortheirvariability;andoutofavastnumberofvaryingtreestherewillalwaysbeagoodchanceoftheappearanceofavaluablekind。InNorthAmerica，wherefruit—treesfrequentlyspringupinwasteplaces，theWashingtonpearwasfoundinahedge，andtheEmperorpeachinawood。（22/20。Downing’Fruit—treesofNorthAmerica’page422;
Foleyin’Transact。Hort。Soc。’volume6page412。）
Withrespecttowheat，somewritershavespoken（22/21。’Gardener’sChronicle’
1847page244。）asifitwereanordinaryeventfornewvarietiestobefoundinwasteplaces;theFentonwheatwascertainlydiscoveredgrowingonapileofbasalticdetritusinaquarry，butinsuchasituationtheplantwouldprobablyreceiveasufficientamountofnutriment。TheChidhamwheatwasraisedfromanearfoundONahedge;andHunter’swheatwasdiscoveredBYtheroadsideinScotland，butitisnotsaidthatthislattervarietygrewwhereitwasfound。（22/22。’Gardener’sChronicle’1841page383;1850page700;
1854page650。）]
Whetherourdomesticproductionswouldeverbecomesocompletelyhabituatedtotheconditionsunderwhichtheynowlive，astoceasevarying，wehavenosufficientmeansforjudging。But，infact，ourdomesticproductionsareneverexposedforagreatlengthoftimetouniformconditions，anditiscertainthatourmostancientlycultivatedplants，aswellasanimals，stillgoonvarying，forallhaverecentlyundergonemarkedimprovement。Insomefewcases，however，plantshavebecomehabituatedtonewconditions。Thus，Metzger，whocultivatedinGermanyduringmanyyearsnumerousvarietiesofwheat，broughtfromdifferentcountries（22/23。’DieGetreidearten’1843s。
66，116，117。），statesthatsomekindswereatfirstextremelyvariable，butgradually，inoneinstanceafteranintervaloftwenty—fiveyears，becameconstant;anditdoesnotappearthatthisresultedfromtheselectionofthemoreconstantforms。
ONTHEACCUMULATIVEACTIONOFCHANGEDCONDITIONSOFLIFE。
Wehavegoodgroundsforbelievingthattheinfluenceofchangedconditionsaccumulates，sothatnoeffectisproducedonaspeciesuntilithasbeenexposedduringseveralgenerationstocontinuedcultivationordomestication。
Universalexperienceshowsusthatwhennewflowersarefirstintroducedintoourgardenstheydonotvary;butultimatelyall，withtherarestexceptions，varytoagreaterorlessextent。Inafewcasestherequisitenumberofgenerations，aswellasthesuccessivestepsintheprogressofvariation，havebeenrecorded，asintheoftenquotedinstanceoftheDahlia。（22/24。
Sabinein’Hort。Transact。’volume3page225;Bronn’GeschichtederNatur’b。
2s。119。）Afterseveralyears’culturetheZinniahasonlylately（1860）
beguntovaryinanygreatdegree。\"Inthefirstsevenoreightyearsofhighcultivation，theSwanRiverdaisy（Brachycomeiberidifolia）kepttoitsoriginalcolour;itthenvariedintolilacandpurpleandotherminorshades。\"
（22/25。’JournalofHorticulture’1861page112;onZinnia’Gardener’sChronicle’1860page852。）AnalogousfactshavebeenrecordedwiththeScotchrose。Indiscussingthevariabilityofplantsseveralexperiencedhorticulturistshavespokentothesamegeneraleffect。Mr。Salter（22/26。
’TheChrysanthemum，itsHistory，etc。’1865page3。）remarks，\"Everyoneknowsthatthechiefdifficultyisinbreakingthroughtheoriginalformandcolourofthespecies，andeveryonewillbeonthelook—outforanynaturalsport，eitherfromseedorbranch;thatbeingonceobtained，howevertriflingthechangemaybe，theresultdependsuponhimself。\"M。deJonghe，whohashadsomuchsuccessinraisingnewvarietiesofpearsandstrawberries（22/27。
’Gardener’sChronicle’1855page54;’JournalofHorticulture’May9，1865
page363。），remarkswithrespecttotheformer，\"Thereisanotherprinciple，namely，thatthemoreatypehasenteredintoastateofvariation，thegreaterisitstendencytocontinuedoingso;andthemoreithasvariedfromtheoriginaltype，themoreitisdisposedtovarystillfarther。\"Wehave，indeed，alreadydiscussedthislatterpointwhentreatingofthepowerwhichmanpossesses，throughselection，ofcontinuallyaugmentinginthesamedirectioneachmodification;forthispowerdependsoncontinuedvariabilityofthesamegeneralkind。ThemostcelebratedhorticulturistinFrance，namely，Vilmorin（22/28。QuotedbyVerlot’DesVarietes’etc。1865page28。），evenmaintainsthat，whenanyparticularvariationisdesired，thefirststepistogettheplanttovaryinanymannerwhatever，andtogoonselectingthemostvariableindividuals，eventhoughtheyvaryinthewrongdirection;forthefixedcharacterofthespeciesbeingoncebroken，thedesiredvariationwillsoonerorlaterappear。
Asnearlyallouranimalsweredomesticatedatanextremelyremoteepoch，wecannot，ofcourse，saywhethertheyvariedquicklyorslowlywhenfirstsubjectedtonewconditions。ButDr。Bachman（22/29。’ExaminationoftheCharacteristicsofGeneraandSpecies’Charleston1855page14。）statesthathehasseenturkeysraisedfromtheeggsofthewildspecieslosetheirmetallictintsandbecomespottedwithwhiteinthethirdgeneration。Mr。
YarrellmanyyearsagoinformedmethatthewildducksbredonthepondsinSt。James’sPark，whichhadneverbeencrossed，asitisbelieved，withdomesticducks，losttheirtrueplumageafterafewgenerations。Anexcellentobserver（22/30。Mr。Hewitt’JournalofHort。’1863page39。），whohasoftenrearedducksfromtheeggsofthewildbird，andwhotookprecautionsthatthereshouldbenocrossingwithdomesticbreeds，hasgiven，aspreviouslystated，fulldetailsonthechangeswhichtheygraduallyundergo。Hefoundthathecouldnotbreedthesewildduckstrueformorethanfiveorsixgenerations，\"astheythenprovedsomuchlessbeautiful。Thewhitecollarroundtheneckofthemallardbecamemuchbroaderandmoreirregular，andwhitefeathersappearedintheducklings’wings。\"Theyincreasedalsoinsizeofbody;theirlegsbecamelessfine，andtheylosttheirelegantcarriage。
Fresheggswerethenprocuredfromwildbirds;butagainthesameresultfollowed。Inthesecasesoftheduckandturkeyweseethatanimals，likeplants，donotdepartfromtheirprimitivetypeuntiltheyhavebeensubjectedduringseveralgenerationstodomestication。Ontheotherhand，Mr。YarrellinformedmethattheAustraliandingos，bredintheZoologicalGardens，almostinvariablyproducedinthefirstgenerationpuppiesmarkedwithwhiteandothercolours;but，theseintroduceddingoshadprobablybeenprocuredfromthenatives，whokeeptheminasemi—domesticatedstate。Itiscertainlyaremarkablefactthatchangedconditionsshouldatfirstproduce，asfaraswecansee，absolutelynoeffect;butthattheyshouldsubsequentlycausethecharacterofthespeciestochange。InthechapteronpangenesisIshallattempttothrowalittlelightonthisfact。
Returningnowtothecauseswhicharesupposedtoinducevariability。Someauthors（22/31。Devay’MariagesConsanguins’pages97，125。InconversationI
havefoundtwoorthreenaturalistsofthesameopinion。）believethatcloseinterbreedinggivesthistendency，andleadstotheproductionofmonstrosities。Intheseventeenthchaptersomefewfactswereadvanced，showingthatmonstrositiesare，asitappears，occasionallythusinduced;andtherecanbenodoubtthatcloseinterbreedingcauseslessenedfertilityandaweakenedconstitution;henceitmayleadtovariability:butIhavenotsufficientevidenceonthishead。Ontheotherhand，closeinterbreeding，ifnotcarriedtoaninjuriousextreme，farfromcausingvariability，tendstofixthecharacterofeachbreed。
Itwasformerlyacommonbelief，stillheldbysomepersons，thattheimaginationofthemotheraffectsthechildinthewomb。（22/32。Mullerhasconclusivelyarguedagainstthisbelief，’ElementsofPhys。’Englishtranslationvolume21842page1405。）Thisviewisevidentlynotapplicabletotheloweranimals，whichlayunimpregnatedeggs，ortoplants。Dr。WilliamHunter，inthelastcentury，toldmyfatherthatduringmanyyearseverywomaninalargeLondonLying—inHospitalwasaskedbeforeherconfinementwhetheranythinghadspeciallyaffectedhermind，andtheanswerwaswrittendown;anditsohappenedthatinnooneinstancecouldacoincidencebedetectedbetweenthewoman’sanswerandanyabnormalstructure;butwhensheknewthenatureofthestructure，shefrequentlysuggestedsomefreshcause。Thebeliefinthepowerofthemother’simaginationmayperhapshavearisenfromthechildrenofasecondmarriageresemblingthepreviousfather，ascertainlysometimesoccurs，inaccordancewiththefactsgivenintheeleventhchapter。
CROSSINGASACAUSEOFVARIABILITY。
InanearlypartofthischapteritwasstatedthatPallas（22/33。’Act。Acad。
St。Petersburg’1780part2page84etc。）andafewothernaturalistsmaintainthatvariabilityiswhollyduetocrossing。Ifthismeansthatnewcharactersneverspontaneouslyappearinourdomesticraces，butthattheyarealldirectlyderivedfromcertainaboriginalspecies，thedoctrineislittlelessthanabsurd;foritimpliesthatanimalslikeItaliangreyhounds，pug—dogs，bull—dogs，pouterandfantailpigeons，etc。，wereabletoexistinastateofnature。Butthedoctrinemaymeansomethingwidelydifferent，namely，thatthecrossingofdistinctspeciesisthesolecauseofthefirstappearanceofnewcharacters，andthatwithoutthisaidmancouldnothaveformedhisvariousbreeds。As，however，newcharactershaveappearedincertaincasesbybud—
variation，wemayconcludewithcertaintythatcrossingisnotnecessaryforvariability。Itis，moreover，certainthatthebreedsofvariousanimals，suchasoftherabbit，pigeon，duck，etc。，andthevarietiesofseveralplants，arethemodifieddescendantsofasinglewildspecies。Nevertheless，itisprobablethatthecrossingoftwoforms，whenoneorbothhavelongbeendomesticatedorcultivated，addstothevariabilityoftheoffspring，independentlyofthecomminglingofthecharactersderivedfromthetwoparent—forms;andthisimpliesthatnewcharactersactuallyarise。Butwemustnotforgetthefactsadvancedinthethirteenthchapter，whichclearlyprovethattheactofcrossingoftenleadstothereappearanceorreversionoflong—
lostcharacters;andinmostcasesitwouldbeimpossibletodistinguishbetweenthereappearanceofancientcharactersandthefirstappearanceofabsolutelynewcharacters。Practically，whetherneworold，theywouldbenewtothebreedinwhichtheyreappeared。
[Gartnerdeclares（22/34。’Bastarderzeugung’s。249，255，295。），andhisexperienceisofthehighestvalueonsuchapoint，that，whenhecrossednativeplantswhichhadnotbeencultivated，heneveroncesawintheoffspringanynewcharacter;butthatfromtheoddmannerinwhichthecharactersderivedfromtheparentswerecombined，theysometimesappearedasifnew。When，ontheotherhand，hecrossedcultivatedplants，headmitsthatnewcharactersoccasionallyappeared，butheisstronglyinclinedtoattributetheirappearancetoordinaryvariability，notinanywaytothecross。Anoppositeconclusion，however，appearstomethemoreprobable。AccordingtoKolreuter，hybridsinthegenusMirabilisvaryalmostinfinitely，andhedescribesnewandsingularcharactersintheformoftheseeds，inthecolouroftheanthers，inthecotyledonsbeingofimmensesize，innewandhighlypeculiarodours，intheflowersexpandingearlyintheseason，andintheirclosingatnight。Withrespecttoonelotofthesehybrids，heremarksthattheypresentedcharactersexactlythereverseofwhatmighthavebeenexpectedfromtheirparentage。（22/35。’NovaActa，St。Petersburg’1794page378;1795
pages307，313，316;1787page407。）
Prof。Lecoq（22/36。’DelaFecondation’1862page311。）speaksstronglytothesameeffectinregardtothissamegenus，andassertsthatmanyofthehybridsfromMirabilisjalapaandmultifloramighteasilybemistakenfordistinctspecies，andaddsthattheydifferedinagreaterdegreethantheotherspeciesofthegenus，fromM。jalapa。Herbert，also，hasdescribed（22/37。
’Amaryllidaceae’1837page362。）certainhybridRhododendronsasbeing\"asUNLIKEALLOTHERSinfoliage，asiftheyhadbeenaseparatespecies。\"Thecommonexperienceoffloriculturistsprovesthatthecrossingandrecrossingofdistinctbutalliedplants，suchasthespeciesofPetunia，Calceolaria，Fuchsia，Verbena，etc。，inducesexcessivevariability;hencetheappearanceofquitenewcharactersisprobable。M。Carriere（22/38。Abstractedin’Gardener’sChronicle’1860page1081。）haslatelydiscussedthissubject:hestatesthatErythrinacristagallihadbeenmultipliedbyseedformanyyears，buthadnotyieldedanyvarieties:itwasthencrossedwiththealliedE。
herbacea，and\"theresistancewasnowovercome，andvarietieswereproducedwithflowersofextremelydifferentsize，form，andcolour。\"
Fromthegeneralandapparentlywell—foundedbeliefthatthecrossingofdistinctspecies，besidescomminglingtheircharacters，addsgreatlytotheirvariability，ithasprobablyarisenthatsomebotanistshavegonesofarastomaintain（22/39。ThiswastheopinionoftheelderDeCandolle，asquotedin’Dic。Class。d’Hist。Nat。’tome8page405。Puvisinhiswork’DelaDegeneration’1837page37，hasdiscussedthissamepoint。）that，whenagenusincludesonlyasinglespecies，thiswhencultivatednevervaries。Thepropositionmadesobroadlycannotbeadmitted;butitisprobablytruethatthevariabilityofmonotypicgenerawhencultivatedisgenerallylessthanthatofgeneraincludingnumerousspecies，andthisquiteindependentlyoftheeffectsofcrossing。Ihaveshowninmy’OriginofSpecies’thatthespeciesbelongingtosmallgeneragenerallyyieldalessnumberofvarietiesinastateofnaturethanthosebelongingtolargegenera。Hencethespeciesofsmallgenerawould，itisprobable，producefewervarietiesundercultivationthanthealreadyvariablespeciesoflargergenera。
Althoughwehavenotatpresentsufficientevidencethatthecrossingofspecies，whichhaveneverbeencultivated，leadstotheappearanceofnewcharacters，thisapparentlydoesoccurwithspecieswhichhavebeenalreadyrenderedinsomedegreevariablethroughcultivation。Hencecrossing，likeanyotherchangeintheconditionsoflife，seemstobeanelement，probablyapotentone，incausingvariability。Butweseldomhavethemeansofdistinguishing，aspreviouslyremarked，betweentheappearanceofreallynewcharactersandthereappearanceoflong—lostcharacters，evokedthroughtheactofcrossing。Iwillgiveaninstanceofthedifficultyindistinguishingsuchcases。ThespeciesofDaturamaybedividedintotwosections，thosehavingwhiteflowerswithgreenstems，andthosehavingpurpleflowerswithbrownstems:nowNaudin（22/40。’ComptesRendus’Novembre21，1864page838。）
crossedDaturalaevisandferox，bothofwhichbelongtothewhitesection，andraisedfromthem205hybrids。Ofthesehybrids，everyonehadbrownstemsandborepurpleflowers;sothattheyresembledthespeciesoftheothersectionofthegenus，andnottheirowntwoparents。Naudinwassomuchastonishedatthisfact，thathewasledcarefullytoobservebothparent—
species，andhediscoveredthatthepureseedlingsofD。ferox，immediatelyaftergermination，haddarkpurplestems，extendingfromtheyoungrootsuptothecotyledons，andthatthistintremainedeverafterwardsasaringroundthebaseofthestemoftheplantwhenold。NowIhaveshowninthethirteenthchapterthattheretentionorexaggerationofanearlycharacterissointimatelyrelatedtoreversion，thatitevidentlycomesunderthesameprinciple。Henceprobablyweoughttolookatthepurpleflowersandbrownstemsofthesehybrids，notasnewcharactersduetovariability，butasareturntotheformerstateofsomeancientprogenitor。
Independentlyoftheappearanceofnewcharactersfromcrossing，afewwordsmaybeaddedtowhathasbeensaidinformerchaptersontheunequalcombinationandtransmissionofthecharacterspropertothetwoparent—forms。
Whentwospeciesorracesarecrossed，theoffspringofthefirstgenerationaregenerallyuniform，butthosesubsequentlyproduceddisplayanalmostinfinitediversityofcharacter。Hewhowishes，saysKolreuter（22/41。’NovaActa，St。Petersburg’1794page391。），toobtainanendlessnumberofvarietiesfromhybridsshouldcrossandrecrossthem。Thereisalsomuchvariabilitywhenhybridsormongrelsarereducedorabsorbedbyrepeatedcrosseswitheitherpureparent—form:andastillhigherdegreeofvariabilitywhenthreedistinctspecies，andmostofallwhenfourspecies，areblendedtogetherbysuccessivecrosses。BeyondthispointGartner（22/42。
’Bastarderzeugung’s。507，516，572。），onwhoseauthoritytheforegoingstatementsaremade，neversucceededineffectingaunion;butMaxWichura（22/43。’DieBastardbefruchtung’etc。1865s。24。）unitedsixdistinctspeciesofwillowsintoasinglehybrid。Thesexoftheparentspeciesaffectsinaninexplicablemannerthedegreeofvariabilityofhybrids;forGartner（22/44。
’Bastarderzeugung’s。452，507。）repeatedlyfoundthatwhenahybridwasusedasafatherandeitheroneofthepureparent—species，orathirdspecies，wasusedasthemother，theoffspringweremorevariablethanwhenthesamehybridwasusedasthemother，andeitherpureparentorthesamethirdspeciesasthefather:thusseedlingsfromDianthusbarbatuscrossedbythehybridD。
chinensi—barbatusweremorevariablethanthoseraisedfromthislatterhybridfertilisedbythepureD。barbatus。MaxWichura（22/45。’DieBastardbefruchtung’s。56。）insistsstronglyonananalogousresultwithhishybridwillows。AgainGartner（22/46。’Bastarderzeugung’s。423。）assertsthatthedegreeofvariabilitysometimesdiffersinhybridsraisedfromreciprocalcrossesbetweenthesametwospecies;andherethesoledifferenceis，thattheonespeciesisfirstusedasthefatherandthenasthemother。Onthewholeweseethat，independentlyoftheappearanceofnewcharacters，thevariabilityofsuccessivecrossedgenerationsisextremelycomplex，partlyfromtheoffspringpartakingunequallyofthecharactersofthetwoparent—
forms，andmoreespeciallyfromtheirunequaltendencytoreverttosuchcharactersortothoseofmoreancientprogenitors。]
ONTHEMANNERANDONTHEPERIODOFACTIONOFTHECAUSESWHICHINDUCE
VARIABILITY。
Thisisanextremelyobscuresubject，andweneedhereonlyconsider，whetherinheritedvariationsareduetocertainpartsbeingactedonaftertheyhavebeenformed，orthroughthereproductivesystembeingaffectedbeforetheirformation;andintheformercaseatwhatperiodofgrowthordevelopmenttheeffectisproduced。Weshallseeinthetwofollowingchaptersthatvariousagencies，suchasanabundantsupplyoffood，exposuretoadifferentclimate，increaseduseordisuseofparts，etc。，prolongedduringseveralgenerations，certainlymodifyeitherthewholeorganisationorcertainorgans;anditisclearatleastinthecaseofbud—variationthattheactioncannothavebeenthroughthereproductivesystem。
[Withrespecttothepartwhichthereproductivesystemtakesincausingvariability，wehaveseenintheeighteenthchapterthatevenslightchangesintheconditionsoflifehavearemarkablepowerincausingagreaterorlessdegreeofsterility。Henceitseemsnotimprobablethatbeingsgeneratedthroughasystemsoeasilyaffectedshouldthemselvesbeaffected，orshouldfailtoinherit，orinheritinexcess，characterspropertotheirparents。Weknowthatcertaingroupsoforganicbeings，butwithexceptionsineachgroup，havetheirreproductivesystemsmuchmoreeasilyaffectedbychangedconditionsthanothergroups;forinstance，carnivorousbirds，morereadilythancarnivorousmammals，andparrotsmorereadilythanpigeons;andthisfactharmoniseswiththeapparentlycapriciousmanneranddegreeinwhichvariousgroupsofanimalsandplantsvaryunderdomestication。
Kolreuter（22/47。’DritteFortsetzung’etc。1766s。85。）wasstruckwiththeparallelismbetweentheexcessivevariabilityofhybridswhencrossedandrecrossedinvariousways，——thesehybridshavingtheirreproductivepowersmoreorlessaffected，——andthevariabilityofancientlycultivatedplants。
MaxWichura（22/48。’DieBastardbefruchtung’etc。1865s。92:seealsotheRev。M。J。Berkeleyonthesamesubjectin’JournalofRoyalHort。Soc。’1866
page80。）hasgoneonestepfarther，andshowsthatwithmanyofourhighlycultivatedplants，suchasthehyacinth，tulip，auricula，snapdragon，potato，cabbage，etc。，whichthereisnoreasontobelievehavebeenhybridised，theantherscontainmanyirregularpollen—grainsinthesamestateasinhybrids。
Hefindsalsoincertainwildforms，thesamecoincidencebetweenthestateofthepollenandahighdegreeofvariability，asinmanyspeciesofRubus;butinR。caesiusandidaeus，whicharenothighlyvariablespecies，thepollenissound。Itisalsonotoriousthatmanycultivatedplants，suchasthebanana，pineapple，bread—fruit，andotherspreviouslymentioned，havetheirreproductiveorganssoseriouslyaffectedastobegenerallyquitesterile;
andwhentheydoyieldseed，theseedlings，judgingfromthelargenumberofcultivatedraceswhichexist，mustbevariableinanextremedegree。Thesefactsindicatethatthereissomerelationbetweenthestateofthereproductiveorgansandatendencytovariability;butwemustnotconcludethattherelationisstrict。Althoughmanyofourhighlycultivatedplantsmayhavetheirpolleninadeterioratedcondition，yet，aswehavepreviouslyseen，theyyieldmoreseeds，andourancientlydomesticatedanimalsaremoreprolific，thanthecorrespondingspeciesinastateofnature。Thepeacockisalmosttheonlybirdwhichisbelievedtobelessfertileunderdomesticationthaninitsnativestate，andithasvariedinaremarkablysmalldegree。Fromtheseconsiderationsitwouldseemthatchangesintheconditionsoflifeleadeithertosterilityortovariability，ortoboth;andnotthatsterilityinducesvariability。Onthewholeitisprobablethatanycauseaffectingtheorgansofreproductionwouldlikewiseaffecttheirproduct，——thatis，theoffspringthusgenerated。
Theperiodoflifeatwhichthecausesthatinducevariabilityact，islikewiseanobscuresubject，whichhasbeendiscussedbyvariousauthors。
（22/49。Dr。P。Lucashasgivenahistoryofopiniononthissubject’Hered。
Nat。’1847tome1page175。）Insomeofthecases，tobegiveninthefollowingchapter，ofmodificationsfromthedirectactionofchangedconditions，whichareinherited，therecanbenodoubtthatthecauseshaveactedonthematureornearlymatureanimal。Ontheotherhand，monstrosities，whichcannotbedistinctlyseparatedfromlesservariations，areoftencausedbytheembryobeinginjuredwhilstinthemother’swomborintheegg。ThusI。
GeoffroySaint—Hilaire（22/50。’Hist。desAnomalies’tome3page499。）assertsthatpoorwomenwhoworkhardduringtheirpregnancy，andthemothersofillegitimatechildrentroubledintheirmindsandforcedtoconcealtheirstate，arefarmoreliabletogivebirthtomonstersthanwomenineasycircumstances。Theeggsofthefowlwhenplaceduprightorotherwisetreatedunnaturallyfrequentlyproducemonstrouschickens。Itwould，however，appearthatcomplexmonstrositiesareinducedmorefrequentlyduringaratherlatethanduringaveryearlyperiodofembryoniclife;butthismaypartlyresultfromsomeonepart，whichhasbeeninjuredduringanearlyperiod，affectingbyitsabnormalgrowthotherpartssubsequentlydeveloped;andthiswouldbelesslikelytooccurwithpartsinjuredatalaterperiod。（22/51。Ibidtome3
pages392，502。TheseveralmemoirsbyM。Darestehereafterreferredtoareofspecialvalueonthiswholesubject。）Whenanypartororganbecomesmonstrousthroughabortion，arudimentisgenerallyleft，andthislikewiseindicatesthatitsdevelopmenthadalreadycommenced。
Insectssometimeshavetheirantennaeorlegsinamonstrouscondition，thelarvaeofwhichdonotpossesseitherantennaeorlegs;andinthesecases，asQuatrefages（22/52。Seehisinterestingwork’Metamorphosesdel’Homme’etc。
1862page129。）believes，weareenabledtoseethepreciseperiodatwhichthenormalprogressofdevelopmentwastroubled。Butthenatureofthefoodgiventoacaterpillarsometimesaffectsthecoloursofthemoth，withoutthecaterpillaritselfbeingaffected;thereforeitseemspossiblethatothercharactersinthematureinsectmightbeindirectlymodifiedthroughthelarvae。Thereisnoreasontosupposethatorganswhichhavebeenrenderedmonstroushavealwaysbeenactedonduringtheirdevelopment;thecausemayhaveactedontheorganisationatamuchearlierstage。Itisevenprobablethateitherthemaleorfemalesexualelements，orboth，beforetheirunion，maybeaffectedinsuchamannerastoleadtomodificationsinorgansdevelopedatalateperiodoflife;innearlythesamemannerasachildmayinheritfromhisfatheradiseasewhichdoesnotappearuntiloldage。
Inaccordancewiththefactsabovegiven，whichprovethatinmanycasesacloserelationexistsbetweenvariabilityandthesterilityfollowingfromchangedconditions，wemayconcludethattheexcitingcauseoftenactsattheearliestpossibleperiod，namely，onthesexualelements，beforeimpregnationhastakenplace。Thatanaffectionofthefemalesexualelementmayinducevariabilitywemaylikewiseinferasprobablefromtheoccurrenceofbud—
variations;forabudseemstobetheanalogueofanovule。Butthemaleelementisapparentlymuchofteneraffectedbychangedconditions，atleastinavisiblemanner，thanthefemaleelementorovuleandweknowfromGartner’sandWichura’sstatementsthatahybridusedasthefatherandcrossedwithapurespeciesgivesagreaterdegreeofvariabilitytotheoffspring，thandoesthesamehybridwhenusedasthemother。Lastly，itiscertainthatvariabilitymaybetransmittedthrougheithersexualelement，whetherornotoriginallyexcitedinthem，forKolreuterandGartner（22/53。’DritteFortsetzung’etc。s。123;’Bastarderzeugung’s。249。）foundthatwhentwospecieswerecrossed，ifeitheronewasvariable，theoffspringwererenderedvariable。]
SUMMARY。
Fromthefactsgiveninthischapter，wemayconcludethatthevariabilityoforganicbeingsunderdomestication，althoughsogeneral，isnotaninevitablecontingentonlife，butresultsfromtheconditionstowhichtheparentshavebeenexposed。Changesofanykindintheconditionsoflife，evenextremelyslightchanges，oftensufficetocausevariability。Excessofnutrimentisperhapsthemostefficientsingleexcitingcause。Animalsandplantscontinuetobevariableforanimmenseperiodaftertheirfirstdomestication;buttheconditionstowhichtheyareexposedneverlongremainquiteconstant。Inthecourseoftimetheycanbehabituatedtocertainchanges，soastobecomelessvariable;anditispossiblethatwhenfirstdomesticatedtheymayhavebeenevenmorevariablethanatpresent。Thereisgoodevidencethatthepowerofchangedconditionsaccumulates;sothattwo，three，ormoregenerationsmustbeexposedtonewconditionsbeforeanyeffectisvisible。Thecrossingofdistinctforms，whichhavealreadybecomevariable，increasesintheoffspringthetendencytofurthervariability，bytheunequalcomminglingofthecharactersofthetwoparents，bythereappearanceoflong—lostcharacters，andbytheappearanceofabsolutelynewcharacters。Somevariationsareinducedbythedirectactionofthesurroundingconditionsonthewholeorganisation，oroncertainpartsalone;othervariationsappeartobeinducedindirectlythroughthereproductivesystembeingaffected，asweknowisoftenthecasewithvariousbeings，whichwhenremovedfromtheirnaturalconditionsbecomesterile。Thecauseswhichinducevariabilityactonthematureorganism，ontheembryo，and，probably，onthesexualelementsbeforeimpregnationhasbeeneffected。
CHAPTER2。XXIII。
DIRECTANDDEFINITEACTIONOFTHEEXTERNALCONDITIONSOFLIFE。
SLIGHTMODIFICATIONSINPLANTSFROMTHEDEFINITEACTIONOFCHANGEDCONDITIONS，INSIZE，COLOUR，CHEMICALPROPERTIES，ANDINTHESTATEOFTHETISSUES。
LOCALDISEASES。
CONSPICUOUSMODIFICATIONSFROMCHANGEDCLIMATEORFOOD，ETC。
PLUMAGEOFBIRDSAFFECTEDBYPECULIARNUTRIMENT，ANDBYTHEINOCULATIONOF
POISON。
LAND—SHELLS。
MODIFICATIONSOFORGANICBEINGSINASTATEOFNATURETHROUGHTHEDEFINITE
ACTIONOFEXTERNALCONDITIONS。
COMPARISONOFAMERICANANDEUROPEANTREES。
GALLS。
EFFECTSOFPARASITICFUNGI。
CONSIDERATIONSOPPOSEDTOTHEBELIEFINTHEPOTENTINFLUENCEOFCHANGED
EXTERNALCONDITIONS。
PARALLELSERIESOFVARIETIES。
AMOUNTOFVARIATIONDOESNOTCORRESPONDWITHTHEDEGREEOFCHANGEINTHE
CONDITIONS。
BUD—VARIATION。
MONSTROSITIESPRODUCEDBYUNNATURALTREATMENT。
SUMMARY。