Soc。Bot。’1862page94。）;andtheseareplantswhichfloristsspeakofasnevercomingtruebyseed:butifcarebetakentopreventcrossing，neitherspeciesisbyanymeansveryinconstant，incolour;thusIraisedtwenty—threeplantsfromapurpleprimrose，fertilisedbyMr。J。Scottwithitspollen，andeighteencameuppurpleofdifferentshades，andonlyfiverevertedtotheordinaryyellowcolour:again，Iraisedtwentyplantsfromabright—redcowslip，similarlytreatedbyMr。Scott，andeveryoneperfectlyresembleditsparentincolour，aslikewisedid，withtheexceptionofasingleplant，72grandchildren。Evenwiththemostvariableflowers，itisprobablethateachdelicateshadeofcolourmightbepermanentlyfixedsoastobetransmittedbyseed，bycultivationinthesamesoil，bylong—continuedselection，andespeciallybythepreventionofcrosses。Iinferthisfromcertainannuallarkspurs（Delphiniumconsolidaandajacis），ofwhichcommonseedlingspresentagreaterdiversityofcolourthananyotherplantknowntome;yetonprocuringseedoffivenamedGermanvarietiesofD。consolida，onlynineplantsoutofninety—fourwerefalse;andtheseedlingsofsixvarietiesofD。ajacisweretrueinthesamemanneranddegreeaswiththestocksabovedescribed。A
distinguishedbotanistmaintainsthattheannualspeciesofDelphiniumarealwaysself—fertilised;thereforeImaymentionthatthirty—twoflowersonabranchofD。consolida，enclosedinanet，yieldedtwenty—sevencapsules，withanaverageof17。2seedineach;whilstfiveflowers，underthesamenet，whichwereartificiallyfertilised，inthesamemannerasmustbeeffectedbybeesduringtheirincessantvisits，yieldedfivecapsuleswithanaverageof35。2fineseed;andthisshowsthattheagencyofinsectsisnecessaryforthefullfertilityofthisplant。Analogousfactscouldbegivenwithrespecttothecrossingofmanyotherflowers，suchascarnations，etc。，ofwhichthevarietiesfluctuatemuchincolour。
Aswithflowers，sowithourdomesticatedanimals，nocharacterismorevariablethancolour，andprobablyinnoanimalmoresothanwiththehorse。Yet，withalittlecareinbreeding，itappearsthatracesofanycolourmightsoonbeformed。Hofackergivestheresultofmatchingtwohundredandsixteenmaresoffourdifferentcolourswithlike—colouredstallions，withoutregardtothecolouroftheirancestors;andofthetwohundredandsixteencoltsborn，elevenalonefailedtoinheritthecolouroftheirparents:AutenriethandAmmonassertthat，aftertwogenerations，coltsofauniformcolourareproducedwithcertainty。（12/50。Hofacker’UeberdieEigenschaften’etc。s。10。）]
Inafewrarecasespeculiaritiesfailtobeinherited，apparentlyfromtheforceofinheritancebeingtoostrong。Ihavebeenassuredbybreedersofthecanary—birdthattogetagoodjonquil—colouredbirditdoesnotanswertopairtwojonquils，asthecolourthencomesouttoostrong，orisevenbrown;butthisstatementisdisputedbyotherbreeders。Soagain，iftwocrestedcanariesarepaired，theyoungbirdsrarelyinheritthischaracter（12/51。Bechstein’Naturgesch。Deutschlands’b。4s。462。Mr。Brent，agreatbreederofcanaries，informsmethathebelievesthatthesestatementsarecorrect。）:forincrestedbirdsanarrowspaceofbareskinisleftonthebackofthehead，wherethefeathersareup—turnedtoformthecrest，and，whenbothparentsarethuscharacterised，thebarenessbecomesexcessive，andthecrestitselffailstobedeveloped。Mr。Hewitt，speakingofLacedSebrightBantams，says（12/52。’ThePoultryBook’byW。B。
Tegetmeier1866page245。）that，\"whythisshouldbesoIknownot，butI