Darwinian；butwhatevidenceotherthaninferentialcanfromthenatureofthecasebeadducedinsupportofthis，asIbelieve，perfectlycorrectjudgment？NoneknowbetterthantheywhoclamourfordirectevidencethattheirmasterwasrightintakingthepositionassignedtohimbyProfessorVines，thattheycannotreasonablylookforit。Withus，aswiththemselves，modificationproceedsverygradually，anditviolatesourprinciplesasmuchastheirowntoexpectvisiblepermanentprogress，inanysinglegeneration，orindeedinanynumberofgenerationsofwildspecieswhichwehaveyethadtimetoobserve。Occasionallywecanfindsuchcases，asinthatofBranchipusstagnalis，quotedbyMr。
Wallace，orinthatoftheNewZealandKeawhoseskin，IwasassuredbythelateSirJuliusvonHaast，hasalreadybeenmodifiedasaconsequenceofitschangeoffood。Herewecanshowthatinevenafewgenerationsstructureismodifiedunderchangedconditionsofexistence，butaswebelievethesecasestooccurcomparativelyrarely，soitisstillmorerarelythattheyoccurwhenandwherewecanwatchthem。Natureiseminentlyconservative，andfixityoftype，evenunderconsiderablechangeofconditions，issurelymoreimportantforthewell-beingofanyspeciesthananover-readypowerofadaptationto，itmaybe，passingchanges。Therecouldbenosteadyprogressifeachgenerationwerenotmainlyboundbythetraditionsofthosethathavegonebeforeit。Itisevolutionandnotincessantrevolutionthatbothpartiesareupholding；andthisbeingso，rapidvisiblemodificationmustbetheexception，nottherule。Ihavequoteddirectevidenceadducedbycompetentobservers，whichis，Ibelieve，sufficienttoestablishthefactthatoffspringcanbeandissometimesmodifiedbytheacquiredhabitsofaprogenitor。Iwillnowproceedtothestillmore，asitappearstome，cogentproofaffordedbygeneralconsiderations。
What，letmeask，aretheprincipalphenomenaofheredity？Theremustbephysicalcontinuitybetweenparent，orparents，andoffspring，sothattheoffspringis，asErasmusDarwinwellsaid，akindofelongationofthelifeoftheparent。
ErasmusDarwinputthemattersowellthatImayaswellgivehiswordsinfull；hewrote：-
\"Owingtotheimperfectionoflanguagetheoffspringistermedanewanimal，butisintruthabranchorelongationoftheparent，sinceapartoftheembryonanimalis，orwas，apartoftheparent，andtherefore，instrictlanguage，cannotbesaidtobeentirelynewatthetimeofitsproduction；andthereforeitmayretainsomeofthehabitsoftheparentsystem。