ItmayhappenthattheOsmia'spatienceisinvainandthatthebarricadethatblocksthewayneverdisappearsatall。Sometimes,theegginacelldoesnotmature;andtheunconsumedprovisionsdryupandbecomeacompact,sticky,mildewedplug,throughwhichtheoccupantsofthefloorsbelowcouldneverclearthemselvesapassage。
　　Sometimes,again,agrubdiesinitscocoon;andthecradleofthedeceased,nowturnedintoacoffin,formsaneverlastingobstacle。
　　Howshalltheinsectcopewithsuchgravecircumstances？
　　Amongthemanybramble—stumpswhichIhavecollected,somefewhavepresentedaremarkablepeculiarity。Inadditiontotheorificeatthetop,theyhadatthesideoneandsometimestworoundaperturesthatlookedasthoughtheyhadbeenpunchedoutwithaninstrument。Onopeningthesestalks,whichwereold,desertednests,Idiscoveredthecauseoftheseveryexceptionalwindows。Aboveeachofthemwasacellfullofmouldyhoney。Theegghadperishedandtheprovisionsremaineduntouched:hencetheimpossibilityofgettingoutbytheordinaryroad。Walledinbytheunsurmountableobstacle,theOsmiaonthefloorbelowhadcontrivedanoutletthroughthesideoftheshaft;andthoseinthelowerstoreyshadbenefitedbythisingeniousinnovation。Theusualdoorbeinginaccessible,aside—windowhadbeenopenedbymeansoftheinsect'sjaws。Thecocoons,torn,butstillinpositioninthelowerrooms,leftnodoubtastothiseccentricmodeofexit。Thesamefact,moreover,wasrepeated,inseveralbramble—
　　stumps,inthecaseofOsmiatridentata;itwaslikewiserepeatedinthecaseofAnthidiumscapulare。Theobservationwasworthconfirmingbyexperiment。
　　Iselectabramble—stemwiththethinnestrindpossible,soastofacilitatetheOsmiae'swork。Isplititinhalf,thusobtainingasmooth—sidedtroughwhichwillenablemetojudgebetteroffutureexits。Thecocoonsarenextlaidoutinoneofthetroughs。I
　　separatethemwithdisksofsorghum,coveringbothsurfacesofthediskwithagenerouslayerofsealing—wax,amaterialwhichtheOsmia'smandiblesarenotabletoattack。Thetwotroughsarethenplacedtogetherandfastened。Alittleputtydoesawaywiththejointandpreventstheleastrayoflightfrompenetrating。Lastly,theapparatusishungupperpendicularly,withthecocoons'headsup。Wehavenowonlytowait。NoneoftheOsmiaecangetoutintheusualmanner,becauseeachofthemisconfinedbetweentwopartitionscoatedwithsealing—wax。Thereisbutoneresourcelefttothemiftheywouldemergeintothelightofday,thatis,foreachofthemtoopenaside—window,providedalwaysthattheypossesstheinstinctandthepowertodoso。
　　InJuly,theresultisasfollows:oftwentyOsmiaethusimmured,sixsucceedinboringaroundholethroughthewallandmakingtheirwayout;theothersperishintheircells,withoutmanagingtoreleasethemselves。But,whenIopenthecylinder,whenIseparatethetwowoodentroughs,Irealizethatallhaveattemptedtoescapethroughtheside,forthewallofeachcellbearstracesofgnawingconcentratedupononespot。All,therefore,haveactedinthesamewayastheirmorefortunatesisters;theydidnotsucceed,becausetheirstrengthfailedthem。Lastly,inmyglasstubes,part—linedwithathickpieceofpacking—paper,Ioftenseeattemptsatmakingawindowinthesideofthecell:thepaperispiercedrightthroughwitharoundhole。
　　ThisthenisyetanotherresultwhichIamgladtorecordinthehistoryofthebramble—dwellers。WhentheOsmia,theAnthidiumandprobablyothersareunabletoemergethroughthecustomaryoutlet,theytakeanheroicdecisionandperforatethesideoftheshaft。Itisthelastresource,resolveduponafterothermethodshavebeentriedinvain。Thebrave,thestrongsucceed;theweakperishintheattempt。
　　SupposingthatalltheOsmiaepossessedthenecessarystrengthofjawaswellastheinstinctforthissidewardboring,itisclearthategressfromeachcellthroughaspecialwindowwouldbemuchmoreadvantageousthanegressthroughthecommondoor。TheBeecouldattendtohisreleaseassoonashewashatched,insteadofpostponingituntilaftertheemancipationofthosewhocomebeforehim;hewouldthusescapelongwaits,whichtoooftenprovefatal。Inpointoffact,itisnouncommonthingtofindbramble—stalksinwhichseveralOsmiaehavediedintheircells,becausetheupperstoreyswerenotvacatedintime。Yes,therewouldbeapreciousadvantageinthatlateralopening,whichwouldnotleaveeachoccupantatthemercyofhisenvironment:manydiethatwouldnotdie。AlltheOsmiae,whencompelledbycircumstances,resorttothissuprememethod;allhavetheinstinctforlateralboring;butveryfewareabletocarrytheworkthrough。Onlythefavouritesoffatesucceed,thosemoregenerouslyendowedwithstrengthandperseverance。
　　Ifthefamouslawofnaturalselection,whichissaidtogovernandtransformtheworld,hadanysurefoundation;ifreallythefittestremovedthelessfitfromthescene;ifthefutureweretothestrongest,tothemostindustrious,surelytheraceofOsmiae,whichhasbeenperforatingbramble—stumpsforages,shouldbythistimehavealloweditsweakermembers,whogoonobstinatelyusingthecommonoutlet,todieoutandshouldhavereplacedthem,downtotheverylastone,bythestalwartdrillersofside—openings。Thereisanopportunityhereforimmenseprogress;theinsectisonthevergeofitandisunabletocrossthenarrowinterveningline。Selectionhashadampletimetomakeitschoice;andyet,thoughtherebeafewsuccesses,thefailuresexceedtheminverylargemeasure。Theraceofthestronghasnotabolishedtheraceoftheweak:itremainsinferiorinnumbers,asdoubtlessithasbeensincealltime。Thelawofnaturalselectionimpressesmewiththevastnessofitsscope;
　　but,wheneverItrytoapplyittoactualfacts,itleavesmewhirlinginspace,withnothingtohelpmetointerpretrealities。Itismagnificentintheory,butitisameregas—bubbleinthefaceofexistingconditions。Itismajestic,butsterile。Thenwhereistheanswertotheriddleoftheworld？Whoknows？Whowilleverknow？
　　Letuswastenomoretimeinthisdarkness,whichidletheorizingwillnotdispel;letusreturntofacts,humblefacts,theonlygroundthatdoesnotgivewayunderourfeet。TheOsmiarespectsherneighbour'scocoon;andherscruplesaresogreatthat,aftervainlytryingtoslipbetweenthatcocoonandthewall,orelsetoopenalateraloutlet,sheletsherselfdieinhercellratherthaneffectanegressbyforcingherwaythroughtheoccupiedcells。Whenthecocoonthatblocksthewaycontainsadeadinsteadofalivegrub,willtheresultbethesame？
　　Inmyglasstubes,IletOsmia—cocoonscontainingalivegrubalternatewithOsmia—cocoonsinwhichthegrubhasbeenasphyxiatedbythefumesofsulphocarbonicacid。Asusual,thestoreysareseparatedbydisksofsorghum。Theanchorites,whenhatched,donothesitatelong。Oncethepartitionispierced,theyattackthedeadcocoons,gorightthroughthem,reducingthedeadgrub,nowdryandshrivelled,todust,andatlastemerge,afterwreckingeverythingintheirpath。Thedeadcocoons,therefore,arenotspared;theyaretreatedaswouldbeanyotherobstaclecapableofattackbythemandibles。TheOsmialooksuponthemasamerebarricadetoberuthlesslyoverturned。Howissheapprisedthatthecocoon,whichhasundergonenooutwardchange,containsadeadandnotalivegrub？Itiscertainlynotbysight。Canitbebysenseofsmell？Iamalwaysalittlesuspiciousofthatsenseofsmellofwhichwedonotknowtheseatandwhichweintroduceontheslightestprovocationasaconvenientexplanationofthatwhichmaytranscendourexplanatorypowers。
　　Mynexttestismadewithastringoflivecocoons。Ofcourse,I
　　cannottakeallthesefromthesamespecies,forthentheexperimentwouldnotdifferfromtheonewhichwehavealreadywitnessed;Itakethemfromtwodifferentspecieswhichleavetheirbramble—stematseparateperiods。Moreover,thesecocoonsmusthavenearlythesamediametertoallowoftheirbeingstackedinatubewithoutleavinganemptyspacebetweenthemandthewall。ThetwospeciesadoptedareSoleniusvagus,whichquitsthebrambleattheendofJune,andOsmiadetrita,whichcomesalittleearlier,inthefirstfortnightofthesamemonth。IthereforealternateOsmia—cocoonsandSolenius—cocoons,withthelatteratthetopoftheseries,eitheringlasstubesorbetweentwobramble—troughsjoinedintoacylinder。
　　Theresultofthispromiscuityisstriking。TheOsmiae,whichmatureearlier,emerge;andtheSolenius—cocoons,aswellastheirinhabitants,whichbythistimehavereachedtheperfectstage,arereducedtoshreds,todust,whereinitisimpossibleformetorecognizeavestige,saveperhapshereandthereahead,oftheexterminatedunfortunates。TheOsmia,therefore,hasnotrespectedthelivecocoonsofaforeignspecies:shehaspassedoutoverthebodiesoftheinterveningSolenii。DidIsaypassedovertheirbodies？Shehaspassedthroughthem,crunchedthelaggardsbetweenherjaws,treatedthemascavalierlyasshetreatsmydisks。Andyetthosebarricadeswerealive。Nomatter:whenherhourcame,theOsmiawentahead,destroyingeverythinguponherroad。Here,atanyrate,isalawonwhichwecanrely:thesupremeindifferenceoftheanimaltoallthatdoesnotformpartofitselfanditsrace。
　　Andwhatofthesenseofsmell,distinguishingthedeadfromtheliving？Here,allarealive;andtheBeepiercesherwayasthrougharowofcorpses。IfIamtoldthatthesmelloftheSoleniimaydifferfromthatoftheOsmiae,Ishallreplythatsuchextremesubtletyintheinsect'solfactoryapparatusseemstomearatherfar—fetchedsupposition。Thenwhatismyexplanationofthetwofacts？Theexplanation？Ihavenonetogive！IamquitecontenttoknowthatI
　　donotknow,whichatleastsparesmemanyvainlucubrations。AndsoIdonotknowhowtheOsmia,inthedensedarknessofhertunnel,distinguishesbetweenalivecocoonandadeadcocoonofthesamespecies;andIknowjustaslittlehowshesucceedsinrecognizingastrangecocoon。Ah,howclearlythisconfessionofignoranceprovesthatIambehindthetimes！Iamdeliberatelymissingagloriousopportunityofstringingbigwordstogetherandarrivingatnothing。
　　Thebramble—stumpisperpendicular,ornearlyso;itsopeningisatthetop。Thisistheruleundernaturalconditions。Myartificesareabletoalterthatstateofthings;Icanplacethetubeverticallyorhorizontally;Icanturnitsoneorificeeitherupordown;
　　lastly,Icanleavethechannelopenatbothends,whichwillgivetwooutlets。Whatwillhappenundertheseseveralconditions？ThatiswhatweshallexaminewiththeThree—prongedOsmia。
　　Thetubeishungperpendicularly,butclosedatthetopandopenatthebottom;infact,itrepresentsabramble—stumpturnedupsidedown。Tovaryandcomplicatetheexperiment,thestringsofcocoonsarearrangeddifferentlyindifferenttubes。Insomeofthem,theheadsofthecocoonsareturneddownwards,towardstheopening;inothers,theyareturnedupwards,towardstheclosedend;inothersagain,thecocoonsalternateindirection,thatistosay,theyareplacedheadtoheadandreartorear,turnandturnabout。Ineednotsaythattheseparatingfloorsareofsorghum。
　　Theresultisidenticalinallthesetubes。IftheOsmiaehavetheirheadspointingupwards,theyattackthepartitionabovethem,ashappensundernormalconditions;iftheirheadspointdownwards,theyturnroundintheircellsandsettoworkasusual。Inshort,thegeneraloutwardtrendistowardsthetop,inwhateverpositionthecocoonbeplaced。
　　Wehereseemanifestlyatworktheinfluenceofgravity,whichwarnstheinsectofitsreversedpositionandmakesitturnround,evenasitwouldwarnusifweourselveshappenedtobehangingheaddownwards。Innaturalconditions,theinsecthasbuttofollowthecounselsofgravity,whichtellsittodigupwards,anditwillinfalliblyreachtheexit—doorsituatedattheupperend。But,inmyapparatus,thesesamecounselsbetrayit:itgoestowardsthetop,wherethereisnooutlet。Thusmisledbymyartifices,theOsmiaeperish,heapeduponthehigherfloorsandburiedintheruins。
　　Itneverthelesshappensthatattemptsaremadetocleararoaddownwards。Butitisrarefortheworktoleadtoanythinginthisdirection,especiallyinthecaseofthemiddleoruppercells。Theinsectislittleinclinedforthisprogress,theoppositetothattowhichitisaccustomed;besides,aseriousdifficultyarisesinthecourseofthisreversedboring。AstheBeeflingstheexcavatedmaterialsbehindher,thesefallbackoftheirownweightunderhermandibles;theclearancehastobebegunanew。ExhaustedbyherSisypheantask,distrustfulofthisnewandunfamiliarmethod,theOsmiaresignsherselfandexpiresinhercell。Iamboundtoadd,however,thattheOsmiaeinthelowerstoreys,thosenearesttheexit——sometimesone,sometimestwoorthree——dosucceedinescaping。
　　Inthatcase,theyunhesitatinglyattackthepartitionsbelowthem,whiletheircompanions,whoformthegreatmajority,persistandperishintheuppercells。
　　Itwaseasytorepeattheexperimentwithoutchanginganythinginthenaturalconditions,exceptthedirectionofthecocoons:allthatI
　　hadtodowastohangupsomebramble—stumpsasIfoundthem,vertically,butwiththeopeningdownwards。OutoftwostalksthusarrangedandpeopledwithOsmiae,notoneoftheinsectssucceededinemerging。AlltheBeesdiedintheshaft,someturnedupwards,othersdownwards。Ontheotherhand,threestemsoccupiedbyAnthidiadischargedtheirpopulationsafeandsound。Theoutgoingwaseffectedatthebottom,fromfirsttolast,withouttheleastimpediment。MustwetakeitthatthetwosortsofBeesarenotequallysensitivetotheinfluencesofgravity？CantheAnthidium,builttopassthroughthedifficultobstacleofhercottonwallets,bebetter—adaptedthantheOsmiatomakeherwaythroughthewreckagethatkeepsfallingundertheworker'sfeet;or,rather,maynotthisverycotton—wasteputastoptothesecataractsofrubbishwhichmustnaturallydrivetheinsectback？Thisisallquitepossible;butIcansaynothingforcertain。
　　Letusnowexperimentwithverticaltubesopenatbothends。Thearrangements,savefortheupperorifice,arethesameasbefore。Thecocoons,insomeofthetubes,havetheirheadsturneddown;others,up;inothersagain,theirpositionsalternate。Theresultissimilartowhatwehaveseenabove。AfewOsmiae,thosenearestthebottomorifice,takethelowerroad,whateverthedirectionfirstoccupiedbythecocoon;theothers,composingbyfarthelargernumber,takethehigherroad,evenwhenthecocoonisplacedupsidedown。Asbothdoorsarefree,theoutgoingiseffectedateitherendwithsuccess。
　　Whatarewetoconcludefromalltheseexperiments？First,thatgravityguidestheinsecttowardsthetop,wherethenaturaldooris,andmakesitturninitscellwhenthecocoonhasbeenreversed。
　　Secondly,Iseemtosuspectanatmosphericinfluenceand,inanycase,somesecondcausethatsendstheinsecttotheoutlet。Letusadmitthatthiscauseistheproximityoftheouterairactingupontheanchoritethroughthepartitions。
　　Theanimalthenissubject,ontheonehand,tothepromptingsofgravity,andthistoanequaldegreeforall,whateverthestoreyinhabited。Gravityisthecommonguideofthewholeseriesfrombasetotop。Butthoseinthelowerboxeshaveasecondguide,whenthebottomendisopen。Thisisthestimulusoftheadjacentair,amorepowerfulstimulusthanthatofgravity。Theaccessoftheairfromwithoutisveryslight,becauseofthepartitions;whileitcanbefeltinthenethermostcells,itmustdecreaserapidlyasthestoreysascend。Whereforethebottominsects,veryfewinnumber,obeyingthepreponderantinfluence,thatoftheatmosphere,makefortheloweroutletandreverse,ifnecessary,theiroriginalposition;thoseabove,onthecontrary,whoformthegreatmajority,beingguidedonlybygravitywhentheupperendisclosed,makeforthatupperend。Itgoeswithoutsayingthat,iftheupperendbeopenatthesametimeastheother,theoccupantsofthetopstoreyswillhaveadoubleincentivetotaketheascendingpath,thoughthiswillnotpreventthedwellersonthelowerfloorsfromobeying,bypreference,thecalloftheadjacentairandadoptingthedownwardroad。
　　Ihaveonemeansleftwherebytojudgeofthevalueofmyexplanation,namely,toexperimentwithtubesopenatbothendsandlyinghorizontally。Thehorizontalpositionhasatwofoldadvantage。
　　Inthefirstplace,itremovestheinsectfromtheinfluenceofgravity,inasmuchasitleavesitindifferenttothedirectiontobetaken,therightortheleft。Inthesecondplace,itdoesawaywiththedescentoftherubbishwhich,fallingundertheworker'sfeetwhentheboringisdonefrombelow,soonerorlaterdiscouragesherandmakesherabandonherenterprise。
　　Thereareafewprecautionstobeobservedforthesuccessfulconductoftheexperiment;Irecommendthemtoanyonewhomightcaretomaketheattempt。ItisevenadvisabletoremembertheminthecaseofthetestswhichIhavealreadydescribed。Themales,thosepunycreatures,notbuiltforwork,aresorrylabourerswhenconfrontedwithmystoutdisks。Mostofthemperishmiserablyintheirglasscells,withoutsucceedinginpiercingtheirpartitionsrightthrough。
　　Moreover,instincthasbeenlessgeneroustothemthantothefemales。Theircorpses,interspersedhereandthereintheseriesofthecells,aredisturbingcauses,whichitiswisetoeliminate。I
　　thereforechoosethelarger,morepowerful—lookingcocoons。These,exceptforanoccasionalunavoidableerror,belongtofemales。Ipackthemintubes,sometimesvaryingtheirpositionineveryway,sometimesgivingthemallalikearrangement。Itdoesnotmatterwhetherthewholeseriescomesfromoneandthesamebramble—stumporfromseveral:wearefreetochoosewhereweplease;theresultwillnotbealtered。
　　ThefirsttimethatIpreparedoneofthesehorizontaltubesopenatbothends,Iwasgreatlystruckbywhathappened。Theseriesconsistedoftencocoons。Itwasdividedintotwoequalbatches。Thefiveontheleftwentoutontheleft,thefiveontherightwentoutontheright,reversing,whennecessary,theiroriginaldirectioninthecell。Itwasveryremarkablefromthepointofviewofsymmetry;
　　moreover,itwasaveryunlikelyarrangementamongthetotalnumberofpossiblearrangements,asmathematicswillshowus。
　　LetustakentorepresentthenumberofOsmiae。Eachofthem,oncegravityceasestointerfereandleavestheinsectindifferenttoeitherendofthetube,iscapableoftwopositions,accordingasshechoosestheexitontherightorontheleft。WitheachofthetwopositionsofthisfirstOsmiacanbecombinedeachofthetwopositionsofthesecond,givingus,inall,2x2=（2squared）
　　arrangements。Eachofthese（2squared）arrangementscanbecombined,initsturn,witheachofthetwopositionsofthethirdOsmia。Wethusobtain2x2x2=（2cubed）arrangementswiththreeOsmiae;andsoon,eachadditionalinsectmultiplyingthepreviousresultbythefactor2。WithnOsmiae,therefore,thetotalnumberofarrangementsis（2tothepowern。）
　　Butnotethatthesearrangementsaresymmetrical,twobytwo:agivenarrangementtowardstherightcorrespondswithasimilararrangementtowardstheleft;andthissymmetryimpliesequality,for,intheprobleminhand,itisamatterofindifferencewhetherafixedarrangementcorrespondwiththerightorleftofthetube。Thepreviousnumber,therefore,mustbedividedby2。Thus,nOsmiae,accordingaseachofthemturnsherheadtotherightorleftinmyhorizontaltube,areabletoadopt（2tothepowern—1）
　　arrangements。Ifn=10,asinmyfirstexperiment,thenumberofarrangementsbecomes（2tothepower9）=512。
　　Consequently,outof512wayswhichmyteninsectscanadoptfortheiroutgoingposition,thereresultedoneofthoseinwhichthesymmetrywasmoststriking。Andobservethatthiswasnotaneffectobtainedbyrepeatedattempts,byhaphazardexperiments。EachOsmiainthelefthalfhadboredtotheleft,withouttouchingthepartitionontheright;eachOsmiaintherighthalfhadboredtotheright,withouttouchingthepartitionontheleft。Theshapeoftheorificesandthesurfaceconditionofthepartitionshowedthis,ifproofwerenecessary。Therehadbeenaspontaneousdecision,onehalfinfavouroftheleft,onehalfinfavouroftheright。
　　Thearrangementpresentsanothermerit,onesuperiortothatofsymmetry:ithasthemeritofcorrespondingwiththeminimumexpenditureofforce。Toadmitoftheexitofthewholeseries,ifthestringconsistsofncells,thereareoriginallynpartitionstobeperforated。Theremightevenbeonemore,owingtoacomplicationwhichIdisregard。Thereare,Isay,atleastnpartitionstobeperforated。WhethereachOsmiapiercesherown,orwhetherthesameOsmiapiercesseveral,thusrelievingherneighbours,doesnotmattertous:thesum—totaloftheforceexpendedbythestringofBeeswillbeinproportiontothenumberofthosepartitions,inwhatevermannertheexitbeeffected。
　　Butthereisanothertaskwhichwemusttakeseriouslyintoconsideration,becauseitisoftenmoretroublesomethantheboringofthepartition:Imeantheworkofclearingaroadthroughthewreckage。Letussupposethepartitionspiercedandtheseveralchambersblockedbytheresultingrubbishandbythatrubbishonly,sincethehorizontalpositionprecludesanymixingofthecontentsofdifferentchambers。Toopenapassageforitselfthroughtheserubbish—heaps,eachinsectwillhavethesmallestefforttomakeifitpassesthroughthesmallestpossiblenumberofcells,inshort,ifitmakesfortheopeningnearesttoit。Thesesmallestindividualeffortsamount,intheaggregate,tothesmallesttotaleffort。
　　Therefore,byproceedingastheydidinmyexperiment,theOsmiaeeffecttheirexitwiththeleastexpenditureofenergy。Itiscurioustoseeaninsectapplythe'principleofleastaction,'sooftenpostulatedinmechanics。
　　Anarrangementwhichsatisfiesthisprinciple,whichconformstothelawofsymmetryandwhichpossessesbutonechancein512,iscertainlynofortuitousresult。Itisdeterminedbyacause;and,asthiscauseactsinvariably,thesamearrangementmustbereproducedifIrenewtheexperiment。Irenewedit,therefore,intheyearsthatfollowed,withasmanyappliancesasIcouldfindbramble—stumps;
　　and,ateachnewtest,IsawoncemorewhatIhadseenwithsuchinterestonthefirstoccasion。Ifthenumberbeeven——andmycolumnatthattimeconsistedusuallyoften——onehalfgoesoutontheright,theotherontheleft。Ifthenumberbeodd——eleven,forinstance——theOsmiainthemiddlegoesoutindiscriminatelybytherightorleftexit。Asthenumberofcellstobetraversedisthesameonbothsides,herexpenditureofenergydoesnotvarywiththedirectionoftheexit;andtheprincipleofleastactionisstillobserved。
　　ItwasimportanttodiscoveriftheThree—prongedOsmiasharedhercapacity,inthefirstplace,withtheotherbramble—dwellersand,inthesecond,withBeesdifferentlyhoused,butalsodestinedlaboriouslytocutanewroadforthemselveswhenthehourcomestoquitthenest。Well,apartfromafewirregularities,dueeithertococoonswhoselarvaperishedinmytubesbeforedeveloping,ortothoseinexperiencedworkers,themales,theresultwasthesameinthecaseofAnthidiumscapulare。Theinsectsdividedthemselvesintotwoequalbatches,onegoingtotheright,theothertotheleft。
　　Tripoxylonfigulusleftmeundecided。Thisfeebleinsectisnotcapableofperforatingmypartitions;itnibblesatthemalittle;
　　andIhadtojudgethedirectionfromthemarksofitsmandibles。
　　Thesemarks,whicharenotalwaysveryplain,donotyetallowmetopronounceanopinion。Soleniusvagus,whoisaskilfulborer,behaveddifferentlyfromtheOsmia。Inacolumnoften,thewholeexoduswasmadeinonedirection。
　　Ontheotherhand,ItestedtheMason—beeoftheSheds,who,whenemergingundernaturalconditions,hasonlytopiercehercementceilingandisnotconfrontedwithaseriesofcells。ThoughastrangertotheenvironmentwhichIcreatedforher,shegavemeamostpositiveanswer。Ofacolumnoftenlaidinahorizontaltubeopenatbothends,fivemadetheirwaytotherightandfivetotheleft。Dioxyscincta,aparasiteinthebuildingsofbothspeciesofMason—bees,theChalicodomaoftheShedsandtheChalicodomaoftheWalls（Cf。"TheMason—bees"byJ。HenriFabre,translatedbyAlexanderTeixeiradeMattos:passim。——Translator'sNote。）,providedmewithnopreciseresult。TheLeaf—cuttingBee（Megachileapicalis,SPIN。（Cf。Chapter8ofthepresentvolume。——Translator'sNote。））,whobuildsherleafycupsintheoldcellsoftheChalicodomaoftheWalls,actsliketheSoleniusanddirectsherwholecolumntowardsthesameoutlet。
　　Incompleteasitis,thissymmetryshowsushowunwiseitweretogeneralizefromtheconclusionstowhichtheThree—prongedOsmialeadsus。WhereassomeBees,suchastheAnthidiumandtheChalicodoma,sharetheOsmia'stalentforusingthetwofoldexit,others,suchastheSoleniusandtheLeaf—cutter,behavelikeaflockofsheepandfollowthefirstthatgoesout。Theentomologicalworldisnotallofapiece;itsgiftsareveryvarious:whatoneiscapableofdoinganothercannotdo;andpenetratingindeedwouldbetheeyesthatsawthecausesofthesedifferences。Bethisasitmay,increasedresearchwillcertainlyshowusalargernumberofspeciesqualifiedtousethedoubleoutlet。Forthemoment,weknowthree;
　　andthatisenoughforourpurpose。
　　Iwilladdthat,whenthehorizontaltubehasoneofitsendsclosed,thewholestringofOsmiaemakesfortheopenend,turningroundtodoso,ifneedbe。
　　Nowthatthefactsaresetforth,letus,ifpossible,tracethecause。Inahorizontaltube,gravitynolongeractstodeterminethedirectionwhichtheinsectwilltake。Isittoattackthepartitionontherightorthatontheleft？Howshallitdecide？ThemoreI
　　lookintothematter,themoredomysuspicionsfallupontheatmosphericinfluencewhichisfeltthroughthetwoopenends。Ofwhatdoesthisinfluenceconsist？Isitaneffectofpressure,ofhygrometry,ofelectricalconditions,ofpropertiesthatescapeourcoarserphysicalattunement？Hewereaboldmanwhoshouldundertaketodecide。Arenotweourselves,whentheweatherisabouttoalter,subjecttosubtleimpressions,tosensationswhichweareunabletoexplain？Andyetthisvaguesensitivenesstoatmosphericchangeswouldnotbeofmuchhelptousincircumstancessimilartothoseofmyanchorites。Imagineourselvesinthedarknessandthesilenceofaprison—cell,precededandfollowedbyothersimilarcells。Wepossessimplementswherewithtopiercethewalls;butwherearewetostriketoreachthefinaloutletandtoreachitwiththeleastdelay？
　　Atmosphericinfluencewouldcertainlyneverguideus。
　　Andyetitguidestheinsect。Feeblethoughitbe,throughthemultiplicityofpartitions,itisexercisedononesidemorethanontheother,becausetheobstaclesarefewer;andtheinsect,sensibletothedifferencebetweenthosetwouncertainties,unhesitatinglyattacksthepartitionwhichisnearertotheopenair。Thusisdecidedthedivisionofthecolumnintotwoconversesections,whichaccomplishthetotalliberationwiththeleastaggregateofwork。Inshort,theOsmiaandherrivals'feel'thefreespace。Thisisyetonemoresensoryfacultywhichevolutionmightwellhaveleftus,forourgreateradvantage。Asithasnotdoneso,arewethenreally,asmanycontend,thehighestexpressionoftheprogressaccomplished,throughouttheages,bythefirstatomofglairexpandedintoacell？
　　CHAPTER2。THEOSMIAE。
　　Februaryhasitssunnydays,heraldingspring,towhichrudewinterwillreluctantlyyieldplace。Insnugcorners,amongtherocks,thegreatspurgeofourdistrict,thecharaciasoftheGreeks,thejusclooftheProvencals,beginstoliftitsdroopinginflorescenceanddiscreetlyopensafewsombreflowers。HerethefirstMidgesoftheyearwillcometoslaketheirthirst。Bythetimethatthetipofthestalksreachestheperpendicular,theworstofthecoldweatherwillbeover。
　　Anothereagerone,thealmond—tree,riskingthelossofitsfruit,hastenstoechothesepreludestothefestivalofthesun,preludeswhicharetoooftentreacherous。Afewdaysofsoftskiesanditbecomesagloriousdomeofwhiteflowers,eachtwinklingwitharoseateeye。Thecountry,whichstilllacksgreen,seemsdottedeverywherewithwhite—satinpavilions。'Twouldbeacallousheartindeedthatcouldresistthemagicofthisawakening。