8
　　Wehavealreadyshownthatwetanddrymustbothgiverisetoanevaporation：earthquakesareanecessaryconsequenceofthisfact。Theearthisessentiallydry，butrainfillsitwithmoisture。Thenthesunanditsownfirewarmitandgiverisetoaquantityofwindbothoutsideandinsideit。Thiswindsometimesflowsoutwardsinasinglebody，sometimesinwards，andsometimesitisdivided。Allthesearenecessarylaws。Nextwemustfindoutwhatbodyhasthegreatestmotiveforce。Thiswillcertainlybethebodythatnaturallymovesfarthestandismostviolent。Nowthatwhichhasthemostrapidmotionisnecessarilythemostviolent;foritsswiftnessgivesitsimpactthegreatestforce。Again，therarestbody，thatwhichcanmostreadilypassthrougheveryotherbody，isthatwhichnaturallymovesfarthest。Windsatisfiestheseconditionsinthehighestdegree（fireonlybecomesflameandmovesrapidlywhenwindaccompaniesit）：sothatnotwaternorearthisthecauseofearthquakesbutwind-thatis，theinrushoftheexternalevaporationintotheearth。
　　Hence，sincetheevaporationgenerallyfollowsinacontinuousbodyinthedirectioninwhichitfirststarted，andeitherallofitflowsinwardsoralloutwards，mostearthquakesandthegreatestareaccompaniedbycalm。Itistruethatsometakeplacewhenawindisblowing，butthispresentsnodifficulty。Wesometimesfindseveralwindsblowingsimultaneously。Ifoneoftheseenterstheearthwegetanearthquakeattendedbywind。Onlytheseearthquakesarelessseverebecausetheirsourceandcauseisdivided。
　　Again，mostearthquakesandtheseverestoccuratnightor，ifbyday，aboutnoon，thatbeinggenerallythecalmestpartoftheday。Forwhenthesunexertsitsfullpower（asitdoesaboutnoon）itshutstheevaporationintotheearth。Night，too，iscalmerthanday。Theabsenceofthesunmakestheevaporationreturnintotheearthlikeasortofebbtide，correspondingtotheoutwardflow;especiallytowardsdawn，forthewinds，asarule，begintoblowthen，andiftheirsourcechangesaboutliketheEuripusandflowsinwardsthequantityofwindintheearthisgreaterandamoreviolentearthquakeresults。
　　Theseverestearthquakestakeplacewheretheseaisfullofcurrentsortheearthspongyandcavernous：sotheyoccurneartheHellespontandinAchaeaandSicily，andthosepartsofEuboeawhichcorrespondtoourdescription-wheretheseaissupposedtoflowinchannelsbelowtheearth。Thehotsprings，too，nearAedepsusareduetoacauseofthiskind。Itistheconfinedcharacteroftheseplacesthatmakesthemsoliabletoearthquakes。Agreatandthereforeviolentwindisdeveloped，whichwouldnaturallyblowawayfromtheearth：buttheonrushoftheseainagreatmassthrustsitbackintotheearth。Thecountriesthatarespongybelowthesurfaceareexposedtoearthquakesbecausetheyhaveroomforsomuchwind。
　　Forthesamereasonearthquakesusuallytakeplaceinspringandautumnandintimesofwetandofdrought-becausethesearethewindiestseasons。Summerwithitsheatandwinterwithitsfrostcausecalm：winteristoocold，summertoodryforwindstoform。Intimeofdroughttheairisfullofwind;droughtisjustthepredominanceofthedryoverthemoistevaporation。Again，excessiveraincausesmoreoftheevaporationtoformintheearth。Thenthissecretionisshutupinanarrowcompassandforcedintoasmallerspacebythewaterthatfillsthecavities。Thusagreatwindiscompressedintoasmallerspaceandsogetstheupperhand，andthenbreaksoutandbeatsagainsttheearthandshakesitviolently。
　　Wemustsupposetheactionofthewindintheearthtobeanalogoustothetremorsandthrobbingscausedinusbytheforceofthewindcontainedinourbodies。Thussomeearthquakesareasortoftremor，othersasortofthrobbing。Again，wemustthinkofanearthquakeassomethinglikethetremorthatoftenrunsthroughthebodyafterpassingwaterasthewindreturnsinwardsfromwithoutinonevolume。
　　Theforcewindcanhavemaybegatherednotonlyfromwhathappensintheair（whereonemightsupposethatitoweditspowertoproducesucheffectstoitsvolume），butalsofromwhatisobservedinanimalbodies。Tetanusandspasmsaremotionsofwind，andtheirforceissuchthattheunitedeffortsofmanymendonotsucceedinovercomingthemovementsofthepatients。Wemustsuppose，then（tocomparegreatthingswithsmall），thatwhathappensintheearthisjustlikethat。Ourtheoryhasbeenverifiedbyactualobservationinmanyplaces。Ithasbeenknowntohappenthatanearthquakehascontinueduntilthewindthatcauseditburstthroughtheearthintotheairandappearedvisiblylikeahurricane。ThishappenedlatelynearHeracleiainPontusandsometimepastattheislandHiera，oneofthegroupcalledtheAeolianislands。Hereaportionoftheearthswelledupandalumplikeamoundrosewithanoise：finallyitburst，andagreatwindcameoutofitandthrewuplivecindersandasheswhichburiedtheneighbouringtownofLiparaandreachedsomeofthetownsinItaly。Thespotwherethiseruptionoccurredisstilltobeseen。
　　Indeed，thismustberecognizedasthecauseofthefirethatisgeneratedintheearth：theairisfirstbrokenupinsmallparticlesandthenthewindisbeatenaboutandsocatchesfire。
　　Aphenomenonintheseislandsaffordsfurtherevidenceofthefactthatwindsmovebelowthesurfaceoftheearth。Whenasouthwindisgoingtoblowthereisapremonitoryindication：asoundisheardintheplacesfromwhichtheeruptionsissue。Thisisbecausetheseaisbeingpushedonfromadistanceanditsadvancethrustsbackintotheearththewindthatwasissuingfromit。Thereasonwhythereisanoiseandnoearthquakeisthattheundergroundspacesaresoextensiveinproportiontothequantityoftheairthatisbeingdrivenonthatthewindslipsawayintothevoidbeyond。
　　Again，ourtheoryissupportedbythefactsthatthesunappearshazyandisdarkenedintheabsenceofclouds，andthatthereissometimescalmandsharpfrostbeforeearthquakesatsunrise。Thesunisnecessarilyobscuredanddarkenedwhentheevaporationwhichdissolvesandrarefiestheairbeginstowithdrawintotheearth。
　　Thecalm，too，andthecoldtowardssunriseanddawnfollowfromthetheory。Thecalmwehavealreadyexplained。Theremustasarulebecalmbecausethewindflowsbackintotheearth：again，itmustbemostmarkedbeforethemoreviolentearthquakes，forwhenthewindisnotpartoutsideearth，partinside，butmovesinasinglebody，itsstrengthmustbegreater。Thecoldcomesbecausetheevaporationwhichisnaturallyandessentiallyhotenterstheearth。（Windisnotrecognizedtobehot，becauseitsetstheairinmotion，andthatisfullofaquantityofcoldvapour。Itisthesamewiththebreathweblowfromourmouth：closebyitiswarm，asitiswhenwebreatheoutthroughthemouth，butthereissolittleofitthatitisscarcelynoticed，whereasatadistanceitiscoldforthesamereasonaswind。）Well，whenthisevaporationdisappearsintotheearththevaporousexhalationconcentratesandcausescoldinanyplaceinwhichthisdisappearanceoccurs。
　　Asignwhichsometimesprecedesearthquakescanbeexplainedinthesameway。Eitherbydayoralittleaftersunset，infineweather，alittle，light，long-drawncloudisseen，likealongverystraightline。Thisisbecausethewindisleavingtheairanddyingdown。
　　Somethinganalogoustothishappensonthesea-shore。Whentheseabreaksingreatwavesthemarksleftonthesandareverythickandcrooked，butwhentheseaiscalmtheyareslightandstraight（becausethesecretionissmall）。Astheseaistotheshoresothewindistothecloudyair;so，whenthewinddrops，thisverystraightandthincloudisleft，asortofwave-markintheair。
　　Anearthquakesometimescoincideswithaneclipseofthemoonforthesamereason。Whentheearthisonthepointofbeinginterposed，butthelightandheatofthesunhasnotquitevanishedfromtheairbutisdyingaway，thewindwhichcausestheearthquakebeforetheeclipse，turnsoffintotheearth，andcalmensues。Forthereoftenarewindsbeforeeclipses：atnightfalliftheeclipseisatmidnight，andatmidnightiftheeclipseisatdawn。Theyarecausedbythelesseningofthewarmthfromthemoonwhenitssphereapproachesthepointatwhichtheeclipseisgoingtotakeplace。Sotheinfluencewhichrestrainedandquietedtheairweakensandtheairmovesagainandawindrises，anddoessolater，thelatertheeclipse。
　　Asevereearthquakedoesnotstopatonceorafterasingleshock，butfirsttheshocksgoon，oftenforaboutfortydays;afterthat，foroneoreventwoyearsitgivespremonitoryindicationsinthesameplace。Theseverityoftheearthquakeisdeterminedbythequantityofwindandtheshapeofthepassagesthroughwhichitflows。Whereitisbeatenbackandcannoteasilyfinditswayouttheshocksaremostviolent，andthereitmustremaininacrampedspacelikewaterthatcannotescape。Anythrobbinginthebodydoesnotceasesuddenlyorquickly，butbydegreesaccordingastheaffectionpassesoff。Soheretheagencywhichcreatedtheevaporationandgaveitanimpulsetomotionclearlydoesnotatonceexhaustthewholeofthematerialfromwhichitformsthewindwhichwecallanearthquake。Sountiltherestofthisisexhaustedtheshocksmustcontinue，thoughmoregently，andtheymustgoonuntilthereistoolittleoftheevaporationlefttohaveanyperceptibleeffectontheearthatall。
　　Subterraneannoises，too，areduetothewind;sometimestheyportendearthquakesbutsometimestheyhavebeenheardwithoutanyearthquakefollowing。Justastheairgivesoffvarioussoundswhenitisstruck，soitdoeswhenitstrikesotherthings;forstrikinginvolvesbeingstruckandsothetwocasesarethesame。Thesoundprecedestheshockbecausesoundisthinnerandpassesthroughthingsmorereadilythanwind。Butwhenthewindistooweakbyreasonofthinnesstocauseanearthquaketheabsenceofashockisduetoitsfilteringthroughreadily，thoughbystrikinghardandhollowmassesofdifferentshapesitmakesvariousnoises，sothattheearthsometimesseemsto’bellow’astheportentmongerssay。
　　Waterhasbeenknowntoburstoutduringanearthquake。Butthatdoesnotmakewaterthecauseoftheearthquake。Thewindistheefficientcausewhetheritdrivesthewateralongthesurfaceorupfrombelow：justaswindsarethecausesofwavesandnotwavesofwinds。Elsewemightaswellsaythatearthwasthecause;foritisupsetinanearthquake，justlikewater（foreffusionisaformofupsetting）。No，earthandwaterarematerialcauses（beingpatients，notagents）：thetruecauseisthewind。
　　Thecombinationofatidalwavewithanearthquakeisduetothepresenceofcontrarywinds。Itoccurswhenthewindwhichisshakingtheearthdoesnotentirelysucceedindrivingofftheseawhichanotherwindisbringingon，butpushesitbackandheapsitupinagreatmassinoneplace。Giventhissituationitfollowsthatwhenthiswindgiveswaythewholebodyofthesea，drivenonbytheotherwind，willburstoutandoverwhelmtheland。ThisiswhathappenedinAchaea。Thereasouthwindwasblowing，butoutsideanorthwind;thentherewasacalmandthewindenteredtheearth，andthenthetidalwavecameonandsimultaneouslytherewasanearthquake。Thiswasthemoreviolentastheseaallowednoexittothewindthathadenteredtheearth，butshutitin。Sointheirstrugglewithoneanotherthewindcausedtheearthquake，andthewavebyitssettlingdowntheinundation。
　　Earthquakesarelocalandoftenaffectasmalldistrictonly;
　　whereaswindsarenotlocal。Suchphenomenaarelocalwhentheevaporationsatagivenplacearejoinedbythosefromthenextandunite;this，asweexplained，iswhathappenswhenthereisdroughtorexcessiverainlocally。Nowearthquakesdocomeaboutinthiswaybutwindsdonot。Forearthquakes，rains，anddroughtshavetheirsourceandorigininsidetheearth，sothatthesunisnotequallyabletodirectalltheevaporationsinonedirection。Butontheevaporationsintheairthesunhasmoreinfluencesothat，whenoncetheyhavebeengivenanimpulsebyitsmotion，whichisdeterminedbyitsvariouspositions，theyflowinonedirection。
　　Whenthewindispresentinsufficientquantitythereisanearthquake。Theshocksarehorizontallikeatremor;exceptoccasionally，inafewplaces，wheretheyactvertically，upwardsfrombelow，likeathrobbing。Itistheverticaldirectionwhichmakesthiskindofearthquakesorare。Themotiveforcedoesnoteasilyaccumulateingreatquantityinthepositionrequired，sincethesurfaceoftheearthsecretesfarmoreoftheevaporationthanitsdepths。Whereveranearthquakeofthiskinddoesoccuraquantityofstonescomestothesurfaceoftheearth（aswhenyouthrowupthingsinawinnowingfan），asweseefromSipylusandthePhlegraeanplainandthedistrictinLiguria，whichweredevastatedbythiskindofearthquake。
　　Islandsinthemiddleoftheseaarelessexposedtoearthquakesthanthosenearland。First，thevolumeoftheseacoolstheevaporationsandoverpowersthembyitsweightandsocrushesthem。
　　Then，currentsandnotshocksareproducedintheseabytheactionofthewinds。Again，itissoextensivethatevaporationsdonotcollectinitbutissuefromit，andthesedrawtheevaporationsfromtheearthafterthem。Islandsnearthecontinentreallyformpartofit：theinterveningseaisnotenoughtomakeanydifference;butthoseintheopenseacanonlybeshakenifthewholeoftheseathatsurroundsthemisshakentoo。
　　Wehavenowexplainedearthquakes，theirnatureandcause，andthemostimportantofthecircumstancesattendantontheirappearance。
　　9
　　Letusgoontoexplainlightningandthunder，andfurtherwhirlwind，fire-wind，andthunderbolts：forthecauseofthemallisthesame。
　　Aswehavesaid，therearetwokindsofexhalation，moistanddry，andtheatmospherecontainsthembothpotentially。It，aswehavesaidbefore，condensesintocloud，andthedensityofthecloudsishighestattheirupperlimit。（Fortheymustbedenserandcolderonthesidewheretheheatescapestotheupperregionandleavesthem。
　　Thisexplainswhyhurricanesandthunderboltsandallanalogousphenomenamovedownwardsinspiteofthefactthateverythinghothasanaturaltendencyupwards。Justasthepipsthatwesqueezebetweenourfingersareheavybutoftenjumpupwards：sothesethingsarenecessarilysqueezedoutawayfromthedensestpartofthecloud。）Nowtheheatthatescapesdispersestotheupregion。
　　Butifanyofthedryexhalationiscaughtintheprocessastheaircools，itissqueezedoutasthecloudscontract，andcollidesinitsrapidcoursewiththeneighbouringclouds，andthesoundofthiscollisioniswhatwecallthunder。Thiscollisionisanalogous，tocomparesmallwithgreat，tothesoundwehearinaflamewhichmencallthelaughterorthethreatofHephaestusorofHestia。Thisoccurswhenthewooddriesandcracksandtheexhalationrushesontheflameinabody。Sointheclouds，theexhalationisprojectedanditsimpactondensecloudscausesthunder：thevarietyofthesoundisduetotheirregularityofthecloudsandthehollowsthatintervenewheretheirdensityisinterrupted。Thisthen，isthunder，andthisitscause。
　　Itusuallyhappensthattheexhalationthatisejectedisinflamedandburnswithathinandfaintfire：thisiswhatwecalllightning，whereweseeasitweretheexhalationcolouredintheactofitsejection。Itcomesintoexistenceafterthecollisionandthethunder，thoughweseeitearlierbecausesightisquickerthanhearing。Therowingoftriremesillustratesthis：theoarsaregoingbackagainbeforethesoundoftheirstrikingthewaterreachesus。
　　However，therearesomewhomaintainthatthereisactuallyfireintheclouds。Empedoclessaysthatitconsistsofsomeofthesun’srayswhichareintercepted：Anaxagorasthatitispartoftheupperether（whichhecallsfire）whichhasdescendedfromabove。Lightning，then，isthegleamofthisfire，andthunderthehissingnoiseofitsextinctioninthecloud。
　　Butthisinvolvestheviewthatlightningactuallyispriortothunderanddoesnotmerelyappeartobeso。Again，thisinterceptingofthefireisimpossibleoneithertheory，butespeciallyitissaidtobedrawndownfromtheupperether。Somereasonoughttobegivenwhythatwhichnaturallyascendsshoulddescend，andwhyitshouldnotalwaysdoso，butonlywhenitiscloudy。Whentheskyisclearthereisnolightning：tosaythatthereis，isaltogetherwanton。
　　Theviewthattheheatofthesun’sraysinterceptedinthecloudsisthecauseofthesephenomenaisequallyunattractive：this，too，isamostcarelessexplanation。Thunder，lightning，andtherestmusthaveaseparateanddeterminatecauseassignedtothemonwhichtheyensue。Butthistheorydoesnothingofthesort。Itislikesupposingthatwater，snow，andhailexistedallalongandwereproducedwhenthetimecameandnotgeneratedatall，asiftheatmospherebroughteachtohandoutofitsstockfromtimetotime。
　　Theyareconcretionsinthesamewayasthunderandlightningarediscretions，sothatifitistrueofeitherthattheyarenotgeneratedbutpre-exist，thesamemustbetrueoftheother。Again，howcananydistinctionbemadeabouttheinterceptingbetweenthiscaseandthatofinterceptionindensersubstancessuchaswater？
　　Water，too，isheatedbythesunandbyfire：yetwhenitcontractsagainandgrowscoldandfreezesnosuchejectionastheydescribeoccurs，thoughitoughtontheirthe。totakeplaceonaproportionatescale。Boilingisduetotheexhalationgeneratedbyfire：butitisimpossibleforittoexistinthewaterbeforehand;andbesidestheycallthenoise’hissing’，not’boiling’。Buthissingisreallyboilingonasmallscale：forwhenthatwhichisbroughtintocontactwithmoistureandisinprocessofbeingextinguishedgetsthebetterofit，thenitboilsandmakesthenoiseinquestion。Some-Cleidemusisoneofthem-saythatlightningisnothingobjectivebutmerelyanappearance。Theycompareittowhathappenswhenyoustriketheseawitharodbynightandthewaterisseentoshine。Theysaythatthemoistureinthecloudisbeatenaboutinthesameway，andthatlightningistheappearanceofbrightnessthatensues。
　　Thistheoryisduetoignoranceofthetheoryofreflection，whichistherealcauseofthatphenomenon。Thewaterappearstoshinewhenstruckbecauseoursightisreflectedfromittosomebrightobject：hencethephenomenonoccursmainlybynight：theappearanceisnotseenbydaybecausethedaylightistooin，tenseandobscuresit。
　　Thesearethetheoriesofothersaboutthunderandlightning：somemaintainingthatlightningisareflection，theothersthatlightningisfireshiningthroughthecloudandthunderitsextinction，thefirenotbeinggeneratedineachcasebutexistingbeforehand。Wesaythatthesamestuffiswindontheearth，andearthquakeunderit，andinthecloudsthunder。Theessentialconstituentofallthesephenomenaisthesame：namely，thedryexhalation。Ifitflowsinonedirectionitiswind，inanotheritcausesearthquakes;intheclouds，whentheyareinaprocessofchangeandcontractandcondenseintowater，itisejectedandcausesthunderandlightningandtheotherphenomenaofthesamenature。
　　Somuchforthunderandlightning。
　　BookIII
　　1
　　LETusexplaintheremainingoperationsofthissecretioninthesamewayaswehavetreatedtherest。Whenthisexhalationissecretedinsmallandscatteredquantitiesandfrequently，andistransitory，anditsconstitutionrare，itgivesrisetothunderandlightning。Butifitissecretedinabodyandisdenser，thatis，lessrare，wegetahurricane。Thefactthatitissuesinbodyexplainsitsviolence：itisduetotherapidityofthesecretion。Nowwhenthissecretionissuesinagreatandcontinuouscurrenttheresultcorrespondstowhatwegetwhentheoppositedevelopmenttakesplaceandrainandaquantityofwaterareproduced。Asfarasthematterfromwhichtheyaredevelopedgoesbothsetsofphenomenaarethesame。Assoonasastimulustothedevelopmentofeitherpotentialityappears，thatofwhichthereisthegreaterquantitypresentinthecloudisatoncesecretedfromit，andthereresultseitherrain，or，iftheotherexhalationprevails，ahurricane。
　　Sometimestheexhalationinthecloud，whenitisbeingsecreted，collideswithanotherundercircumstanceslikethosefoundwhenawindisforcedfromanopenintoanarrowspaceinagatewayoraroad。
　　Itoftenhappensinsuchcasesthatthefirstpartofthemovingbodyisdeflectedbecauseoftheresistancedueeithertothenarrownessortoacontrarycurrent，andsothewindformsacircleandeddy。Itispreventedfromadvancinginastraightline：atthesametimeitispushedonfrombehind;soitiscompelledtomovesidewaysinthedirectionofleastresistance。Thesamethinghappenstothenextpart，andthenext，andsoon，tilltheseriesbecomesone，thatis，tillacircleisformed：forifafigureisdescribedbyasinglemotionthatfiguremustitselfbeone。Thisishoweddiesaregeneratedontheearth，andthecaseisthesameinthecloudsasfarasthebeginningofthemgoes。Onlyhere（asinthecaseofthehurricanewhichshakesoffthecloudwithoutcessationandbecomesacontinuouswind）thecloudfollowstheexhalationunbroken，andtheexhalation，failingtobreakawayfromthecloudbecauseofitsdensity，firstmovesinacircleforthereasongivenandthendescends，becausecloudsarealwaysdensestonthesidewheretheheatescapes。Thisphenomenoniscalledawhirlwindwhenitiscolourless;anditisasortofundigestedhurricane。Thereisneverawhirlwindwhentheweatherisnortherly，norahurricanewhenthereissnow。Thereasonisthatallthesephenomenaare’wind’，andwindisadryandwarmevaporation。Nowfrostandcoldprevailoverthisprincipleandquenchitatitsbirth：thattheydoprevailisclearortherecouldbenosnowornortherlyrain，sincetheseoccurwhenthecolddoesprevail。
　　Sothewhirlwindoriginatesinthefailureofanincipienthurricanetoescapefromitscloud：itisduetotheresistancewhichgeneratestheeddy，anditconsistsinthespiralwhichdescendstotheearthanddragswithitthecloudwhichitcannotshakeoff。Itmovesthingsbyitswindinthedirectioninwhichitisblowinginastraightline，andwhirlsroundbyitscircularmotionandforciblysnatchesupwhateveritmeets。
　　Whenthecloudburnsasitisdrawndownwards，thatis，whentheexhalationbecomesrarer，itiscalledafire-wind，foritsfirecolourstheneighbouringairandinflamesit。
　　Whenthereisagreatquantityofexhalationanditisrareandissqueezedoutintheclouditselfwegetathunderbolt。Iftheexhalationisexceedinglyrarethisrarenesspreventsthethunderboltfromscorchingandthepoetscallit’bright’：iftherarenessislessitdoesscorchandtheycallit’smoky’。Theformermovesrapidlybecauseofitsrareness，andbecauseofitsrapiditypassesthroughanobjectbeforesettingfiretoitordwellingonitsoastoblackenit：thesloweronedoesblackentheobject，butpassesthroughitbeforeitcanactuallyburnit。Further，resistingsubstancesareaffected，unresistingonesarenot。Forinstance，ithashappenedthatthebronzeofashieldhasbeenmeltedwhilethewoodworkremainedintactbecauseitstexturewassoloosethattheexhalationfilteredthroughwithoutaffectingit。Soithaspassedthroughclothes，too，withoutburningthem，andhasmerelyreducedthemtoshreds。
　　Suchevidenceisenoughbyitselftoshowthattheexhalationisatworkinallthesecases，butwesometimesgetdirectevidenceaswell，asinthecaseoftheconflagrationofthetempleatEphesuswhichwelatelywitnessed。Thereindependentsheetsofflameleftthemainfireandwerecarriedbodilyinmanydirections。Nowthatsmokeisexhalationandthatsmokeburnsiscertain，andhasbeenstatedinanotherplacebefore;butwhentheflamemovesbodily，thenwehaveocularproofthatsmokeisexhalation。Onthisoccasionwhatisseeninsmallfiresappearedonamuchlargerscalebecauseofthequantityofmatterthatwasburning。Thebeamswhichwerethesourceoftheexhalationsplit，andaquantityofitrushedinabodyfromtheplacefromwhichitissuedforthandwentupinablaze：
　　sothattheflamewasactuallyseenmovingthroughtheairawayandfallingonthehouses。Forwemustrecognizethatexhalationaccompaniesandprecedesthunderboltsthoughitiscolourlessandsoinvisible。Hence，wherethethunderboltisgoingtostrike，theobjectmovesbeforeitisstruck，showingthattheexhalationleadsthewayandfallsontheobjectfirst。Thunder，too，splitsthingsnotbyitsnoisebutbecausetheexhalationthatstrikestheobjectandthatwhichmakesthenoiseareejectedsimultaneously。Thisexhalationsplitsthethingitstrikesbutdoesnotscorchitatall。
　　Wehavenowexplainedthunderandlightningandhurricane，andfurtherfirewinds，whirlwinds，andthunderbolts，andshownthattheyareallofthemformsofthesamethingandwhereintheyalldiffer。
　　2
　　Letusnowexplainthenatureandcauseofhalo，rainbow，mocksuns，androds，sincethesameaccountappliestothemall。
　　Wemustfirstdescribethephenomenaandthecircumstancesinwhicheachofthemoccurs。Thehalooftenappearsasacompletecircle：itisseenroundthesunandthemoonandbrightstars，bynightaswellasbyday，andatmiddayorintheafternoon，morerarelyaboutsunriseorsunset。
　　Therainbowneverformsafullcircle，noranysegmentgreaterthanasemicircle。Atsunsetandsunrisethecircleissmallestandthesegmentlargest：asthesunriseshigherthecircleislargerandthesegmentsmaller。Aftertheautumnequinoxintheshorterdaysitisseenateveryhouroftheday，inthesummernotaboutmidday。Therearenevermorethantworainbowsatonetime。Eachofthemisthree-coloured;thecoloursarethesameinbothandtheirnumberisthesame，butintheouterrainbowtheyarefainterandtheirpositionisreversed。Intheinnerrainbowthefirstandlargestbandisred;intheouterrainbowthebandthatisnearesttothisoneandsmallestisofthesamecolour：theotherbandscorrespondonthesameprinciple。Thesearealmosttheonlycolourswhichpainterscannotmanufacture：fortherearecolourswhichtheycreatebymixing，butnomixingwillgivered，green，orpurple。Thesearethecoloursoftherainbow，thoughbetweentheredandthegreenanorangecolourisoftenseen。
　　Mocksunsandrodsarealwaysseenbythesideofthesun，notaboveorbelowitnorintheoppositequarterofthesky。Theyarenotseenatnightbutalwaysintheneighbourhoodofthesun，eitherasitisrisingorsettingbutmorecommonlytowardssunset。Theyhavescarcelyeverappearedwhenthesunwasonthemeridian，thoughthisoncehappenedinBosporuswheretwomocksunsrosewiththesunandfolloweditallthroughthedaytillsunset。
　　Thesearethefactsabouteachofthesephenomena：thecauseofthemallisthesame，fortheyareallreflections。Buttheyaredifferentvarieties，andaredistinguishedbythesurfacefromwhichandthewayinwhichthereflectiontothesunorsomeotherbrightobjecttakesplace。
　　Therainbowisseenbyday，anditwasformerlythoughtthatitneverappearedbynightasamoonrainbow。Thisopinionwasduetotherarityoftheoccurrence：itwasnotobserved，forthoughitdoeshappenitdoessorarely。Thereasonisthatthecoloursarenotsoeasytoseeinthedarkandthatmanyotherconditionsmustcoincide，andallthatinasingledayinthemonth。Forifthereistobeoneitmustbeatfullmoon，andthenasthemooniseitherrisingorsetting。Sowehaveonlymetwithtwoinstancesofamoonrainbowinmorethanfiftyyears。
　　Wemustacceptfromthetheoryofopticsthefactthatsightisreflectedfromairandanyobjectwithasmoothsurfacejustasitisfromwater;alsothatinsomemirrorstheformsofthingsarereflected，inothersonlytheircolours。Ofthelatterkindarethosemirrorswhicharesosmallastobeindivisibleforsense。Itisimpossiblethatthefigureofathingshouldbereflectedinthem，forifitisthemirrorwillbesensiblydivisiblesincedivisibilityisinvolvedinthenotionoffigure。Butsincesomethingmustbereflectedinthemandfigurecannotbe，itremainsthatcolouraloneshouldbereflected。Thecolourofabrightobjectsometimesappearsbrightinthereflection，butitsometimes，eitherowingtotheadmixtureofthecolourofthemirrorortoweaknessofsight，givesrisetotheappearanceofanothercolour。
　　However，wemustaccepttheaccountwehavegivenofthesethingsinthetheoryofsensation，andtakesomethingsforgrantedwhileweexplainothers。
　　3
　　Letusbeginbyexplainingtheshapeofthehalo;whyitisacircleandwhyitappearsroundthesunorthemoonoroneoftheotherstars：theexplanationbeinginallthesecasesthesame。
　　Sightisreflectedinthiswaywhenairandvapourarecondensedintoacloudandthecondensedmatterisuniformandconsistsofsmallparts。Henceinitselfitisasignofrain，butifitfadesaway，offineweather，ifitisbrokenup，ofwind。Forifitdoesnotfadeawayandisnotbrokenupbutisallowedtoattainitsnormalstate，itisnaturallyasignofrainsinceitshowsthataprocessofcondensationisproceedingwhichmust，whenitiscarriedtoanend，resultinrain。Forthesamereasonthesehaloesarethedarkest。Itisasignofwindwhenitisbrokenupbecauseitsbreakingupisduetoawindwhichexiststherebuthasnotreachedus。Thisviewfindssupportinthefactthatthewindblowsfromthequarterinwhichthemaindivisionappearsinthehalo。Itsfadingawayisasignoffineweatherbecauseiftheairisnotyetinastatetogetthebetteroftheheatitcontainsandproceedtocondenseintowater，thisshowsthatthemoistvapourhasnotyetseparatedfromthedryandfirelikeexhalation：andthisisthecauseoffineweather。
　　Somuchfortheatmosphericconditionsunderwhichthereflectiontakesplace。Thereflectionisfromthemistthatformsroundthesunorthemoon，andthatiswhythehaloisnotseenoppositethesunliketherainbow。
　　Sincethereflectiontakesplaceinthesamewayfromeverypointtheresultisnecessarilyacircleorasegmentofacircle：forifthelinesstartfromthesamepointandendatthesamepointandareequal，thepointswheretheyformananglewillalwayslieonacircle。
　　LetAGBandAZBandADBbelineseachofwhichgoesfromthepointA
　　tothepointBandformsanangle。LetthelinesAG，AZ，ADbeequalandthoseatB，GB，ZB，DBequaltoo。（Seediagram。）
　　DrawthelineAEB。Thenthetrianglesareequal;fortheirbaseAEBisequal。DrawperpendicularstoAEBfromtheangles;GEfromG，ZEfromZ，DEfromD。Thentheseperpendicularsareequal，beinginequaltriangles。Andtheyareallinoneplane，beingallatrightanglestoAEBandmeetingatasinglepointE。SoifyoudrawthelineitwillbeacircleandEitscentre。NowBisthesun，Atheeye，andthecircumferencepassingthroughthepointsGZDthecloudfromwhichthelineofsightisreflectedtothesun。
　　Themirrorsmustbethoughtofascontiguous：eachofthemistoosmalltobevisible，buttheircontiguitymakesthewholemadeupofthemalltoseemone。Thebrightbandisthesun，whichisseenasacircle，appearingsuccessivelyineachofthemirrorsasapointindivisibletosense。Thebandofcloudnexttoitisblack，itscolourbeingintensifiedbycontrastwiththebrightnessofthehalo。Thehaloisformedratherneartheearthbecausethatiscalmer：
　　forwherethereiswinditisclearthatnohalocanmaintainitsposition。
　　Haloesarecommonerroundthemoonbecausethegreaterheatofthesundissolvesthecondensationsoftheairmorerapidly。
　　Haloesareformedroundstarsforthesamereasons，buttheyarenotprognosticinthesamewaybecausethecondensationtheyimplyissoinsignificantastobebarren。
　　4
　　Wehavealreadystatedthattherainbowisareflection：wehavenowtoexplainwhatsortofreflectionitis，todescribeitsvariousconcomitants，andtoassigntheircauses。
　　Sightisreflectedfromallsmoothsurfaces，suchasareairandwateramongothers。Airmustbecondensedifitistoactasamirror，thoughitoftengivesareflectionevenuncondensedwhenthesightisweak。Suchwasthecaseofamanwhosesightwasfaintandindistinct。Healwayssawanimageinfrontofhimandfacinghimashewalked。Thiswasbecausehissightwasreflectedbacktohim。Itsmorbidconditionmadeitsoweakanddelicatethattheairclosebyactedasamirror，justasdistantandcondensedairnormallydoes，andhissightcouldnotpushitback。Sopromontoriesinthesea’loom’whenthereisasouth-eastwind，andeverythingseemsbigger，andinamist，too，thingsseembigger：so，too，thesunandthestarsseembiggerwhenrisingandsettingthanonthemeridian。Butthingsarebestreflectedfromwater，andeveninprocessofformationitisabettermirrorthanair，foreachoftheparticles，theunionofwhichconstitutesaraindrop，isnecessarilyabettermirrorthanmist。Nowitisobviousandhasalreadybeenstatedthatamirrorofthiskindrendersthecolourofanobjectonly，butnotitsshape。
　　Henceitfollowsthatwhenitisonthepointofrainingandtheairinthecloudsisinprocessofformingintoraindropsbuttherainisnotyetactuallythere，ifthesunisopposite，oranyotherobjectbrightenoughtomakethecloudamirrorandcausethesighttobereflectedtotheobjectthenthereflectionmustrenderthecolouroftheobjectwithoutitsshape。Sinceeachofthemirrorsissosmallastobeinvisibleandwhatweseeisthecontinuousmagnitudemadeupofthemall，thereflectionnecessarilygivesusacontinuousmagnitudemadeupofonecolour;eachofthemirrorscontributingthesamecolourtothewhole。Wemaydeducethatsincetheseconditionsarerealizabletherewillbeanappearanceduetoreflectionwheneverthesunandthecloudarerelatedinthewaydescribedandwearebetweenthem。Butthesearejusttheconditionsunderwhichtherainbowappears。Soitisclearthattherainbowisareflectionofsighttothesun。
　　Sotherainbowalwaysappearsoppositethesunwhereasthehaloisroundit。Theyarebothreflections，buttherainbowisdistinguishedbythevarietyofitscolours。Thereflectionintheonecaseisfromwaterwhichisdarkandfromadistance;intheotherfromairwhichisnearerandlighterincolour。Whitelightthroughadarkmediumoronadarksurface（itmakesnodifference）looksred。
　　Weknowhowredtheflameofgreenwoodis：thisisbecausesomuchsmokeismixedwiththebrightwhitefirelight：so，too，thesunappearsredthroughsmokeandmist。Thatiswhyintherainbowreflectiontheoutercircumferenceisred（thereflectionbeingfromsmallparticlesofwater），butnotinthecaseofthehalo。Theothercoloursshallbeexplainedlater。Again，acondensationofthiskindcannotpersistintheneighbourhoodofthesun：itmusteitherturntorainorbedissolved，butoppositetothesunthereisanintervalduringwhichthewaterisformed。Iftherewerenotthisdistinctionhaloeswouldbecolouredliketherainbow。Actuallynocompleteorcircularhalopresentsthiscolour，onlysmallandfragmentaryappearancescalled’rods’。Butifahazeduetowateroranyotherdarksubstanceformedthereweshouldhavehad，aswemaintain，acompleterainbowlikethatwhichwedofindlamps。A
　　rainbowappearsroundtheseinwinter，generallywithsoutherlywinds。
　　Personswhoseeyesaremoistseeitmostclearlybecausetheirsightisweakandeasilyreflected。Itisduetothemoistnessoftheairandthesootwhichtheflamegivesoffandwhichmixeswiththeairandmakesitamirror，andtotheblacknesswhichthatmirrorderivesfromthesmokynatureofthesoot。Thelightofthelampappearsasacirclewhichisnotwhitebutpurple。Itshowsthecoloursoftherainbow;butbecausethesightthatisreflectedistooweakandthemirrortoodark，redisabsent。Therainbowthatisseenwhenoarsareraisedoutoftheseainvolvesthesamerelativepositionsasthatinthesky，butitscolourismorelikethatroundthelamps，beingpurpleratherthanred。Thereflectionisfromverysmallparticlescontinuouswithoneanother，andinthiscasetheparticlesarefullyformedwater。Wegetarainbow，too，ifamansprinklesfinedropsinaroomturnedtothesunsothatthesunisshininginpartoftheroomandthrowingashadowintherest。Thenifonemansprinklesintheroom，another，standingoutside，seesarainbowwherethesun’sraysceaseandmaketheshadow。Itsnatureandcolourislikethatfromtheoarsanditscauseisthesame，forthesprinklinghandcorrespondstotheoar。
　　Thatthecoloursoftherainbowarethosewedescribedandhowtheothercolourscometoappearinitwillbeclearfromthefollowingconsiderations。Wemustrecognize，aswehavesaid，andlaydown：
　　first，thatwhitecolouronablacksurfaceorseenthroughablackmediumgivesred;second，thatsightwhenstrainedtoadistancebecomesweakerandless;third，thatblackisinasortthenegationofsight：anobjectisblackbecausesightfails;soeverythingatadistancelooksblacker，becausesightdoesnotreachit。Thetheoryofthesemattersbelongstotheaccountofthesenses，whicharethepropersubjectsofsuchaninquiry;weneedonlystateaboutthemwhatisnecessaryforus。Atallevents，thatisthereasonwhydistantobjectsandobjectsseeninamirrorlookdarkerandsmallerandsmoother，whythereflectionofcloudsinwaterisdarkerthanthecloudsthemselves。Thislatterisclearlythecase：thereflectiondiminishesthesightthatreachesthem。Itmakesnodifferencewhetherthechangeisintheobjectseenor。inthesight，theresultbeingineithercasethesame。Thefollowingfactfurtherisworthnoticing。
　　Whenthereisacloudnearthesunandwelookatitdoesnotlookcolouredatallbutwhite，butwhenwelookatthesamecloudinwateritshowsatraceofrainbowcolouring。Clearly，then，whensightisreflecteditisweakenedand，asitmakesdarklookdarker，soitmakeswhitelooklesswhite，changingitandbringingitnearertoblack。Whenthesightisrelativelystrongthechangeistored;thenextstageisgreen，andafurtherdegreeofweaknessgivesviolet。Nofurtherchangeisvisible，butthreecompletestheseriesofcolours（aswefindthreedoesinmostotherthings），andthechangeintotherestisimperceptibletosense。Hencealsotherainbowappearswiththreecolours;thisistrueofeachofthetwo，butinacontraryway。Theouterbandoftheprimaryrainbowisred：forthelargestbandreflectsmostsighttothesun，andtheouterbandislargest。
　　Themiddlebandandthethirdgoonthesameprinciple。Soiftheprincipleswelaiddownabouttheappearanceofcoloursaretruetherainbownecessarilyhasthreecolours，andthesethreeandnoothers。Theappearanceofyellowisduetocontrast，fortherediswhitenedbyitsjuxtapositionwithgreen。Wecanseethisfromthefactthattherainbowispurestwhenthecloudisblackest;andthentheredshowsmostyellow。（Yellowintherainbowcomesbetweenredandgreen。）Sothewholeoftheredshowswhitebycontrastwiththeblacknessofthecloudaround：foritiswhitecomparedtothecloudandthegreen。Again，whentherainbowisfadingawayandtheredisdissolving，thewhitecloudisbroughtintocontactwiththegreenandbecomesyellow。Butthemoonrainbowaffordsthebestinstanceofthiscolourcontrast。Itlooksquitewhite：thisisbecauseitappearsonthedarkcloudandatnight。So，justasfireisintensifiedbyaddedfire，blackbesideblackmakesthatwhichisinsomedegreewhitelookquitewhite。Brightdyestooshowtheeffectofcontrast。
　　Inwovenandembroideredstuffstheappearanceofcoloursisprofoundlyaffectedbytheirjuxtapositionwithoneanother（purple，forinstance，appearsdifferentonwhiteandonblackwool），andalsobydifferencesofillumination。Thusembroidererssaythattheyoftenmakemistakesintheircolourswhentheyworkbylamplight，andusethewrongones。
　　Wehavenowshownwhytherainbowhasthreecoloursandthattheseareitsonlycolours。Thesamecauseexplainsthedoublerainbowandthefaintnessofthecoloursintheouteroneandtheirinvertedorder。Whensightisstrainedtoagreatdistancetheappearanceofthedistantobjectisaffectedinacertainway：andthesamethingholdsgoodhere。Sothereflectionfromtheouterrainbowisweakerbecauseittakesplacefromagreaterdistanceandlessofitreachesthesun，andsothecoloursseenarefainter。Theirorderisreversedbecausemorereflectionreachesthesunfromthesmaller，innerband。Forthatreflectionisnearertooursightwhichisreflectedfromthebandwhichisnearesttotheprimaryrainbow。Nowthesmallestbandintheouterrainbowisthatwhichisnearest，andsoitwillbered;andthesecondandthethirdwillfollowthesameprinciple。LetBbetheouterrainbow，Atheinnerone;letRstandfortheredcolour，Gforgreen，Vforviolet;yellowappearsatthepointY。Threerainbowsormorearenotfoundbecauseeventhesecondisfainter，sothatthethirdreflectioncanhavenostrengthwhateverandcannotreachthesunatall。（Seediagram。）
　　5
　　Therainbowcanneverbeacirclenorasegmentofacirclegreaterthanasemicircle。Theconsiderationofthediagramwillprovethisandtheotherpropertiesoftherainbow。（Seediagram。）
　　LetAbeahemisphererestingonthecircleofthehorizon，letitscentrebeKandletHbeanotherpointappearingonthehorizon。
　　Then，ifthelinesthatfallinaconefromKhaveHKastheiraxis，and，KandMbeingjoined，thelinesKMarereflectedfromthehemispheretoHoverthegreaterangle，thelinesfromKwillfallonthecircumferenceofacircle。Ifthereflectiontakesplacewhentheluminousbodyisrisingorsettingthesegmentofthecircleabovetheearthwhichiscutoffbythehorizonwillbeasemi-circle;iftheluminousbodyisabovethehorizonitwillalwaysbelessthanasemicircle，anditwillbesmallestwhentheluminousbodyculminates。
　　FirstlettheluminousbodybeappearingonthehorizonatthepointH，andletKMbereflectedtoH，andlettheplaneinwhichAis，determinedbythetriangleHKM，beproduced。Thenthesectionofthespherewillbeagreatcircle。LetitbeA（foritmakesnodifferencewhichoftheplanespassingthroughthelineHKanddeterminedbythetriangleKMHisproduced）。NowthelinesdrawnfromHandKtoapointonthesemicircleAareinacertainratiotooneanother，andnolinesdrawnfromthesamepointstoanotherpointonthatsemicirclecanhavethesameratio。ForsinceboththepointsHandKandthelineKHaregiven，thelineMHwillbegiventoo;
　　consequentlytheratioofthelineMHtothelineMKwillbegiventoo。SoMwilltouchagivencircumference。LetthisbeNM。Thentheintersectionofthecircumferencesisgiven，andthesameratiocannotholdbetweenlinesinthesameplanedrawnfromthesamepointstoanyothercircumferencebutMN。
　　DrawalineDBoutsideofthefigureanddivideitsothatD：B=MH：MK。ButMHisgreaterthanMKsincethereflectionoftheconeisoverthegreaterangle（foritsubtendsthegreaterangleofthetriangleKMH）。ThereforeDisgreaterthanB。ThenaddtoBalineZsuchthatBZ：D=D：B。ThenmakeanotherlinehavingthesameratiotoBasKHhastoZ，andjoinMI。
　　ThenIisthepoleofthecircleonwhichthelinesfromKfall。FortheratioofDtoIMisthesameasthatofZtoKHandofBtoKI。Ifnot，letDbeinthesameratiotoalineindifferentlylesserorgreaterthanIM，andletthislinebeIP。ThenHKandKIandIPwillhavethesameratiostooneanotherasZ，B，andD。ButtheratiosbetweenZ，B，andDweresuchthatZB：D=D：B。ThereforeIH：IP=IP：IK。Now，ifthepointsK，HbejoinedwiththepointPbythelinesHP，KP，theselineswillbetooneanotherasIHistoIP，forthesidesofthetrianglesHIP，KPIabouttheangleIarehomologous。Therefore，HPtoowillbetoKPasHIistoIP。ButthisisalsotheratioofMHtoMK，fortheratiobothofHItoIPandofMHtoMKisthesameasthatofDtoB。Therefore，fromthepointsH，Ktherewillhavebeendrawnlineswiththesameratiotooneanother，notonlytothecircumferenceMNbuttoanotherpointaswell，whichisimpossible。SincethenDcannotbearthatratiotoanylineeitherlesserorgreaterthanIM（theproofbeingineithercasethesame），itfollowsthatitmuststandinthatratiotoMI
　　itself。ThereforeasMIistoIKsoIHwillbetoMIandfinallyMHtoMK。
　　If，then，acirclebedescribedwithIaspoleatthedistanceMIitwilltouchalltheangleswhichthelinesfromHandKmakebytheirreflection。Ifnot，itcanbeshown，asbefore，thatlinesdrawntodifferentpointsinthesemicirclewillhavethesameratiotooneanother，whichwasimpossible。If，then，thesemicircleAberevolvedaboutthediameterHKI，thelinesreflectedfromthepointsH，KatthepointMwillhavethesameratio，andwillmaketheangleKMHequal，ineveryplane。Further，theanglewhichHMandMI
　　makewithHIwillalwaysbethesame。SothereareanumberoftrianglesonHIandKIequaltothetrianglesHMIandKMI。TheirperpendicularswillfallonHIatthesamepointandwillbeequal。
　　LetObethepointonwhichtheyfall。ThenOisthecentreofthecircle，halfofwhich，MN，iscutoffbythehorizon。（Seediagram。）
　　NextletthehorizonbeABGbutletHhaverisenabovethehorizon。LettheaxisnowbeHI。Theproofwillbethesamefortherestasbefore，butthepoleIofthecirclewillbebelowthehorizonAGsincethepointHhasrisenabovethehorizon。Butthepole，andthecentreofthecircle，andthecentreofthatcircle（namelyHI）
　　whichnowdeterminesthepositionofthesunareonthesameline。ButsinceKHliesabovethediameterAG，thecentrewillbeatOonthelineKIbelowtheplaneofthecircleAGdeterminedthepositionofthesunbefore。SothesegmentYXwhichisabovethehorizonwillbelessthanasemicircle。ForYXMwasasemicircleandithasnowbeencutoffbythehorizonAG。Sopartofit，YM，willbeinvisiblewhenthesunhasrisenabovethehorizon，andthesegmentvisiblewillbesmallestwhenthesunisonthemeridian;forthehigherHisthelowerthepoleandthecentreofthecirclewillbe。
　　Intheshorterdaysaftertheautumnequinoxtheremaybearainbowatanytimeoftheday，butinthelongerdaysfromthespringtotheautumnequinoxtherecannotbearainbowaboutmidday。Thereasonforthisisthatwhenthesunisnorthoftheequatorthevisiblearcsofitscourseareallgreaterthanasemicircle，andgoonincreasing，whiletheinvisiblearcissmall，butwhenthesunissouthoftheequatorthevisiblearcissmallandtheinvisiblearcgreat，andthefartherthesunmovessouthoftheequatorthegreateristheinvisiblearc。Consequently，inthedaysnearthesummersolstice，thesizeofthevisiblearcissuchthatbeforethepointHreachesthemiddleofthatarc，thatisitspointofculmination，thepointiswellbelowthehorizon;thereasonforthisbeingthegreatsizeofthevisiblearc，andtheconsequentdistanceofthepointofculminationfromtheearth。Butinthedaysnearthewintersolsticethevisiblearcsaresmall，andthecontraryisnecessarilythecase：forthesunisonthemeridianbeforethepointHhasrisenfar。
　　6
　　Mocksuns，androdstoo，areduetothecauseswehavedescribed。
　　Amocksuniscausedbythereflectionofsighttothesun。Rodsareseenwhensightreachesthesunundercircumstanceslikethosewhichwedescribed，whentherearecloudsnearthesunandsightisreflectedfromsomeliquidsurfacetothecloud。Herethecloudsthemselvesarecolourlesswhenyoulookatthemdirectly，butinthewatertheyarefullofrods。Theonlydifferenceisthatinthislattercasethecolourofthecloudseemstoresideinthewater，butinthecaseofrodsontheclouditself。Rodsappearwhenthecompositionofthecloudisuneven，denseinpartandinpartrare，andmoreandlesswateryindifferentparts。Thenthesightisreflectedtothesun：themirrorsaretoosmallfortheshapeofthesuntoappear，but，thebrightwhitelightofthesun，towhichthesightisreflected，beingseenontheunevenmirror，itscolourappearspartlyred，partlygreenoryellow。Itmakesnodifferencewhethersightpassesthroughorisreflectedfromamediumofthatkind;thecolouristhesameinbothcases;ifitisredinthefirstcaseitmustbethesameintheother。
　　Rodsthenareoccasionedbytheunevennessofthemirror-asregardscolour，notform。Themocksun，onthecontrary，appearswhentheairisveryuniform，andofthesamedensitythroughout。Thisiswhyitiswhite：theuniformcharacterofthemirrorgivesthereflectioninitasinglecolour，whilethefactthatthesightisreflectedinabodyandisthrownonthesunalltogetherbythemist，whichisdenseandwaterythoughnotyetquitewater，causesthesun’struecolourtoappearjustasitdoeswhenthereflectionisfromthedense，smoothsurfaceofcopper。Sothesun’scolourbeingwhite，themocksuniswhitetoo。This，too，isthereasonwhythemocksunisasurersignofrainthantherods;itindicates，morethantheydo，thattheairisripefortheproductionofwater。
　　Furtheramocksuntothesouthisasurersignofrainthanonetothenorth，fortheairinthesouthisreadiertoturnintowaterthanthatinthenorth。
　　Mocksunsandrodsarefound，aswestated，aboutsunsetandsunrise，notabovethesunnorbelowit，butbesideit。Theyarenotfoundveryclosetothesun，norveryfarfromit，forthesundissolvesthecloudifitisnear，butifitisfaroffthereflectioncannottakeplace，sincesightweakenswhenitisreflectedfromasmallmirrortoaverydistantobject。（Thisiswhyahaloisneverfoundoppositetothesun。）Ifthecloudisabovethesunandclosetoitthesunwilldissolveit;ifitisabovethesunbutatadistancethesightistooweakforthereflectiontotakeplace，andsoitwillnotreachthesun。Butatthesideofthesun，itispossibleforthemirrortobeatsuchanintervalthatthesundoesnotdissolvethecloud，andyetsightreachesitundiminishedbecauseitmovesclosetotheearthandisnotdissipatedintheimmensityofspace。Itcannotsubsistbelowthesunbecauseclosetotheearththesun’srayswoulddissolveit，butifitwerehighupandthesuninthemiddleoftheheavens，sightwouldbedissipated。
　　Indeed，evenbythesideofthesun，itisnotfoundwhenthesunisinthemiddleofthesky，forthenthelineofvisionisnotclosetotheearth，andsobutlittlesightreachesthemirrorandthereflectionfromitisaltogetherfeeble。
　　Someaccounthasnowbeengivenoftheeffectsofthesecretionabovethesurfaceoftheearth;wemustgoontodescribeitsoperationsbelow，whenitisshutupinthepartsoftheearth。
　　Justasitstwofoldnaturegivesrisetovariouseffectsintheupperregion，sohereitcausestwovarietiesofbodies。Wemaintainthattherearetwoexhalations，onevaporoustheothersmoky，andtherecorrespondtwokindsofbodiesthatoriginateintheearth，’fossiles’andmetals。Theheatofthedryexhalationisthecauseofall’fossiles’。Sucharethekindsofstonesthatcannotbemelted，andrealgar，andochre，andruddle，andsulphur，andtheotherthingsofthatkind，most’fossiles’beingeithercolouredlyeor，likecinnabar，astonecompoundedofit。Thevaporousexhalationisthecauseofallmetals，thosebodieswhichareeitherfusibleormalleablesuchasiron，copper，gold。Alltheseoriginatefromtheimprisonmentofthevaporousexhalationintheearth，andespeciallyinstones。Theirdrynesscompressesit，anditcongealsjustasdeworhoar-frostdoeswhenithasbeenseparatedoff，thoughinthepresentcasethemetalsaregeneratedbeforethatsegregationoccurs。Hence，theyarewaterinasense，andinasensenot。Theirmatterwasthatwhichmighthavebecomewater，butitcannolongerdoso：norarethey，likesavours，duetoaqualitativechangeinactualwater。Copperandgoldarenotformedlikethat，butineverycasetheevaporationcongealedbeforewaterwasformed。Hence，theyall（exceptgold）areaffectedbyfire，andtheypossessanadmixtureofearth;fortheystillcontainthedryexhalation。
　　Thisisthegeneraltheoryofallthesebodies，butwemusttakeupeachkindofthemanddiscussitseparately。
　　BookIV
　　1
　　WEhaveexplainedthatthequalitiesthatconstitutetheelementsarefour，andthattheircombinationsdeterminethenumberoftheelementstobefour。
　　Twoofthequalities，thehotandthecold，areactive;two，thedryandthemoist，passive。Wecansatisfyourselvesofthisbylookingatinstances。Ineverycaseheatandcolddetermine，conjoin，andchangethingsofthesamekindandthingsofdifferentkinds，moistening，drying，hardening，andsofteningthem。Thingsdryandmoist，ontheotherhand，bothinisolationandwhenpresenttogetherinthesamebodyarethesubjectsofthatdeterminationandoftheotheraffectionsenumerated。Theaccountwegiveofthequalitieswhenwedefinetheircharactershowsthistoo。Hotandcoldwedescribeasactive，for’congregating’isessentiallyaspeciesof’beingactive’：moistanddryarepassive，foritisinvirtueofitsbeingacteduponinacertainwaythatathingissaidtobe’easytodetermine’or’difficulttodetermine’。Soitisclearthatsomeofthequalitiesareactiveandsomepassive。
　　Nextwemustdescribetheoperationsoftheactivequalitiesandtheformstakenbythepassive。Firstofall，truebecoming，thatis，naturalchange，isalwaystheworkofthesepowersandsoisthecorrespondingnaturaldestruction;andthisbecomingandthisdestructionarefoundinplantsandanimalsandtheirparts。Truenaturalbecomingisachangeintroducedbythesepowersintothematterunderlyingagiventhingwhentheyareinacertainratiotothatmatter，whichisthepassivequalitieswehavementioned。Whenthehotandthecoldaremastersofthemattertheygenerateathing：iftheyarenot，andthefailureispartial，theobjectisimperfectlyboiledorotherwiseunconcocted。Butthestrictestgeneraloppositeoftruebecomingisputrefaction。Allnaturaldestructionisonthewaytoit，asare，forinstance，growingoldorgrowingdry。
　　Putrescenceistheendofallthesethings，thatisofallnaturalobjects，exceptsuchasaredestroyedbyviolence：youcanburn，forinstance，flesh，bone，oranythingelse，butthenaturalcourseoftheirdestructionendsinputrefaction。Hencethingsthatputrefybeginbybeingmoistandendbybeingdry。Forthemoistandthedryweretheirmatter，andtheoperationoftheactivequalitiescausedthedrytobedeterminedbythemoist。