TWO NEW SCIENCESBY GALILEO

_o
THE MACMILLAN COMPANY _W YORJKBOSTON CItlCAGO DALLAS ATLANTA SANyP.ANCISCO MACMILLAN & CO., Lmrr_ LONDON BOMBAY CALCUTTA _T,_ OUP._NE T_L_; MACMILLAN CO. OF CANADA, LTD. TORONTO

GALILEO GALILEI. Subterman's portrait, aintedaboutI64O; owin the Galler_a Pittiat Florence. p n dr'

DIALOGUES
CONCERNING

TWO NEW SCIENCES

BY

GALILEOGALILEI
Translatedfromthe Italian and Latin into Englishby HENRY CREW AND ALFONSODE SALVIO of Northwestern University
WITH AN INTRODUCTION BY

ANTONIOFAVARO of the University of Padua.

"I think with your friend that it has been of late too much the mode to slight the learning of the ancients." Benjamin Franklin, Phil. Trans.

64,445.(I774.)

N_ _ork
THE MACMILLAN COMPANY I914
All rightsreserved

COPX_IGBT. X914 B_ THEMACMILLAN COMPANY Set up ande[ectrotyped. Published May_I914.

"La Dynamiquest lascience e desforcesaccN6ratricesor retardatrices, des mouvemens et varies qu'ellesdoiventproduire.Cette science due est enti_rement auxmodernes, Galilee cduiqui et est en a jet_les premiers fondemens." agrange ec. L M dad. zzl, I.

TRANSLATORS'

PREFACE

OR morethana centuryEnglishspeaking students avebeenplaced theanomalous h in position hearingGalileo of constantlyreferred asthefounder fmodern to o physical science, without avingany chance read, h to in theirownlanguage, whatGalileo himself has to say. Archimedes as been made h availableyHeath; uygens' hasbeen b H Light turnedinto Englishby Thompson, hileMotte has put the w Principia Newtonbackintothe language whichit was of in conceived.To renderthe Physics f Galileo o alsoaccessible to English American and studentsis the purpose ofthefollowing translation. The lastof the greatcreators the Renaissance of wasnot a prophetwithouthonorin his owntime;for it wasonlyone groupofhis country-menhat failedto appreciateim. Even t h during hislifetime,hisMechanics hadbeenrenderedntoFrench i byoneofthe leading physicists oftheworld, ersenne. M Withintwenty-five earsafterthe deathofGalileo, Diay his logues nAstronomy, o andthoseon TwoNewSciences, hadbeen doneinto Englishby ThomasSalusbury wereworthily and printed in two handsome quarto volumes. The TwoNew Sciences, contains racticallyllthat Galileo which p a hastosayon the subje&ofphysics,ssuedfromthe English i pressin I665.

vi TRANSLATORS' PREFACE It issupposedhatmostofthecopies ere t w destroyed inthegreat London firewhichoccurred inihe yearfollowing. eare not W aware ofanycopyinAmerica: venthat belonging e totheBritish Museum isan imperfecCt one. Againin 173 the TwoNewSciences wasdoneintoEnglish by Thomas Weston;but this book,nownearlytwo centuries old, is scarceand expensive.Moreover,he literalness ith t w whichthis translation wasmaderenders manypassages ither e ambiguousrunintelligible themodernreader. Otherthan o to thesetwo,noEnglish version hasbeenmade. Quite recentlyan eminentItalian scholar,after spending thirtyofthebestyearsofhislifeuponthe sul_jecCt, hasbrought to completion greatNationalEdition of the Worksof the Galileo.Wereferto the twentysuperb volumesn whichProi fessor ntonioFavaro A ofPaduahasgivena definitive resentap tionofthelaborsofthe manwhocreatedthemodern science f o physics. The followingendition r includes neitherLe Mechanichef o Galileo his paperDe MotuAccelerato, nor sincethe formerof thesecontainslittlebut the Staticswhichwascurrentbefore thetimeofGalileo, thelatter isessentiallyncluded the and i in Dialogue oftheThirdDay. Dynamics asthe onesubjecCt w to whichundervariousforms,suchas Ballistics, coustics, A Astronomy,he consistentlynd persistently a devotedhis whole life. Into the one volumeheretranslatedhe seemsto have gathered,duringhis last years,pracCtically allthat is of value either to the engineeror the physicist. The historian,the philosopher, the astronomer illfind the other volumes and w replete withinteresting aterial. m tt ishardlynecessary addthatwehavestricdy to followedhe t textofthe NationalEdition---essentially Elzevireditionof the 1638. Allcomments nd annotations avebeenomittedsave a h hereand therea foot-note intendedto economize reader's the time. To eachofthesefootnotes as beenattachedthe signah ture[Trans.]norderto preserve i theoriginal snearly a intacCt as possible. Muchofthevalueofanyhistorical ocumentiesin the land l guage employed, this is doubly and truewhenoneattemptsto

TRANSLATORS' PREFACE vii trace the rise and growthof any set of conceptssuch as those employedin modem physics. We have thereforemade this translationasliteral as is consistentwith clearnessandmodemity. In caseswherethere is any importantdeviationfromthis rule, and in the caseof many technicaltermswherethere is no deviationfrom it, we have giventhe originalItalian or Latin phrase in italics enclosedin square brackets. The intention here is to illustratethe great variety of termsemployedby the earlyphysiciststo describea singledefiniteidea,and conversely, to illustratethe numeroussensesin which,then as now,a single wordis used. For the fewexplanatoryEnglishwordswhichare placedin squarebrackets without italics,the translatorsalone are responsible.The pagingofthe NationalEditionis indicated in squarebrackets inserted alongthe medianlineof the page. The imperfecCtions the followingpages wouldhave been of manymorebut for the aidof three of our colleagues.Professor D. R. Curtiss was kind enoughto assist in the translationof thosepageswhichdiscussthe nature ofInfinity:ProfessorO.H. Basquin gave valuable help in the renditionof the chapteron Strength of Materials;and ProfessorO. F. Long clearedup the meaningofa numberof Latin phrases. To ProfessorA. Favaroof the UniversityofPaduathe translators share, with every reader, a feelingof sincereobligation for his IntroducCtion.

H.C.

A. DE S.
EVANSTON) ILLINOIS)

x5February, I9x4.

INTRODUCTION
.................. '............................... hisfaithful riend RITINGto f EliaDiodati, hehad in mindto print as being"superior to everythingelseof minehithertopublished";elsewhere e says"they contain h results-which I consider themostimportant ofallmy studies";andthis opinion which _ Galileo speaks o w he expressed fthe"NewSciences" hich concerningis ownworkhas h beenconfirmedy posterity: he "NewSciences" re,indeed, b t a the masterpiece fGalileo o whoat the timewhenhemadethe above remarks adspentuponthemmorethanthirty laborious h years. Onewhowishes tracethe history to ofthis remarkable ork w willfindthat the greatphilosopher laidits foundations uring d the eighteen best yearsof his lifc thosewhichhe spentat Padua. As we learnfromhis last scholar, incenzio iviani, V V the numerousesultsatwhichGalileo r hadarrivedwhile this in city,awakenedntenseadmiration the friends i in whohad witnessed variousexperiments meansof whichhe wasaccusby tomed investigate to interestinguestions q inphysics.FraPaolo Sarpiexclaimed: giveus the Science fMotion,Godand To o Naturehavejoinedhandsand createdthe intellecCt ofGalileo. Andwhenthe "NewSciences"amefromthe pressoneofhis c foremost pupils, aolo P Aproino, wrotethatthevolume ontained c muchwhichhehad "alreadyheardfromhisownlips"during student aysat Padua. d Limitingourselveso onlythe moreimportant t documents which mightbecitedin support fourstatement,t willsuffice o i to mention letter,writtentoGuidobaldoelMonteon the the d 29thofNovember, I6O2, concerning thedescent fheavy o bodies

x INTRODUCTION alongthearcsofcircles andthe chords ubtended ythem;that s b to Sarpi,datedI6thofOctober, 6o4,dealing I withthefreefall of heavybodies;he letterto Afltonio Medicion the IIth t de' ofFebruary, 6o9,in which I hestatesthat hehas"completed all the theorems anddemonstrations pertaining forcesand reto sistances f beamsof various o lengths,thicknessesnd shapes, a proving theyareweaker t the middle that a thanneartheends, that theycancarrya greaterloadwhenthat loadisdistributed throughout lengthof the beamthanwhenconcentratedt the a onepoint,demonstratinglsowhatshapeshould givento a a be beamin orderthat it mayhavethe samebendingstrengthat everypoint,"andthat hewasnowengaged "uponsomequestionsdealing withthemotionofprojeCtiles"; andfinallyinthe letter to Belisario Vinta,dated7th of May, x6IO, oncerning c hisreturnfromPaduato Florence, heenumerates various ieces p ofworkwhichwerestillto becompleted, mentioningxplicitly e threebooks onan entirelynewscience dealing withthe theory ofmotion. Although varioustimesafterthe returnto his at nativestatehe devoted considerable thoughttotheworkwhich, evenat that date,he had in mindas isshown certainfragby mentswhichclearlybelongto different eriods his lifeand p of whichhave,for the first time,beenpublished the National in Edition;and although thesestudieswerealwaysuppermostn i histhoughtit doesnotappear that hegavehimself eriouslyo s t themuntilafterthe publication f the Dialoguend the como a pletion that trial which of wasrightly described thedisgrace as ofthecentury. InfaCtaslate asOctober,x63,he barely o mentionsto Aggiuntihis discoveries the theoryof motion,and in onlytwoyearslater,in a lettertoMarsiliconcerning themotion ofprojeCtiles, hehintsat a booknearlyreadyfor publication in whichhewilltreat alsoofthis subject;and onlya yearafter this hewritesto Arrighettihat hehas inhanda treatiseonthe t resistance fsolids. o But the workwasgivendefinite formby Galileo duringhis enforced residence t Siena: n thesefivemonthsspentquietly a i with the Archbishope himself ritesthat he has completed h w "a treatiseona newbranchofmechanics fullofinteresting and usefulideas";sothat a fewmonthslaterhe wasableto send

INTRODUCTION xi wordto Micanzio the "workwasready";as soonashis that friendsearned l ofthis,theyurgeditspublication.It was,however,no easymatterto print the workofa manalreadycondemned bythe HolyOffice: andsince Galileo could nothopeto printit eitherinFlorence r inRome,heturnedto the faithful o Micanzio asking himto findoutwhether thiswould bepossible inVenice, romwhence f hehadreceived offerso printthe Diat logue onthePrincipal Systems, soonasthenewshadreached as therethat hewasencounteringifficulties. firsteverything d At wentsmoothly;othat Galileo s commenced sendingoMicanzio t someofthe manuscript hichwasreceived the latterwith w by an enthusiasm whichhewassecond noneof thewarmest in to admirers the greatphilosopher.But whenMicanzio of consultedthe Inquisitor, e received answerthat therewas h the an express orderprohibiting printingor reprinting any the of workof Galileo,eitherin Veniceor in any otherplace,hullo excepto. As soonasGalileo received discouraging this newshebegan to lookwithmore favoruponoffers hich w hadcome tohimfrom Germany wherehis friend,andperhapsalsohis scholar, GiovanniBattistaPieroni,wasin the service the Emperor, s of a militaryengineer; onsequently c Galileo gaveto PrinceMattia de'Medici howasjust leaving w forGermany firsttwoDiathe logues be handedto Pieroniwhowasundecided hether to w to publish themat Vienna orPrague at some or placeinMoravia; in themeantime, however, hehadobtained permission print to bothat Vienna andat Olmtitz.But Galileo recognizedanger d at everypointwithinreachof the longarmof the Courtof Rome; ence, vailing h a himself oftheopportunityffered o bythe arrivalof LouisElzevirin Italyin 1636, lsoofthe friendship a betweenthe latter and Micanzio, ot to mentiona visit at n Arcetri,he decided abandon to allotherplansand entrustto the Dutch publisherthe printingof his newworkthe manuscriptofwhich,although notcomplete, Elzevirtookwithhim onhisreturnhome. In the course oftheyear1637,hepriming t wasfinished, and at the beginning ofthe followingeartherewaslacking y only the index,the title-page the dedication.This last had, and

xii INTRODUCTION through thegoodoffices ofDiodati, eenoffered theCountof b to Noailles, formerscholar fGalileo Padua,and since1634 a o at ambassador ofFranceatRome,a manwhodidmuch alleviate to the distressing consequences the celebratedrial; and the of t offerwasgratefully accepted.The phrasing the dedication of deserves riefcomment. Since b Galileo wasaware,onthe one hand,ofthe prohibitiongainstthe printingofhis worksand a since, n theotherhand,he didnotwishto irritatethe Court o ofRome fromwhose handshewasalways hoping complete for freedom, pretends the dedicatoryetter (where, robably he in l p through excess f caution, e gives o h onlymainoutlines)hat he t hadnothingto dowiththeprintingofhis book,asserting that he willneveragainpublishany of his researches, willat and mostdistributehereand there a manuscript opy. He even c expresses reatsurprisehat hisnewDialogues avefalleninto g t h the handsof the Elzevirsand weresoonto be published; o s that, having beenaskedtowritea dedication, hecould thinkof no manmoreworthywhocouldalsoon this occasion defend himagainsthisenemies. As to the title whichreads:Discourses Mathematical and Demonstrations concerning TwoNew Sciences ertaining p toMechanics andLocal otions, onlyisknown, amely,hat the M this n t titleisnot theonewhichGalileo haddevised andsuggested; in fac_he protestedagainstthe publishersakingthe libertyof t changing and substituting lowand common forthe it "a title noble anddignified onecarried uponthetitle-page." In reprintinghis workin the National dition,I havefolt E lowed Leydentextof 1638faithfully notslavishly, the but becauseI wishedto utilizethe largeamountofmanuscript am terialwhich come has downto us,forthepurpose f colTeccting o a considerable numberof errorsin this firstedition,and also for thesakeofinserting ertainadditions esired c d bythe author himself.IntheLeyden Edition,hefourDialoguesrefollowed t a by an"Appendix containing theorems some andtheir proofs, dealzngwithcenters f gravity solidbodies, rittenbythesame o of w Author anearlier ate,"whichhas noimmediate at d connecCtion withthesubjeccs treatedintheDialogues; thesetheorems ere w foundby Galileo, hetellsus, "at the ageoftwenty-twond as a

INTRODUCTION _5ii aftertwoyearsstudyofgeometry" andwerehereinserted only to savethemfromoblivion. But it wasnot the intentionof Galileo that the Dialogues ontheNewSciences should contain onlythefourDaysand the above-mentioned appendix whichconstitute LeydenEdithe tion;while,on the one hand,the Elzevirs werehastening the printing striving completet attheearliest ossible ate, and to i p d Galileo, n the otherhand,kepton speaking f anotherDay, o o besideshe four,thusembarrassing t andperplexing theprinters. Fromthe correspondence whichwenton between authorand publisher, t appears i that this FifthDaywasto havetreated "of the forceof percussionnd the useofthe catenary";, a but as the typographical workapproachedompletion, pnnter c the became anxious the bookto issuefromthe presswithout for furtherdelay;and thus it cameto passthat the Discorsi e Dimostrazioni appeared containingnlythe fourDaysandthe o Appendix, spiteof the fact that inApril,I638,Galileo in had plunged moredeeply thanever"into the profound question of percussion" nd "had almost eached completeolution." a r a s The "NewSciences" owappearin an edition n following the text whichI, afterthemostcareful nd devoted a study,determineduponfor the National dition. It appears E alsoin that languagenwhich, bove i a allothers, havedesired seeit. In I to thistranslation, thelastandripestwork ofthegreatphilosopher makesits first appearancen the NewWorld:if towardthis i important esultI mayhopetohavecontributednsome r i measureI shallfeelamplyrewardedorhavinggiven thisfieldof f to research bestyearsofmylife. the AwroNm FAv_a_o.
UNIVERSITY OF PADUA_

2_ o] October, I9x3.

DISCOR,
E

SI

DIMOSTR.AZIONI
MI.A.TEMA TI C H E, intorno _duenuoue fiiene_e
Atxenenti alla M_-CANICA_ i Mo VIMENTI LOCALt; delSignor GALILEO GALILEI LINCEO, FilofofoMatematico e primario delSereniilimo Grand Duca diTofcana. co_ z_a_ppendiceceutrograuit_ del di d'klc-_ni ._olidi._

IN

LEIDA,

ApprdroEl/_virii. c,xxxv_. gli _.D.

' [431 TO THE MOST ILLUSTRIOUS LORD COUNT OF NOAILLES

Counsellor ofhisMostChristian ajesty, M KnightoftheOrder of theHolyGhost, ieldMarshal nd Commander, F a Seneschal andGovernor ofRouergue, andHis Majesty's ieutenant L inAuvergne, my LordandFForshipful Patron OSTILLUSTRIOUS LORD:In the pleasure whichyou derivefrom the possessionf this workof mineI reco i]_/i[ ll_r_-_[_ yourLordship's agnanimity. he ognize m T i[lV/[]_-4_}disappointment anddiscouragement I have _'_[,_ feltoverthe ill-fortune hichhasfollowed w __ myotherbooksare already known you. to Indeed,I had decided to publish not any moreof my work. Andyet in order to save it fromcomplete oblivion,it seemedto me wiseto leavea manuscript copyin some place whereit would availabletleasttothose be a who followintelligently sub the jet% whichI have treated. Accordingly chosefirstto placemyworkin yourLordship's I hands,askingno moreworthydepository, andbelieving that, onaccount fyouraffecCtion o forme,youwould haveat heartthe preservation mystudiesand labors. Therefore, of whenyou werereturning homefromyourmission Rome, came pay to I to myrespecCts inpersonasI had already donemanytimesbefore byletter. At thismeeting presented yourLordship copy I to a ofthesetwoworks whichat thattimeI happened tohaveready. Inthegraciouseception hich r w yougavetheseI found assurance of

xviii TO THE COUNT OF NOAILLES oftheirpreservation. hefacet T ofyourcarrying themto France and showing themto friendsofyourswhoare skilled these in sciences gaveevidencehatmysilence asnotto beinterpreted t w ascomplete idleness.Alittlelater,just asI wasonthepointof sending othercopieso Germany, landers, ngland, painand t F E S possiblyo some t placesin Italy,I wasnotified theElzevirs by that theyhad theseworksofminein pressand that I oughtto decide upona dedication andsendthema replyat once. This sudden andunexpecCted newsledmeto thinkthat theeagerness of your Lordship reviveand spreadmy nameby passing to theseworkson to variousfriendswasthe real causeof their falling intothe handsofprinterswho,becauseheyhad already t published therworks o ofmine,nowwishedto honormewitha beautiful ndornateedition a ofthiswork. But thesewritings f o minemusthavereceived additional aluefromthe criticism v of so excellent judgeas yourLordship,whoby the unionof a manyvirtueshas wonthe admiration all. Your desireto of enlarge therenown ofmyworkshows ourunparalleledenerosy g ity and your zeal for the publicwelfarewhichyou thought wouldthus be promoted. Under these circumstances is it eminently fittingthat I should,in unmistakable terms,gratefullyacknowledge thisgenerosityn thepartofyourLordship, o whohasgiven myfame to wings thathavecarried intoregions it moredistantthanI had daredto hope. It is,therefore, proper that I dedicateto yourLordship childof my brain. To this this courseI amconstrained otonlybythe weightof obligan tion underwhichyou haveplacedme,but also,if I may so speak,by the interestwhichI havein securing yourLordship as the defender fmyreputation o againstadversaries whomay attackit while remain I underyourprotecCtion. Andnow,advancing underyourbanner,I pay my respecCts to youbywishinghatyoumayberewardedorthesekindnesses t f bythe achievement ofthe highest appiness nd greatness. h a I amyourLordship's Most devotedServant, GALILEO GALILEn _lrcetri, 6March, 638. I

[44]

THE PUBLISHER TO THE READER

INCEsociety isheldtogether ythemutual b services hichmenrenderoneto another, w and sinceto this endthe arts and sciences have largelycontributed, investigations in thesefields havealways eenheldin great b esteem andhave beenhighly regardedyour b wiseforefathers.Thelarger heutilityand t excellence ofthe inventions, greater the has beenthehonorandpraise bestowedpontheinventors.Indeed, u menhaveevendeified themand haveunitedin theattemptto perpetuate memory theirbenefafftors the of bythe bestowal of this supreme onor. h Praiseand admiration likewise to thosecleverinare due tellecCts who, confining their attentionto the known,have discovered corre&edfallaciesand errorsin many and and many a propositionenunciated men of distincCtion by and acceptedorages f asfacet.Althoughhese t menhaveonlypointed outfalsehoodndhavenotreplaced bytruth,theyarenevera it thelessworthyof commendation whenwe considerhe wellt knowndifficulty discovering a difficulty hichledthe of facet, w princeof oratorsto exclaim: tinara U tamfacilepossem vera reperire, uamfalsa convincere.* nd indeed,these latest q A centuries meritthis praisebecause is duringthemthat the it arts andsciences, discovered bytheancients, havebeenreduced to so great and constantly increasingerle&ion p throughthe investigations nd experiments f clear-seeing a o minds. This development particularly is evidentin the caseof the mathematicalsciences.Here,without entioning arious m v menwho haveachieved success, wemustwithout esitation h andwiththe *Cicero. eNatura eorum,9I. [Trans.] d D I,

xx THE PUBLISHER TO THE READER unanimousapprovalof scholarsassignthe first placeto Galileo Galilei,Memberof theAcademyoftheLincei. This hedeserves not only because he has effectivelydemonstrated fallaciesin many of our current conclusions,as is amply shownby his publishedworks, but also becauseby means of the telescope (inventedin this countrybut greatly perfectedby him) he has discoveredthe four satellitesof Jupiter, has shownus the true charaCterof the MilkyWay, and has made us acquaintedwith spots on the Sun, with the rough and cloudyportions of the lunar surface, with the threefold nature of Saturn, with the phases of Venus and with the physical charaCterof comets. Thesematterswereentirelyunknownto the ancientastronomers andphilosophers;so that wemay truly say that he has restored to the worldthe scienceof astronomyand has presentedit in a newlight. Remembering that the wisdomand powerand goodness ofthe Creator are nowhereexhibitedso well as in the heavensand celestialbodies,we can easilyrecognizethe great merit of him who has brought these bodies to our knowledgeand has, in spite of their almost infinite distance, rendered them easily visible. For, accordingto the commonsaying,sight can teach moreandwith greatercertainty in a singleday than can precept even though repeated a thousand times; or, as another says, intuitiveknowledge keepspacewith accuratedefinition. But the divine and natural gifts of this man are shownto best advantage in the present 'workwhere he is seen to have discovered,though not w_6houtmany labors and long vigils, twoentirely new sciencesand to have demonstratedthem in a rigid, that is, geometric,manner: and what is even more remarkablein this work is the facetthat one of the two sciences dealswith a subjeCtof never-endinginterest, perhapsthe most importantin nature, onewhichhas engagedthe mindsof all the great philosophersand one concerningwhichan extraordinary number of books have been written. I refer to motion [moto locale], phenomenonexhibitingvery many wonderfulpropera ties,noneofwhichhas hithertobeen discovered demonstrated or by any one. The other science whichhehas alsodevelopedfrom its

THE PUBLISHER TO THE READER xxi its very foundationsdealswith the resistancewhichsolidbodies offer to fracture by external forces_er violenza], subjectof a great utility, especially in the sciencesand mechanicalarts, and onealsoaboundingin propertiesand theoremsnot hitherto observed. In this volume one finds the first treatment of these two sciences,full of propositionsto which, as time goes on, able thinkers willadd many more;alsoby meansof a large number of clear demonstrationsthe author points the way to many other theoremsas will be readilyseen andunderstoodby allintelligent readers.

TABLE

OF CONTENTS

I Page Firstnewscience, treatingf theresistance solid o which bodies offer tofracture.FirstDay.......................... I II Concerning thecause f cohesion. o Second ay D ............ lO 9

III Second science, new treating motion[movimenti of locah]. ThirdDay........................................ 153 Uniform motion ...................................... 154 Naturally ccelerated ........................... a motion 16o IV Violent otions.ProjeFtiles. m Fourth Day............... 244

V Appendix;heorems t anddemonstrations concerning thecenters of gravityfsolids................................. o . 295

TWO NEW SCIENCESBY GALILEO

FIRST DAY
INTERLOCUTORS: SALVIATI, SAGREDO AND SIMPLICIO
:'............. . :'.............................. ALV. heconstant ctivity T a which youVenetothestudious inda large m fieldforinvestigation,especially that part of the work which involves echanics; m forinthisdepartmentalltypesofinstrumentsndmachines a are constantly beingconstrue'tedy many b _ tiansdisplay s artisans, mongwhom arsenaluggests a inyourfamous theremustbesome who,partlybyinherited experience andpartlyby theirownobservations, havebecome ighly h expert ndclevern explanation. a i SAc_.Youare quitefight. Indeed,I myself, eingcurious b by nature,frequently visitthis placeforthe merepleasure of observing theworkofthosewho,onaccount ftheirsuperiority o overotherartisans, ecall"firstrankmen." Conference w with themhasoftenhelped mein the investigation ofcertaineffec2s includingotonlythosewhich n arestriking, butalsothose which arerecondite nd almost ncredible.AttimesalsoI havebeen a i put to confusionnddrivento despair feverexplainingomea o s thingfor whichI couldnotaccount, utwhich b mysenses told meto be true. Andnotwithstanding thefadtthat whatthe old man told us a little whileago is proverbial nd commonly a accepted, it seemedo mealtogether likemanyanother yet t false, sayingwhichis currentamongthe ignorant; or I thinkthey f introduce theseexpressions orderto givethe appearancef in o knowingomethingbout atters s a m which theydonotunderstand. Salv.

THE TWO NEW SCIENCESOF GALILEO

SAJ_v. Yourefer, erhaps,to that lastremarkofhiswhenwe p askedthe reasonwhy they employed stocks,scaffoldingnd a bracing oflargerdimensions forlaunching bigvesselhanthey a t dofora small ne;andheanswered theydidthisinorderto o that avoid danger ftheshippartingunderitsownheavyweight the o [vasta ole], dangerowhichsmall oatsarenotsubject? m a t b SAcR. Yes,that iswhatI mean;and I referespeciallyo his t lastassertion whichI havealways regarded false,though asa current,opinion; amely,that in speaking theseand other n of similar machines necannotarguefromthe smallto the large, o because many devices whichsucceed a smallscaledo not on workona largescale. Now,since mechanics hasitsfoundation in geometry, where meresizecutsnofigure, donotseethat the I properties circles, riangles, ylinders, of t c conesand othersolid figures illchange w withtheirsize. If,therefore, largemachine a be construcCted insucha way that itspartsbearto oneanother thesameratioasina smaller ne,andifthesmallerssufficiently o i strongfor the purpose whichit wasdesigned, donot see for I whythe largeralsoshould notbeableto withstandanysevere anddestrucCtive teststowhich maybesubjected. it SAT.v. Thecommonpinion o ishereabsolutely wrong. Indeed, it is so far wrongthat precisely opposite true, namely, the is thatmanymachines canbeconstrucCted evenmoreperfecCtly ona largescale thanona small;hus,forinstance, clock t a which indicatesand strikes thehourcanbemademoreaccurate large ona scalethan on a small. Thereare someintelligent eople p who maintainthis sameopinion, ut on morereasonable rounds, b g whenthey cut loose fromgeometry arguethat the better and performance ofthe largemachine isowing the imperfecCtions to andvariations fthematerial. HereI trustyouwillnotcharge o mewith arrogancef I say that imperfections the material, i in eventhosewhichare great enough invalidate clearest to the mathematical proof,arenotsufficient explain deviations to the observed etween b machinesnthe concrete nd inthe abstra_. i a Yet I shallsayit and willaffirm that,evenif the imperfecCtions did

[3o]

Is1]

FIRST DAY 3 didnot exist andmatterwereabsolutelyperfecCt, unalterable and freefrom all accidentalvariations,stillthe mere fact that it is matter makes the larger machine,built of the same material and in the same proportionas the smaller, correspondwith exacCtness the smallerin everyrespecCt to exceptthat it willnot be so strong or so resistant against violent treatment; the larger the machine,the greater its weakness. SinceI assume matter to be unchangeableand alwaysthe same,it is clearthat weare nolessable to treat this constantandinvariableproperty in a rigidmanner than ifit belongedto simpleand pure mathematics. Therefore,Sagredo,you woulddo wellto changethe opinionwhich you, and perhapsalso many other students of mechanics,have entertainedconcerning abilityof machines the and structures to resist external disturbances,thinkingthat whenthey are builtof the samematerialand maintainthesame ratio betweenparts, they are able equally, or rather proportionally, to resist or yield to such external disturbancesand blows. For we can demonstrateby geometrythat the large machineis not proportionately strongerthan thesmall. Finally, we may say that, for every machineand strucCture, whether artificialor natural, there is set a necessarylimit beyondwhich neither art nor nature canpass; it is hereunderstood,of course, that the material is the same and the proportionpreserved. SAGI_. brain alreadyreels. My mind,likea cloudmomenMy tarily illuminatedby a lightning-flash, s for an instant filled i with an unusuallight,whichnowbeckonsto meand whichnow suddenly mingles and obscures strange, crude ideas. From what you have said it appears to me impossible build two to similarstrucCtures ofthe samematerial,but ofdifferentsizesand have themproportionatelystrong; and if this wereso, it would [52] not be possibleto tind twosinglepolesmade of the same-wood which shall be alike in strength and resistancebut unlikein size. SALv. oit is,Sagredo. _And makesurethat weunderstand S to each other, I say that if we take a woodenrod of a certain length and size, fitted, say, into a wall at right angles,i. e., parallel

4 THE TWO NEW SCIENCES OF GALILEO parallelto the horizon,it may be reducedto sucha lengththat it willjust supportitself; so that if a ha_r'sbreadth be addedto its lengthit willbreakunder its ownweightandwillbe the only rodof the kind in the world.* Thus if,for instance,its lengthbe a hundredtimesits breadth,youwillnot be able to findanother rodwhoselengthis alsoa hundred timesits breadth and which, like the former, is just able to sustain its own weight and no more:allthe largeroneswillbreakwhileallthe shorteroneswill be strong enough to support somethingmore than their own weight. And thiswhichI havesaidabout the abilityto support itselfmust beunderstoodto applyalsoto other tests; so that if a pieceof scantling[corrente] willcarrythe weightoften similarto itself, a beam [trave] aving the same proportionswill not be h ableto supportten similarbeams. Please observe, gentlemen,how faCtswhich at first seem improbablewill, even on scant explanation,drop the cloak which has hidden them and stand forth in naked and simple beauty. Who doesnot knowthat a horsefallingfrom a height of three or four cubitswill break his bones,whilea dog falling fromthe sameheightor a cat froma heightof eightor ten cubits will sufferno injury? Equally harmlesswouldbe the fall of a grasshopper from a toweror the fallof an ant fromthe distance of the moon. Do not childrenfallwith impunityfromheights whichwouldcosttheir eldersabroken legor perhapsa fraCtured skull? And just assmalleranimalsare proportionatelystronger and more robustthan the larger,so alsosmallerplants are able to stand upbetter than larger. I amcertainyouboth knowthat an oak two hundred cubits [braccia] wouldnot be able to high sustainits ownbranchesif they weredistributedas in a tree of ordinarysize;and that nature cannotproducea horseas largeas twenty ordinary horses or a giant ten times taller than an ordinary man unless by miracle or by greatly altering the proportionsofhis limbsand especiallyofhis bones,whichwould have to be considerablyenlargedover the ordinary. Likewise the currentbeliefthat, in the caseof artificialmachinesthevery *The authorhereapparently meansthat the solutionis unique. [Trans.]

'

[53]

FIRST DAY 5 largeand the smallare equallyfeasibleand lastingis a man_fest error. Thus, for example,a smallobeliskor columnor other solidfigurecan certainlybe laiddownor setup without danger ofbreaking,whilethe very largeoneswillgoto piecesunderthe slightestprovocation,and that purely on accountof their own weight. AndhereI must relatea circumstance whichis worthy ofyourattention asindeedare alleventswhichhappencontrary to expecCtation, especiallywhen a precautionarymeasureturns out to be a causeof disaster. A large marble columnwas laid out so that its two ends rested each upon a pieceof beam; a little later it occurredto a mechanicthat, in orderto be doubly sureof its not breakingin the middleby its ownweight,it would be wise to lay a third support midway;this seemedto all an excellentidea;but the sequelshowedthat it wasquitethe opposite, for not many monthspassedbeforethe columnwas found crackedand broken exadtlyabovethe newmiddlesupport. Sn_P.A very remarkable and thoroughly unexpectedaccident, especiallyif causedby placingthat new support in the middle. SALV. Surely this is the explanation,and the moment the cause is knownour surprisevanishes;for when the two pieces of the columnwereplacedon levelgroundit wasobservedthat one of the end beamshad, after a long while,becomedecayed and sunken,but that the middleone remainedhard and strong, thus causingone halfof the columnto projecCt the air without in any support. Under these circumstancesthe body therefore behaveddifferentlyfrom what it wouldhave doneif supported only upon the first beams; becauseno matter howmuch they might have sunken the columnwouldhave gonewith them. This is an accidentwhichcouldnot possiblyhave happenedto a smallcolumn,eventhoughmadeofthe samestoneand havinga length corresponding its thickness,i. e., preservingthe ratio to betweenthicknessand lengthfoundin the large pillar. SAc_.I am quite convinced ofthe fa_s of the case,but I do not understandwhy the strength and resistanceare not multipliedin the sameproportionas the material;and I am the more puzzled

[541

6 THE TWO NEW SCIENCES OF GALILEO puzzledbecause,on the contrary, I have noticedin other cases that the strength and resistanceagainstbreaking increasein a largerratio than the amountof material. Thus, for instance,if two nails be driven into a wall, the one which is twice as big as the other will supportnot only twiceas muchweightas the other,but three or fourtimes asmuch. SALv.Indeedyouwillnot be far wrongifyousay eighttimes as much; nor doesthis phenomenoncontradicCt other even the thoughin appearancethey seemso different. SACR. Will you not then, Salviati,remove these difficulties and clear awaythese obscuritiesif possible:for I imaginethat this problemofresistanceopensup a fieldofbeautifuland useful ideas;and if youare pleasedto make thisthe subjecCt to-day's of discourseyou will placeSimplicioand me under many obligations. SALV. am at your serviceif only I can call to mindwhat I I learned from our Academician* who had thought much upon this subjecCt and accordingto his custom had demonstrated everything by geometricalmethods so that one might fairly call this a new science. For, althoughsomeof his conclusions had been reachedby others, first of all by Aristotle,these are not the most beautiful and, what is more important,they had not been proven in a rigidmanner fromfundamentalprinciples. Now, since I wish to convinceyou by demonstrativereasoning rather than to persuadeyou by mere probabilities,I shall supposethat youare familiarwith present-daymechanicsso far as it is needed in our discussion. First of all it is necessary,to considerwhat happenswhena pieceofwoodor any other solid which coheres firmly is broken; for this is the fundamental facet,nvolvingthe firstand simpleprinciplewhichwemusttake i for grantedas wellknown. To grasp this more clearly,imaginea cylinderor prism,AB, made of wood or other solid coherent material. Fasten the upper end, A, so that the cylinder hangs vertically. To the lower end, B, attach the weight C. It is clear that however great they may be, the tenacity and coherence [tenacit_e I. e. Galileo: he authorfrequently T refersto himself nderthis u name. [Tran_r.]

FIRST DAY 7 [55] eoeren_] between partsofthis solid, o longastheyarenot the s infinite, anbeovercomey thepullofthe weight , a weight c b C which canbeincreasedndefinitely i untilfinally solid the breaks likea rope. Andas in the caseofthe ropewhose strength we knowto be derived froma multitudeof hempthreadswhich composet, soin thecaseofthe wood, eobservetsfibres nd i w i a filaments runlengthwisend render muchstronger a it than a hempropeof the samethickness.But in the caseof a stoneor metallic cylinder herethe' w coherence seemsto be stillgreater cement the whichholdsthe parts together mustbe somethingotherthan filamentsnd fibres;and yet a eventhis canbebroken strong bya pull. Srme.If thismatterbe asyousay I canwell understand the fibres that ofthewood,beingas longas thepieceofwooditself,renderit strong and resistant againstlargeforces tendingto breakit. But howcan one makea ropeone hundredcubitslongoutofhempen fibres which arenotmorethantwoor threecubitslong,and stillgiveit somuchstrength Besides, should ? I begladto hearyouropinion to themannern as i whichthe parts of metal,stone,and othermaterialsnot showing filamentoustrucCture a s are Fig.i put together;for, if I mistakenot, theyexhibitevengreater tenacity. SALV. solve To theproblems hich w youraiseit willbenecessaryto makea digression intosubjecCts havelittlebearing which uponourpresentpurpose. SAcg.But if, by digressions, canreachnewtruth, what we harm is there in makingone now,so that we maynot lose this knowledge, remembering suchan opportunity, that once omitted, aynotreturn;remembering m alsothatwearenottied down fixed toa andbriefmethod butthat wemeetsolely forour ownentertainment? Indeed,whoknows utthat wemaythus b

[S6]

frequently

8 THE TWO NEW SCIENCESOF GALILEO frequently discover something ore interesting beautiful m and thanthe solution originallyought._ begofyou,therefore, s I to grantthe request f Simplicio, o whichis alsomine;forI amno lesscurious desirous and thanhe to learnwhatis the binding materialwhichholdstogetherthe partsof solidsso that they canscarcely separated.Thisinformations alsoneededto be i understand coherence f the parts of fibresthemselves the o of Whichome s solids arebuiltup. SAJ_V.am at yourservice,sinceyou desireit. The first I question is,Howare fibres, achnot morethantwo or three e cubitsin length, o tightlyboundtogetherinthe caseof a rope s onehundredcubitslongthat greatforce[violent] isrequired to breakit? Nowtellme,Simplicio, canyounotholda hempen fibreso tightlybetween yourfingers I, pullingby the otherend, that wouldbreakit beforedrawing awayfromyou? Certainly it youcan. Andnowwhenthefibres ofhempareheldnotonlyat the ends,but aregrasped bythe surrounding medium throughouttheirentirelengthisit notmanifestly oredii_cult o tear m t themloose fromwhatholdsthemthanto breakthem? But in the caseofthe ropethe veryacCt oftwisting causes threads the tobindoneanothern sucha waythatwhen i theropeisstretched witha greatforcethe fibresbreakratherthan separatefrom eachother. At the pointwherea ropepartsthe fibresare,as everyone knows, eryshort,nothing v likea cubitlong, stheywoaldbeif a the partingof the ropeoccurred, ot by the breaking the n of filaments, bytheirslipping neovertheother. but o SAGR. confirmation In ofthis it maybe remarkedhat ropes t sometimes reak not by a lengthwise ull but by excessive b p twisting.This,it seemso me,isa conclusive t argument because the threadsbindone anotherso tightlythat the compressing fibres donotpermitthose whicharecompressed lengthen to the spiralseventhat littlebit bywhichit is necessaryor themto f lengthen inorder surroundherope to t which, ntwisting, rows o g shorterandthicker. SALv. Youarequiteright. Nowseehowonefa_ suggests
another

FIRST DAY 9 another. Thethreadheldbetween fingers the doesnot yield to onewhowisheso draw awayevenwhenpulledwithcont it siderableorce, ut resistsbecause is heldbackby a double f b it compression, seeingthat the upperfinger presses againstthe lower s hamasthe lower gainsttheupper.Now,ifwecould a a retainonlyoneof thesepressures thereis no doubtthat only half the original esistance r wouldremain;but sincewe are _ not able,by lifting,say, the upperfinger,o removeoneof t thesepressures ithoutalsoremoving other,it becomes w the necessary preservene of themby meansof a newdevice to o whichcausesthe threadto pressitselfagainstthe finger or againstsome othersolid bodyuponwhichit rests; ndthusit is a brought boutthat theveryforce hich a w pulls it in orderto snatchit awaycompresses it moreand moreas the pullincreases. his T is accomplishedy wrappingthe thread b aroundthe solidin the mannerof a spiral; _I_ andwillbebetterunderstood ymeansofa b figure.LetABandCDbetwocylinderseb tween whichis stretched thethreadEF: and _ O for the sakeof greaterclearness ewillimw agineit to be a smallcord. If these two cylinders e pressedstronglytogether,the b cordEF, whendrawnbytheendF, willundoubtedly standa considerable pullbeforeit slipsbetweenthe two compressingolids. s But if weremove oneofthesecylindershe t cord,thoughremainingn contacCt the i with other, willnot therebybe prevented from slipping freely. On the other hand,if one holdsthecordloosely againstthe topof the Fig. 2 cylinderA, windsit in the spiralformAFLOTR, nd then a pullsit by the endR, it is evident hat the cordwillbeginto t bindthe cylinder; greaterthe number f spiralsthe more the o tightlywillthe cordbe pressedagainstthe cylinderby any given pull. Thusasthe number f turnsincreases, lineof o the contacCt

[ST]

Io THE TWO NEW SCIENCESOF GALILEO contactbecomesongerand in consequence l moreresistant;so that the cordslipsandyieldsto thetractiveforce withincreasingdifficulty. [58] Isitnotclearthatthisisprecisely thekindofresistance hich w onemeetsin thecaseofa thickhempropewhere thefibres form thousands nd thousands similar a of spirals?And,indeed,the qbinding effecCt theseturnsis sogreatthat a fewshortrushes of woventogetherintoa fewinterlacingpiralsformoneof the s strongestof ropes whichI believetheycallpack rope[susta]. SAoR. Whatyou sayhas cleared twopointswhichI did up notpreviously understand.Onefact is howtwo,or at most three,turnsofa ropearoundtheaxleofa windlass annot nly c o holdit fast,but canalsopreventit fromslipping henpulled w by the immense forceof the weight[forzadelpeso] hichit w sustains; ndmoreover byturningthewindlass, same a how, this axle,by merefricCtion the ropearoundit, canwindup and of lift hugestoneswhilea mereboy i'sableto handle the slack therope. TheotherfaCt asto dowith of h a simple utclever evice,nvented young b d i bya kinsmanof mine,forthe purpose descendingroma of f window meansof a ropewithoutlacerating by the palmsofhishands,ashad happened himshortly to before andgreatlytohisdiscomfort.Asmall ketch s willmakethis clear. He took a woodencylinder, AB,aboutasthickasa walking stickandaboutone spanlong:on this he cut a spiralchannel f about o oneturnand a half,andlargeenough just receive to theropewhich hewished touse. Havingintroduced theropeat theendAandledit outagainat the end BB, he enclosed oth the cylinderand the ropein a b caseofwoodor tin, hingedalongthe 81deo that it s Fig. couldbe easilyopenedand closed. After he had 3 Iastenedtheropeto a firmsupportabove,he could, graspon ingandsqueezing thecasewith bothhands,hangbyhis arms. The pressure therope,lyingbetweenthe caseand the cylon inder,wassuchthat he could,at will,eithergrasp the case
more

FIRST DAY II moretightly and holdhimselffromslipping, r slackenhis o hold anddescendsslowly a ashewished. SALV. truly ingenious evice! I feel,however, hat for A d t a complete xplanation e otherconsiderations mightwellenter; yet I mustnotnowdigress uponthisparticularopicsinceyou t arewaitingto hearwhatI thinkaboutthebreaking strength of other materials which,unlikeropesand mostwoods, o not d showa filamentoustructure. The coherencef thesebodies s o is,in my estimation, roduced othercauses hichmaybe p by w grouped undertwoheads. Oneis that much-talked-of repugnance whichnatureexhibits towards vacuum; ut thishorror a b of a vacuumnot beingsufficient, is necessaryo introduce it t anothercause theformof a gluey viscousubstance hich in or s w bindsfirmly togetherthecomponent partsofthebody. FirstI shallspeakofthe vacuum, emonstrating definite d by experiment qualityand quantityofits force[o/rt_].If you the taketwohighlypolished andsmooth platesofmarble, metal,or glassandplacethemfaceto face,onewillslideoverthe other withthe greatestease,showing conclusively thereisnoththat ingof a viscous aturebetween n them. But whenyouattempt to separatethemandkeepthemat a constantdistanceapart, youfindtheplatesexhibit ucha repugnance separationhat s to t the upperonewillcarrythe lower newithit andkeepit lifted o indefinitely, evenwhenthelatterisbigandheavy. This experiment howsthe aversionof nature for empty s space,evenduring thebriefmomentrequired fortheoutside air to rushin andfillup theregion between twoplates. It is the alsoobserved if twoplatesare not thoroughly that polished, theircontactisimperfectothat whenyouattemptto separate s them slowly onlyresistance fferedis that of weight;if, the o however,he pullbe sudden, hen the lowerplate rises,but t t quickly fallsback,havingfollowed theupper lateonlyforthat p very shortintervalof time required the expansion f the for o smallamountof air remaining etweenthe plates,in conseb quence oftheirnotfitting, ndfortheentrance a ofthesurroundingair. This resistance whichis exhibited between two the plates

IS9]

Iz THE TWO NEW SCIENCESOF GALILEO platesisdoubtlessikewise resent etween l p b thepartsofa solid, and enters,at leastin par[, as a concomitant causeof their coherence. SAGR. Allow to interruptyoufor a moment, lease; or me p f I wantto speakof something hich occurs me,namely, w just to whenI seehowthe lower latefollowshe upperoneandhow p t rapidlyit is lifted,I feelsurethat, contrary the opinion to of manyphilosophers, including perhapsevenAristotlehimself, motionin a vacuumis notinstantaneous.If this wereso the two platesmentioned abovewouldseparatewithoutany resistance whatever, eeingthat the sameinstantof timewould s sufficeor theirseparation for the surrounding edium f and m to rushin and fillthe vacuum between them. The fa&that the lowerplatefollowshe upperone allows s to infer,not only t u that motionin a vacuumis not instantaneous, alsothat, but betweenthe twoplates,a vacuum reallyexists,at leastfor a veryshorttime,sufficient allow surrounding ediumto to the m rushin andfillthe vacuum; or iftherewerenovacuumthere f would benoneedofanymotion inthemedium.Onemustadmit thenthat a vacuumis sometimes roduced violentmotion p by [violenza] contraryto the lawsof nature,(although my or in opinion nothing occurs contrary natureexcept heimpossible, to t andthat never occurs). But hereanotherdifficulty arises. Whileexperiment onc vincesmeofthe correcCtness of thisconclusion, mymindis not entirelysatisfied to the causeto whichthis effe&is to be as attributed. For the separationof the plates precedesthe formation the vacuumwhichis produced a consequence of as ofthisseparation; andsince appearsomethat,in theorderof it t nature,the causemustprecedethe effe&, venthoughit ape pearsto follow pointoftime,and sinceevery in positive effecCt musthavea positive cause,I do notseehowthe adhesion of twoplatesand theirresistance separation--acCrual to fa_s---can be referredto a vacuumas cause whenthis vacuumis yet to follow. According the infallible aximof the Philosopher, to m thenon-existent canproduce oeffe&. n Simp.

[6o]

FIRST DAY 13 Sire,. Seeingthat youacceptthis axiomofAristotle,I hardly thinkyouwillreje_ another excellent nd reliable a maximof his, namely,Nature undertakesonly that which happenswithout resistance;and in this saying,it appearsto me,youwillfindthe solutionof your difficulty. Sincenature abhorsa vacuum,she preventsthat fromwhicha vacuumwouldfollowas a necessary consequence.Thus it happensthat nature preventsthe separation ofthe twoplates. SACR. Nowadmittingthat what Simplicio saysis an adequate solutionof my difficulty,it seemsto me, ifI may be allowedto resume my former argument, that this very resistanceto a vacuumought to be sufficientto hold togetherthe parts either of stoneor of metalor the parts of any other solidwhichis knit togethermorestronglyandwhichismoreresistantto separation. If for one effe_ there be only one cause,or if, more beingassigned,they canbe reducedto one, thenwhyis not this vacuum whichreally existsa sufficientcausefor allkinds of resistance ? SALV. do not wish just nowto enter this discussionas to I whether the vacuum alone is sufficientto hold together the separateparts of a solidbody;but I assureyouthat the vacuum whichacCts a sufficient ausein the caseofthetwoplatesisnot as c alone sufficientto bind togetherthe parts of a solidcylinderof marble or metal which, when pulled violently,separates and divides. Andnow if I finda methodof distinguishing thiswell known resistance,dependingupon the vacuum, from every other kind which might increasethe coherence,and if I show you that the aforesaidresistancealone is not nearly sufficient for such an effect, will you not grant that we are bound to introduceanother cause. Help him, Simplicio,since he does _ not knowwhat replytomake. SIMP. urely,Sagredo'shesitationmust be owingto another S reason,for therecan be nodoubt concerning conclusion a which isat oncesoclearandlogical. SACra. Youhave guessedrightly,Simplicio. I waswondering whether, if a million of gold each year from Spain were not sufficientto pay the army, it might not be necessary to make

[6i]

I4 THE TWO NEW SCIENCES OF GALILEO make provision other than small coin for the pay of the soldiers.* But go ahead, Salviati;assumethat I admit your conclusion and showus yourmethodof separatingtheacCtion thevacuum of from other causes;and by measuringit showus how it is not sufficiento producethe effectin question. t SALV. Your good angel assist you. I will tell you how to separate the force of the vacuum from the others, and a{terwards how to measure it. For this purposelet us consider a continuoussubstancewhoseparts lack all resistanceto separation exceptthat derivedfroma vacuum,suchas is the casewith water,a fact fullydemonstratedbyourAcademician oneof his in treatises. Whenever cylinderofwaterissubjectedto apull and a [62] offersa resistanceto the separation itsparts this canbe attribof uted to noothercausethantheresistance the of I have invented a devicewhichI can better explainby meansof a sketch than by mere words. Let CABDrepresentthe crosssection of a cylindereither of metal or, preferably, vacuum. In orderto try suchan experiment of glass,hollowinsideand accuratelyturned.

/k_j C G .

I-I i Dcylinderofis introduced a perfecdyfitting Into this wood,representedin crosssection by EGHF, and capableof up-and-downmotion. Through the middleof this cylinderis boreda holeto receivean ironwire,carrying a hook at the end K, while the upper end of the wire, I, is providedwith a conical head. The wooden cylinder is countersunk Fig.4 at the top so as to receive,witha perfect fit, the conical head I of the wire, IK,when pulled down by theendK. NowinsertthewoodencylinderEH in the hollow cyllnderAD, so as not to touchthe upper end of the latterbut to leavefree a spaceof two or three finger-breadths; his spaceis to be filled t *Thebearing fthisremark o becomes onreading hatSalviati clear w says onp. 18below.[Trans.]

FIRST DAY 15 withwaterbyholding thevessel iththemouthCD upwards, w pushing down onthestopper EH,andat thesametimekeeping theconical eadofthewire,I, away h fromthe hollow portion of thewooden cylinder.Theairisthusallowedoescape t alongside the ironwire(which doesnotmakea close assoonas one fit) pressesdownon the woodenstopper. The air havingbeen allowedo escape theironwirehaving t and beendrawn backso that it fits snugly againstthe conical epression the wood, d in invert thevessel, ringingtmouthdownwards, b i andhangonthe hookK a vesselwhichcanbe filled with sandor any heavy materialin quantity sufficiento finallyseparatethe upper t surface ofthe stopper, F, fromthelowersurface E ofthewater to whichit wasattached onlyby theresistance ofthevacuum. Next weighthe stopperand wiretogetherwith the attached vesseland its contents; e shallthenhave the forceof the w vacuum [forza ddvacuo].Ifoneattachesoa cylinder fmarble t o [63] or glassa weight hich, ogether iththe weight fthemarble w t w o or glassitself,is just equalto the sumof the weights before mentioned, if breaking and occurs shallthenbejustified we in sayingthat thevacuum aloneholdsthepartsofthe marble and glasstogether; ut ifthisweight oesnotsufficendifbreaking b d a occursonlyafteradding, ay,fourtimesthis weight, eshall s w thenbe compelled say that the vacuumfurnishes nlyone to o fifthofthetotalresistanceresf._ema]. [ SLurP. one candoubtthe cleverness No ofthe device; et it y presentsmanydifficulties which makemedoubtits reliability. Forwhowillassure thattheair doesnotcreepinbetween us the glassand stopperevenif it is wellpackedwith towor other yielding material._ question I alsowhether oiling withwaxor turpentine willsufficeo makethecone, , fit snugly t I onitsseat. Besides, aynot the parts of the waterexpandand dilate? m Whymaynotthe airorexhalations orsome othermoresubtile substances enetrate p theporesofthewood,orevenoftheglass itself? SAT.v. Withgreatskillindeed hasSimplicio laidbefore the us difficulties; andhehasevenpartlysuggestedowtopreventthe h
air

x6 THE TWO NEW SCIENCESOF GALILEO airfrompenetratinghewood passing t or between thewood and the glass. Butnowletmepoifitoutthat, asour experience increases, shalllearnwhetheror notthesealleged we difficulties reallyexist. For if, as is the casewithair,wateris bynature expansible, although onlyunderseveretreatment, e shallsee w the stopperdescend; nd if we put a smallexcavation the a in upperpart ofthe glassvessel, uchas indicated s byV, thenthe air or any other tenuousand gaseous substance, whichmight penetratethe poresof glassor wood,wouldpassthroughthe waterandcolle& inthis receptacle Butifthesethingsdonot V. happenwemayrest assured ourexperknent that hasbeenperformedwithpropercaution;andwe shalldiscoverhat water t doesnot dilate and that glassdoesnot allowany material, however tenuous, penetrate to it. SAGm Thanks thisdiscussion,havelearned to I thecause ofa certaineffe& whichI havelongwondered and despaired at of understanding.I oncesawa cistern whichhad beenprovided with a pumpunderthe mistakenimpression the water that mightthusbedrawnwithlesseffort rin greater o quantitythan bymeansofthe ordinary bucket. Thestockofthe pumpcar[64] riedits sucker ndvalveintheupperpartsothatthewaterwas a liftedby attra&ion andnotbya pushasisthe case withpumps in whichthe suckerisplacedlowerdown.Thispumpworked peffedtly solong asthewaterinthecistern stoodabove certain a level;but belowthis levelthe pumpfailedto work. WhenI firstnoticed thisphenomenonthoughtthemachine asoutof I w order;but the workman whomI called to repairit toldme in the defecCt wasnot in the pumpbut in the waterwhichhad fallentoolowto beraisedthrough sucha height;andheadded that it wasnot possible, itherby a pumpor by any other e machine working the principle f attra&ion,o liftwatera on o t hair's breadthaboveeighteencubits;whetherthe pump be largeor smallthis is the extreme limitof the lift. Up to this timeI had beensothoughtlesshat,although knewa rope,or t I rodofwood,orof iron,if sufficiently long,wouldbreakby its ownweight henheldbytheupperend,it never w occurred me to " that

FIRST DAY 17 that the samethingwould happen, nlymuchmoreeasily, o a o t columnof water. And reallyis not that thing whichis attra_ed in the pumpa column ofwaterattachedat the upper endandstretched moreandmoreuntilfinally pointisreached a whereit breaks,likea rope,onaccount fits excessive o weight ? SALV. ispreciselyhe wayit works;hisfixed That t t elevation ofeighteen cubits istrueforanyquantityofwaterwhatever, be the pumplargeor smallor evenasfineas a straw. Wemay therefore that, onweighing watercontained a tube say the in eighteen cubitslong,no matterwhat the diameter, e shall w obtainthevalueoftheresistancefthevacuum cylinder o ina of any solidmaterialhavinga boreof this samediameter.And havinggoneso far, let us seehoweasyit is to findto what lengthcylinders fmetal,stone,wood, lass,etc.,ofanydiamo g eter canbe elongated withoutbreaking theirownweight. by [6S] Takefor instance copper ireofanylengthand thickness; a w fixthe upperend and to the otherend attacha greaterand greaterloaduntilfinally thewirebreaks;letthemaximum load be, say,fiftypounds. Then it is clearthat if fiftypoundsof copper,in additionto the weightofthe wireitselfwhichmay be, say,z/sounce,is drawnout intowireof this samesizewe shallhavethegreatest ength l ofthiskindofwirewhich cansustain its ownweight. Supposehewirewhichbreaks beone t to cubitin lengthand I/souncein weight;thensinceit supports 5olbs.in additionto its ownweight,i.e.,48ooeighths-of-anounce,it followshat allcopper ires, ndependent t w i ofsize,can sustainthemselves to a lengthof48Olcubitsandnomore. up Sincethen a copperrod can sustainits ownweightup to a lengthof48Olcubitsit followshat that part ofthe breaking t strength [resistenza] depends ponthevacuum, omparing which u c itwiththe remainingacetors f ofresistance, isequalto theweight ofa rodofwater,eighteen cubits longandasthickasthecopper rod. If,forexample, coppersninetimesasheavyaswater,the i breaking strength[resistenza allostrappars.z]anycopperrod, of inso farasit depends ponthevacuum, sequalto theweight u a of twocubitsof this samerod. By a similar methodonecan find

I8 THE TWO NEW SCIENCESOF GALILEO findthe maximum lengthofwireorrod ofany material hich w willjust sustainits ownweight,and canat the sametimediscover partwhichthevacuum the playsinits breaking strength. SACR. stillremains youto tellus uponwhatdepends It for theresistanceo breaking, t otherthanthatofthevacuum; hat w is the glueyor viscous substance whichcementstogetherthe partsof the solid? For I cannotimagine gluethat willnot a burnup in a highlyheatedfurnace twoor threemonths,or in certainly withintenor a hundred. Forif gold,silverandglass are keptfor a longwhilein the moltenstateand are removed fromthe furnace,theirparts,on cooling,mmediatelyeunite i r and bind themselvesogetheras before. Not only so, but t whatever ifficultyrises d a withrespe_to thecementationfthe o partsof the glass arisesalsowithregard thepartsoftheglue; to in otherwords, hatisthat whichholdsthesepartstogetherso w firmly? SALv. littlewhile A ago,I expressed thehopethatyourgood angelmightassistyou. I nowfindmyselfn the samestraits. i Experiment eavesno doubtthat the reasonwhytwo plates l cannotbe separated, exceptwithviolent ffort,isthat theyare e heldtogetherby the resistance fthe vacuum; nd the same o a canbe saidof two largepiecesof a marble bronzecolumn. or Thisbeingso,I donotseewhythissame cause maynotexplain the coherence fsmaller artsandindeed the verysmallest o p of particles f thesematerials.Now,since o eacheffe_musthave onetrueandsufficient andsince findnoothercement, m cause I a I notjustified intryingto discover whetherhevacuum t isnot a sufficient cause? S_. But seeingthat youhavealready provedthat the resistancewhichthe largevacuumoffersto the separation of twolargepartsofa solid isreally verysmall n comparison i with thatcohesiveorce f which bindstogetherhemostminute t parts, whydo you hesitateto regardthis latter as something very different fromtheformer ? S_v. Sagredo asalready h [p.I3 above] nswered a thisquestionwhenhe remarked eachindividual oldier asbeing that s w paid

[661

FIRST DAY 19 paidfromcoincoiled-ted a generalaxofpennies by t andfarthings,whileevena million ofgoldwould sufficeo pay the not t entirearmy. Andwhoknowsbut that theremay be other extremely minutevacuawhichaffecCt smallest articles the p so that that which bindstogetherhe contiguous t partsisthroughoutofthesamemintage Letmetellyousomething hich ? w has just occurredo meandwhich donotoffer t I asanabsoluteacet, f but ratheras a passing thought, tillimmature calling s and for morecareful onsideration. c Youmaytakeofit whatyoulike; andjudgethe rest asyouseefit. Sometimes whenI haveobservedhowfirewindsits way in betweenthe mostminute particles fthisorthatmetaland,eventhough o thesearesolidly cemented together,tearsthemapartand separateshem,and t whenI haveobserved that,onremovinghefire,theseparticles t reunitewiththe sametenacityas at first,withoutany lossof quantityin the caseofgoldand withlittlelossin the caseof othermetals, venthoughthesepartshavebeenseparated e fora longwhile,I havethoughtthat theexplanation ightlieinthe m fact that the extremely fineparticles fire,penetrating of the slenderporesof the metal(toosmallto admiteventhe finest particles air or of manyother fluids), ouldfill the small of w intervening acuaandwould v setfreethesesmallparticles rom f the attracCtion whichthesesamevacuaexertuponthemand which prevents theirseparation.Thustheparticles reableto a movefreely sothat the mass[rnassa] becomes fluidandremains soaslong astheparticles ffireremain o inside; utiftheydepart b and leavethe former acuathentheoriginal ttraction v a [attrazzione]eturnsandthepartsareagaincementedogether. r t In replytothe question raised bySimplicio, onemaysaythat althougheachparticularvacuumis exceedingly inuteand m therefore easily overcome, yettheirnumber ssoextraordinarily i greatthat theircombined resistances,so to speak,multipled i almostwithoutlimit. The nature and the amountof force [forza] hichresults[risulta] w fromadding togetheran immense numberof smallforces[debolissimi rnornent_] clearlyillusis tratedbythefa_ that a weight fmillions fpounds, uspended o o s by

[671

20 THE TWO NEW SCIENCES OF GALILEO by great cables,is overcomeand lifted, when the south wind carries innumerableatoms of water, suspendedin thin mist, whichmovingthroughthe air penetratebetweenthefibresof the tense ropes in spite of the tremendousforce of the hanging weight. When these particles enter the narrow pores they swell the ropes, thereby shorten them, and perforcelift the heavymass[mole]. SAcR. here can be no doubt that any resistance,so long as T it is not infinite,may be overcomeby a multitudeof minute forces. Thus a vast numberof ants might carry ashorea ship laden with grain. And since experienceshowsus daily that one ant can easilycarry onegrain,it is clearthat the numberof rains in the ship is not infinite,but fallsbelowa certainlimit. you take another number four or six times as great, and if you set to work a corresponding numberof ants theywillcarry the grain ashoreand the boat also. It is true that thiswill call for a prodigiousnumberof ants, but in my opinionthis is preciselythe casewith the vacua which bind together the least particlesofa metal. SALV. even if this demandedan infinitenumberwould But you stillthink it impossible ? SACR. if the mass [mole] metal were infinite;otherNot of wise .... SAT.V. Otherwise what? Now since we have arrived at paradoxeslet us see if we cannot provethat within a finiteextent it ispossibleto discoveran infinitenumberofvacua. Atthe sametime we shall at least reacha solutionof the mostremarkable of all that list of problemswhichAristotle himselfcalls wonderful;I referto his Questionsn Mechanics.This solution i may be no lessclearand conclusive than that whichhe himself givesand quitedifferentalsofromthat so cleverly expoundedby themostlearnedMonsignordi Guevara.* First it is necessaryto considera proposition, ot treated by n others, but uponwhichdependsthe solutionofthe problemand from which, if I mistake not, we shall derive other new and remarkable facts. For the sake of clearnesslet us draw an *Bishop fTeano; . x56xd.I64I. [Trans.] o b ,

[68]

FIRST DAY 2I accurate figure. AboutG as a centerdescribe equiangular an andequilateral polygon ofanynumber fsides, aythehexagon o s ABCDEF. Similarto this and concentric ith it, describe w anothersmaller newhich o weshallcallHIIZT.MN. Prolong the

F ,

................., ff"i 4
-,,I !

,
t

.
T
, I

Fig. 5 sideAB,of the largerhexagon,ndefinitely i towardS; in like mannerprolong corresponding the sideHI ofthe smaller exh agon,in the samedirecCtion, that the lineHT isparallelto so AS;and throughthe centerdrawthe lineGV parallelo the t othertwo. Thisdone,imagine largerpolygon rollupon the to [69] thelineAS,carrying withit thesmaller olygon.It isevident p that,if thepointB,the endofthesideAB,remainsixed the f at beginningfthe rotation,thepointAwillriseandthe pointC o willfalldescribinghearc CQuntilthesideBCcoincides t with thelineBQ,equaltoBC. Butduring thisrotation thepointI, onthe smaller olygon, p willriseabove thelineIT becauseBis I obliqueoAS;andit willnotagainreturn thelineITuntilthe t to pointC shallhavereachedtheposition . ThepointI, having Q describedhe arcIO above lineHT,willreachtheposition t the Oat

":
iim]

2z THE TWO NEW SCIENCES OF GALILEO 0 at the sametime the sideIK assumesthe position0P; but in the meantimethe centerG has traverseda path aboveGVand doesnot return to it until it has _ompletedthe arc GC. This stephavingbeen taken,the largerpolygonhas been broughtto rest with its sideBC coinciding with the lineBQ whilethe side IK of the smallerpolygonhas been made to coincidewith the lineOP,havingpassedoverthe portion I0 withouttouchingit; alsothe centerG willhave reachedthe positionC after having traversed allits courseabovethe parallellineGV. Andfinally the entire figure will assumea positionsimilarto the first, so that ifwe continuethe rotationand cometo the next step, the sideDC of the largerpolygonwillcoincidewith the portionQX and the sideKL of the smallerpolygon, avingfirstskippedthe h arc PY,will fallon YZ, whilethe centerstillkeepingabovethe lineGV will return to it at R after havingjumpedthe interval CR. At the endofonecompleterotationthe largerpolygonwill have traced upon the lineAS,without break, six linestogether equal to its perimeter; the lesser polygonwill likewisehave imprintedsix linesequal to its perimeter,but separatedby the interpositionof five arcs, whose chords represent the parts of HT not touchedby the polygon:the centerG neverreaches the lineGV exceptat sixpoints. From this it is clearthat the spacetraversedby the smallerpolygonis almostequal to that traversed by the larger, that is, the lineHT approximatesthe lineAS,differingfrom it onlyby the lengthof one chordof one ofthese arcs,provided understandhe lineI-ITto include we t the fiveskippedarcs. Now this exposition whichI have givenin the caseof these hexagonsmust be understoodto be applicableto all other polygons, hatever numberof sides,providedonlythey are w the similar, concentric,and rigidly connecCted, that when the so greateronerotates the lesserwillalsoturn howeversmallit may be. Youmust alsounderstandthat the linesdescribedby these two are nearlyequal providedwe includein the spacetraversed by the smallerone the intervalswhichare not touchedby any part ofthe perimeterofthis smallerpolygon.

[70]

Let

:
:

FIRST DAY z3 Let a large polygon of, say, one thousandsides make one completerotationand thus layoff a lineequal to its perimeter; at the sametime the smallonewillpassover an approximately equal distance, made up of a thousand small portions each equal to oneof its sides,but interrupted by a thousandspaces which,in contrastwith the portionsthat coincide with the sides of the polygon,we may call empty. So far the matter is free fromdifficulty doubt. or But nowsupposethat about any center, say A, we describe two concentricand rigidlyconneCtedcircles;and supposethat from the points C and B, on their radii, there are drawn the tangentsCE and BF and that throughthe centerAthe lineAD is drawn parallel to them, then if the large circlemakes one completerotation alongthe lineBF, equal not only to its circumference alsoto the other two linesCE andAD, tell me but what the smallercirclewilldoand alsowhat the centerwilldo. Asto the center it will certainlytraverseand touch the entire lineAD whilethe circumference the smallercirclewillhave of measuredoffby its pointsof contaCtthe entire lineCE,just as wasdoneby theabovementionedpolygons.The onlydifference is that the lineI-ITwas not at everypoint in contact with the perimeterof the smallerpolygon,but there wereleftuntouched as manyvacant spacesas there werespacescoinciding with the sides. But herein the caseofthe circles circumferencefthe the o smalleroneneverleavesthe lineCE, sothat nopart of the latter isleftuntouched,noris thereevera timewhensomepointonthe circleis not in contaCt ith the straightline. Hownowcan the w smallercircletraverse a lengthgreater than its circumference unlessit goby jumps? 8AGmIt seemsto methat onemaysaythat just as thecenter ofthe circle,by itself,carriedalongthe lineAD is constantlyin contac2 with it, althoughit isonlya singlepoint,sothepointson the circumference the smaller circle,carried alongby the of motionof the largercircle,wouldslideover somesmallparts of the lineCE. [7I] SALV. There are two reasonswhy this cannot happen. First because

24 THE TWO NEW SCIENCESOF GALILEO because thereis nogroundforthinking onepointof conthat taCt,suchas that at C, ratherthan another, houldslipover s certain portionsofthe lineCE. But if suchslidings longCE a didoccurtheywould infinitein number incethe pointsof be s contaCt being ( merepoints)are infinite number:an infinite in number ffiniteslipswillhowever o makeaninfinitely longline, while matteroffaCtthelineCEisfinite. Theotherreason asa isthat asthe greatercircle,in its rotation, hangestspointof c i contactcontinuously lesser ircle the c mustdothe same because Bistheonlypointfrom whicha straightlinecanbedrawntoA andpassthrough Accordingly C. thesmallcircle mustchange itspointofcontactwheneverhelargeonechanges: t nopointof the smallcircletouches straightlineCE in morethanone the point. Not onlyso,but evenin the rotationof the polygons therewasnopointon theperimeter fthe smaller hichcoino w cidedwithmorethanonepointonthe linetraversed that by perimeter; his is at onceclearwhenyou rememberhat the t t lineIK isparallelo BCandthatthereforeK willremain t I above IP untilBCcoincides withBQ,andthatIK willnotlieuponIP except ttheveryinstant henBCoccupies a w theposition Q;at B thisinstanttheentireline coincides IK withOPandimmediately afterwardsisesabove r it. SAOl_. Thisisa veryintricate atter. I seenosolution.Pray m explain to us. it SALV. us returnto the consideration Let ofthe above mentionedpolygons hosebehavior ealready w w understand.Now in the caseofpolygons ith IOOOOO thelinetraversed w sides, by the perimeterof the greater,i. e., the line laid downby its IOOCXX) oneafteranother,sequalto thelinetracedoutby sides i the IOCX:_ sidesofthe smaller, provided weincludehe IO(Xx_ t vacantspaces interspersed.Soin thecaseofthe circles, olyp gonshavingan infinitude sides,the linetraversed the of by continuouslyistributed[continuamente d dispostz] infinitude of sidesisin the greatercircle equalto thelinelaiddownby the infinitude f sidesin the smaller o circle withthe exception but that these latter alternatewith emptyspaces;and sincethe sidesarenotfinitein number, utinfinite, oalsoaretheinterb s vening

FIRST DAY 25 veningempty spacesnot finitebut infinite. The linetraversed by the largercircleconsiststhen of an infinitenumberof points whichcompletely fillit; whilethat whichistracedbythe smaller circleconsistsof an infinitenumberof pointswhichleaveempty spacesand only partly fill the line. And here I wish you to observethat after dividingand resolvinga line into a finite numberof parts, that is,into a numberwhichcan becounted,it [72] is not possibleto arrangethem again intoa greaterlengththan that whichthey occupiedwhen they formeda continuum[continuate]and were conne_ed without the interpositionof as many empty spaces. But if we considerthe lineresolvedinto an infinitenumberof infinitelysmalland indivisible parts, we shall be ableto conceivethe lineextendedindefinitely the by interposition,not of a finite, but of an infinitenumberof infinitelysmallindivisible emptyspaces. Nowthiswhichhasbeen saidconcerning simplelinesmust be understoodto hold alsoin the caseof surfacesand solidbodies, it being assumedthat they are made up of an infinite,not a finite, number of atoms. Such a body once divided into a finitenumberofparts it is impossible toreassemble themsoas to occupy more space than before unless we interpose a finite number of empty spaces,that is to say, spacesfree from the substanceof whichthe solidis made. But if we imaginethe body, by someextreme and final analysis, resolvedinto its primaryelements,infinitein number,then we shall be able to think of them as indefinitelyextended in space, not by the interpositionof a finite, but of an infinite number of empty spaces. Thus one can easilyimaginea smallball of gold expanded into a very large spacewithout the introducCtion a of finite number of empty spaces,always provided the gold is madeupof aninfinitenumberof indivisible parts. SIM1,. t seemsto me that you are travellingalongtoward I thosevacuaadvocatedby a certainancientphilosopher. SAzv.But youhavefailedto add, "whodeniedDivineProvidence," an inapt remarkmade on a similaroccasionby a certain antagonistofour Academician. Simp.

26 THE TWO NEW SCIENCESOF GALILEO Sr_P.I noticed,and notwithoutindignation, rancorof the this ill-natured opponent;furtherreferenceso theseaffairsI t omit,not onlyas a matterof goodform,but alsobecause I know howunpleasantheyareto the goodtemperedandwell t orderedmindof one so religious nd pious,so orthodox a and God-fearingsyou. a But to returnto our subject,yourprevious iscourseeaves d l with me manydifficulties whichI am unableto solve. First among theseis that,if the circumferences twocircles re ofthe a equalto the two straightlines,CE and BF, the latter consideredasa continuum, formeras interrupted the with an infinityofemptypoints,I donotseehowit ispossible saythat to thelineAD describedythecenter,andmadeupofan infinity b ofpoints, sequalto thiscenter hich single oint. Besides, i w isa p thisbuilding upof linesoutofpoints, ivisiblesutofindivisid o bles,andfinites outofinfinites, meanobstacle ifficult offers d to avoid;and the necessity introducing vacuum,so concluof a sively refutedbyAristotle, presentshesame t difficulty. [73] SAr.V. Thesedifficulties real;and theyare not the only are ones. But let us rememberhat weare dealing t withinfinities and indivisibles, both of whichtranscendour finiteunderstanding,he former n account ftheirmagnitude, latter t o o the because ftheirsmallness.In spiteofthis,mencannotrefrain o fromdiscussing them,eventhough mustbe donein a roundit aboutway. Therefore alsoshould I liketo taketheliberty present ome to s of my ideaswhich,thoughnot necessarilyonvincing, c would, onaccount f theirnovelty, t least,provesomewhat o a startling. But sucha diversion ightperhaps m carryus toofarawayfrom the subjectunderdiscussion mighttherefore and appear you to inopportune andnotverypleasing. SACR. Prayletus enjoythe advantagesndprivileges a which comefromconversationetween b friends, speciallyponsube u jects freelychosenand not forceduponus, a matter vastly differentromdealing f withdeadbooks hichgiveriseto many w doubtsbutremove one. Share n withus,therefore, thethoughts which

FIRST DAY z7 which discussion our hassuggested you;for since to wearefree fromurgentbusinessherewillbe abundantimeto pursue t t the topics alreadymentioned; nd in particularthe objetions a raised bySimplicio notinanywisetobenegle&ed. ought S_J_v. Granted,sinceyouso desire.Thefirst question was, Howcana single pointbe equalto a line? Since cannotdo I moreat presentI shallattemptto remove, rat leastdiminish, o one improbability introducing similaror a greaterone, by a just as sometimeswondersdiminished a i byamiracle.* Andthis I shalldoby showing youtwo equalsurfaces, togetherwithtwoequalsolidslocated uponthesesamesurfaces asbases, llfourofwhichdiminish ontinuously a c anduniformly in sucha waythat theirremainderslways a preserve equality among themselves, finally and boththe surfaces ndthe solids a terminate theirpreviousonstant quality degenerating, c e by the one solidand the one surfaceintoa verylongline,the other solidand the other surface a singlepoint;that is, the into latterto onepoint,theformero aninfinite umber fpoints. t n o [74] SACR. proposition This appearsto me wonderfial, indeed; butletusheartheexplanationnddemonstration. a SALV. Sincethe proofis purelygeometrical shallneed we a figure. Let _FB be a semicircle withcenterat C; aboutit describe re&angle DEBand fromthe centerdrawthe the A straightlinesCDand CEto thepointsD andE. Imaginehe t radius CFto bedrawn perpendicular eitherofthelines to ABor DE, andtheentirefigure rotateaboutthisradiusasanaxis. to It isclearthatthere&angle DEBwillthusdescribe cylinder, A a thesemicircle AFBa hemisphere, andthetriangle CDE,a cone. Nextletus removehe hemisphere leave coneand the t but the restofthecylinder, which, naccount fitsshape, ewillcalla o o w "bowl." Firstwe shallprovethat the bowland the coneare equal;thenweshall how s thataplane drawn parallelothecircle t which forms thebaseofthebowlandwhich hasthelineDEfor diameter nd F for a centerwaplane a whose traceisGN---cuts thebowlinthepoints O,N,andtheconeinthepoints L, G,I, I-I, sothat thepartofthe cone indicated byCHLisalways qualo e t *Cf.p.3obelow.Trans.] [

z8 THE TWO NEW SCIENCESOF GALILEO thepartofthebowlwhose rofile p isrepresentedythetriangles b GAIandBON. Besideshisweshallprovethat thebaseofthe t cone,i.e.,thecircle whose iametersHL,isequaltothecircular d i A C 5 surface whichforms thebaseof onemightsay,equaltoa ribbon N whosewidthis OI. (Noteby the waythe natureof mathematicaldefinitions hichconw ___ this portionof the bowl,or as . sistmerely the impositionf in o D F _ names or,ifyouprefer, bbrevia Fig. 6 ations ofspeech establishednd a introduced orderto avoidthe tediousdrudgerywhich ou in y and I now experience simplybecause have not agreed we to call this surfacea "circularband" and that sharp solid portionof the bowla "round razor.") Now call them by [75] whatnameyouplease, tsufficeso understand theplane, i t that drawnat any heightwhatever,so longas it is parallelto the base,i. e., to the circle whose diameter isDE, alwayscuts thetwosolids othattheportion s CHLofthecone isequalo the t upperportionofthebowl;likewise thetwoareaswhich arethe basesofthesesolids, amely n thebandandthecircle-IL, realso I a equal. Herewehavethemiracle entioned bove; sthe cutm a a tingplaneapproaches lineABthe portions fthe solids ut the o c offarealways qual,soalsotheareasoftheirbases.Andasthe e cuttingplanecomes earthe top,thetwosolids n (always qual) e aswellastheirbases(areas which arealsoequal) finally vanish, onepairofthemdegenerating intothecircumference circle, ofa theotherintoa single point,namely,heupperedge t ofthebowl andthe apexof the cone. Now,sinceas thesesolids diminish equality ismaintained etween b themupto theverylast,weare justified sayingthat, at the extreme finalend of this in and diminution, theyare still equaland that oneis not infinitely greaterthan the other. It appearsthereforethat we may equatethecircumference largecircle a single ofa to point. And thiswhichistrueofthe solids istruealsoofthe surfaces hich w form G

FIRST DAY 29 formtheirbases;forthese alsopreserveequalitybetweenthemselvesthroughout their diminutionand in the end vanish, the one into the circumference a circle,the other into a single of point. Shallwenot then callthemequalseeing thatthey arethe last tracesand remnantsof equalmagnitudes Note alsothat, ? even if these vesselswere large enoughto contain immense celestial emispheres, oth their upper edgesand the apexesof h b the cones therein contained would always remain equal and wouldvanish,the formerinto circleshavingthe dimensions of the largestcelestialorbits, the latter into singlepoints. Hence in conformitywith the precedingwe may say that all circumferencesof circles,howeverdifferent,are equal to each other, andareeachequalto a singlepoint. SAtin.This presentationstrikes me as so clever and novel that, even if I were able, I wouldnot be willingto opposeit; for to defacesobeautifula stru_ure by a bluntpedanticattack wouldbe nothingshortofsinful. But for our completesatisfac[76] tion pray give us this geometricalproof that there is always equality between these solidsand between their bases; for it cannot,I think, fail to be very ingenious, eeinghow subtle is s the philosophical argumentbaseduponthis result. SAJ_v. demonstrationis both short and easy. Referring The to the precedingfigure,sinceIPC is a right anglethe squareof the radiusIC is equalto the sumofthe squareson thetwo sides IP, PC; but the radiusIC is equalto AC and alsoto GP, while CP isequalto PH. Hencethe squareof the lineGP is equalto the sumof the squaresof IP andPH, ormuklplyingthroughby 4,wehave thesquareof the diameterGN equalto the sumofthe squareson IO and HL. And, sincethe areasof circlesare to eachother as the squaresof their diameters,it followsthat the areaofthe circle whosediameterisGN isequalto the sumofthe areasof circles avingdiametersIO andI-i-L, h sothat ifweremove the commonarea of the circlehavingIO for diameterthe remainingarea of the circleGN willbe equal to the area of the circlewhosediameterisHL. Somuchfor thefirstpart. Asfor the otherpart, weleaveitsdemonstration forthe present,partly because

30 THE TWO NEW SCIENCES OF GALILEO because thosewhowishto follow willfindit in the twelfth it propositionfthesecond o bookof centro ravitatis De g solidorum bytheArchimedes ofourage,LucaValerio,* homadeuseofit w for a different bjecc, partly because,or our purpose, t o and f i sufficeso have seenthat the above-mentioned t surfacesare always equalandthat, as theykeepon diminishing uniformly, theydegenerate, oneintoa single the point,theotherintothe circumference circle ofa largerthanany assignable; thisfa& in liesourmiracle.t SACR. demonstration ingenious nd the inferences The is a drawnfromit areremarkable. ndnowletus hearsomething A concerning otherdifficultyaised Simplicio, youhave the r by if anythingspecialto say,which,however, eemsto me hardly s possible, incethe matterhas already s beenso thoroughly dis-

cussed.

S_mv. ut I dohavesomethingpecial o say,and willfirst B s t of all repeatwhatI saida littlewhileago,namely,that infinityand indivisibility arein theirverynatureincomprehensibleto us; imagine thenwhattheyarewhencombined.Yet if [77] wewishto buildup a lineout of indivisibleoints,wemust p take an infinite numberof them,and are,therefore, oundto b understand both the infiniteand the indivisible t the same a time. Manyideas havepassed through mymindconcerning this subjecCt, ofwhich, ossiblyhemoreimportant, maynot some p t I be ableto recallonthe spurof themoment;but in thecourse ofourdiscussion mayhappen I shallawaken it that inyou,and especially Simplicio, in objecCtions difficulties hichin and w turn willbringto memory which, ithoutsuchstimulus, that w wouldhavelaindormantinmymind. Allow methereforehe t customaryibertyofintroducing ofourhumanfancies, l some for indeedwemayso callthemin comparison ith supernatural w truth whichfurnisheshe onetrue and saferecourse decit for sionin our discussions whichis an infallible and guidein the darkanddubious athsofthought. p *Distinguished mathematician; Ferrarabout 5S2; Italian bornat a I admitted totheAccademiaLincel died dei I612; I618.[Trans.] Jf f.p.27above.[Trans.] C

FIRST DAY 3I Oneof the mainobjeccions urgedagainsthis building t up of continuousuantities of indivisible q out quantities[continuo d' Cndivisibih] that the additionof one indivisibleo anis t othercannotproduce divisible,or if this wereso it would a f renderthe indivisible ivisible.Thusif twoindivisibles, d say twopoints,canbe unitedto forma quantity,say a divisible line,thenan evenmoredivisibleinemightbe formed the l by unionofthree,five,seven, r anyotheroddnumber fpoints. o o Since howeverheselinescanbe cut intotwo equalparts,it t becomes ossibleo cutthe indivisible p t whichliesexacdy the in middle oftheline. In answero thisandotherobjecdons t ofthe sametypewe replythat a divisible agnitude m cannot e conb stru(tedoutoftwoortenora hundred thousandndivisibleS, ora i butrequires ninfinite a number fthem. o Sire,.Herea difficulty resentsitselfwhichappears me p to insoluble.Sinceit is clearthat wemayhaveonelinegreater than another,eachcontaining infinitenumberof points, an we are forcedto admitthat, withinoneand the sameclass, we mayhavesomething reaterthaninfinity, ecause ing b the finityof pointsin the longlineis greaterthan the infinity of pointsin theshortline. Thisassigning an infinite to quantity a valuegreaterthaninfinity isquitebeyond mycomprehension. SALv. This is one of the difficulties whicharisewhenwe attempt,withourfiniteminds, o discusshe infinite, t t assigning to itthoseproperties which wegivefothefinite andlimited; but this I thinkiswrong, forwecannot peak s ofinfinite uantities q as beingtheonegreater rlessthanorequalto another.To o provethis I haveinmindanargument which, forthe sakeof clearness,shallput intheform I ofquestions Simplicio to who raised thisdifficulty. I takeit forgrantedthatyouknow which ofthenumbers re a squares andwhich arenot. Sn_P. amquiteaware I thata squared numbersonewhich i resultsfromthemultiplication another umber yitself;thus of n b 4,9,etc.,aresquared numbers hich w come multiplying,3, from 2 etc.,bythemselves. Salv.

[78]

37 THE TWO NEW SCIENCES OF GALILEO SALV. Very well;and youalsoknowthat just as the products are calledsquaresso the favors are calledsidesor roots; while on the other hand those numberswhichdo not consistof two equal facCtorsre not squares. Thereforeif I assert that all a numbers, including both squares and non-squares,are more than the squares alone, I shall speak the truth, shall I not? Snvn,. ost certainly. M SALV. t shouldask furtherhowmanysquaresthere areone If might reply truly that there are as many as file corresponding numberof roots,sinceeverysquarehas its own rootand every root its own square, whileno squarehas more than one root and norootmorethan onesquare. SIMP.Preciselyso. SALV. if I -inquirehowmany rootsthere are, it cannot But be deniedthat there are as manyas there are numbersbecause every number is a root of some square. This beinggranted we must say that there are as many squaresas there are numbers becausethey are just as numerousas their roots, and all the numbers are roots. Yet at the outset we said there are many more numbersthan squares, sincethe largerportion of them are not squares. Not only so, but the proportionate number of squares diminishesas we pass to larger numbers. Thusup to IoowehaveIOsquares,that is,the squaresconstitute I/IO part of all the numbers;up to IOOOO, findonly I/IO0 we [79] part to be squares;and up to a milliononly I/IOOO on the part; otherhand in an infinitenumber,ifone couldconceive sucha of thing, he wouldbe forced to admit that there are as many squaresas there arenumbersalltaken together. SAGR. What then must one concludeunder these circumstances? SALV. far as I see we can only infer that the totality of So all numbersis infinite,that the number of squares is infinite, and that the number of their roots is infinite; neither is the number of squaresless than the totality of all numbers, nor the latter greater than the former; and finally the attributes "equal," "greater," and "less," are not applicableto infinite, but

FIRST DAY 33 but only to finite, quantkies. When thereforeSimpllc[ointroducesseverallines of differentlengthsand asksme how it is possible that the longerones do not contain more points than the shorter,I answerhim that one linedoesnot contain more or lessor just as many points as another,but that each line containsan infinite number. Or if I had repliedto him that the pointsin one llnewereequalin numberto the squares; in another,greaterthan thetotality ofnumbers;andin the little one, asmany asthe numberof cubes,mightI not, indeed,have satisfiedhim by thus placingmore points in one line than in another and yet maintainingan infinite numberin each? So muchforthe firstdifficulty. SAGg. ray stop a momentand let me add to what has alP readybeen said an idea whichjust occursto me. If the precedingbe true, it seemsto me impossible say either that one to infinitenumberis greater than anotheror eventhat it isgreater than a finitenumber,becauseifthe infinitenumberweregreater than, say, a millionit wouldfollowthat on passing from the millionto higher and highernumberswe wouldbe approaching the infinite;but this is not so; on the contrary, the larger the number to which we pass, the more we recedefrom [this property of]infinity,becausethe greater the numbersthe fewer [relatively] re the squares contained in them; but the a squares in infinity cannot be less than the totality of all the numbers,as we havejust agreed;hencethe approachto greater and greaternumbersmeansa departurefrominfinity.* SAT.v. thus fromyour ingeniousargumentwe areled to And concludethat the attributes "larger," "smaller,"and "equal" have no placeeither in comparinginfinitequantitieswith each other or in comparinginfinitewith finitequantities. I pass now to another consideration. Sincelines and all continuousquantities are divisibleinto parts whichare themselvesdivisiblewithout end, I do not see how it is possible *Acertain confusion ofthought ppearsobeintroduced a t herethrough a failure distinguishetween to b thenumber andtheclass fthefirstn n o numbers; andlikewise froma failure distinguish to infinitysa number a from infinity astheclass ofallnumbers.[Trans.]

[8o]

34 THE TWO NEW SCIENCESOF GALILEO to avoidthe conclusion theselinesare builtup of an inthat finitenumberofindivisibleuantities q because division a a and subdivision hichcan be carriedon indefinitely resupposes w p that theparts areinfinite number, therwisehe subdivision in o t wouldreachan end;and if thepartsareinfinite number, e in w mustconclude theyare not finitein size,because inthat an finitenumberoffinitequantities would givean infinite magnitude. Andthuswe havea continuous quantitybuiltupof an infinite number f indivisibles. o Shay.But if we can carryon indefinitelyhe division t into finiteparts what necessity therethen for the introduction is ofnon-finlte parts? SALV. very facet hat one is ableto continue, ithout The t w end,the divisionntofiniteparts[inpattiquante] i makesit necessaryto regardthe quantityas composed an infinite of number of immeasurablymallelements[di infinitinon quanta]. s Nowin orderto settlethis matterI shallask you to tellme whether,in youropinion, continuum madeup of a finite a is orofaninfinite number ffiniteparts[parti uante]. o q SIMI,. answeris that their numberis both infiniteand My finite;potentiallyinfinitebut afftually finite[infinite, n po.. i tenza;efinite,in atto];hat is to say,potentially t infinite before division actually and finiteafterdivision; ecause artscannot b p be saidto existin a bodywhichis notyet divided at least or marked out;ifthisisnotdone wesaythattheyexistpotentially. SALV. that a linewhichis, for instance,twentyspans So longis notsaidto contain afftually twentylineseachonespan in lengthexceptafterdivision intotwentyequalparts; before division is saidto containthemonlypotentially.Suppose it thefacets areasyousay;tellmethenwhether, whenthedivision is oncemade,the sizeof the original uantityis therebyinq creased, diminished, orunaffecCted. SIMV. neither It increasesordiminishes. n SALV. is my opinionalso. Therefore finiteparts That the [pattiquante]n a continuum, i whethera&uallyor potentially present,donotmakethe quantityeitherlargeror smaller; ut b it is perfecCtly clearthat, if thenumberoffiniteparts aCtually contained

FIRST DAY 35 containedn thewhole i isinfinite number, in theywillmakethe magnitudenfinite.Hencethenumber ffinite i o parts,although existing nlypotentially, o cannot einfinite nlesshemagnitude b u t containing thembeinfinite; ndconversely themagnitudes a if i finiteit cannotcontain infinite an number ffinitepartseither o actually potentially. or SAGe. Howthenisit possibleo divide continuum t a without limitintopartswhicharethemselves always apable fsubdivic o sion? SAT.V. ThisdistinCtion ofyours between actualandpotential appearso rendereasybyonemethod t whatwould beimpossible by another. But I shallendeavoro reconcile t thesematters in anotherway;and as to the querywhether finiteparts the of a limitedcontinuum [continuoerminato] finiteor int are finitein numberI will,contraryto the opinion Simplicio, of answerhattheyareneitherinite t f norinfinite. SIMP. Thisanswerwould neverhaveoccurred mesinceI to didnotthinkthat thereexisted intermediate any stepbetween the finiteand the infinite, o that the classification distincs or tionwhich assumeshat a thingmustbeeitherfinite infinite t or isfaulty anddefective. SALv. seems Soit tome. Andifweconsider discrete uantiq ties I thinkthereis, between finiteand infinite quantities, a third intermediateermwhichcorresponds everyassigned t to number;so that if asked,as in the presentcase,whether the finitepartsof a continuum arefiniteor infinite number he in t best replyisthat theyareneither finitenorinfinite correbut spondto everyassigned number. In orderthat this maybe possible, isnecessary those it that partsshould notbeincluded withina limited number,orin that casetheywould f notcorrespondto a number hichisgreater; orcantheybeinfinite w n in numbersincenoassigned number s infinite; nd thus at the i a pleasure ofthe questioner wemay,to any givenline,assign a hundred finite parts,a thousand, hundred a thousand, indeed or anynumber emayplease w solongasit benotinfinite.I grant, therefore, the philosophers_ the continuum to that contains as many

36 THE TWO NEW SCIENCESOF GALILEO manyfinitepartsas theypleaseand concedelsothat it conI a tainsthem,eitheraCtuallyr potentially, o astheymaylike;but I mustaddthat justasa lineten fathoms Jeanne] inlengthcontainsten lineseachof one fathomand fortylineseachof one cubit[braccia] eightylineseachof halfa cubit,etc.,soit and containsan infinitenumber f points;callthemaCtual r poo o tential,asyoulike, oras to thisdetail,Simplicio, f I defer o your t opinion to yourjudgment. and SL_P. cannothelp admiring I your discussion; I fear but that this parallelism betweenthe pointsand the finiteparts contained linewillnotprovesatisfaCtory, ina andthat youwill notfindit so easyto divide givenlineintoan infinite a numberofpointsasthe philosophers cut it intotenfathoms r doto o fortycubits;notonlyso,butsucha divisions quiteimpossible i to realizein praCtice, sothat thiswillbe one of thosepotentialities which cannot ereduced actuality. b to SALV. Thefact that somethinganbe doneonlywitheffort c ordiligencerwithgreatexpenditureftimedoesnotrenderit o o impossible; I thinkthat youyourself ould for c noteasilydivide a lineinto a thousand parts,and muchlessif the numberof parts were937or any other largeprimenumber. But if I wereto accomplish thisdivision hichyoudeemimpossible w as readily anotherperson as woulddivide lineintofortyparts the would youthenbemore willing, inourdiscussion, toconcedehe t possibilityfsucha division o ? Snvn,. general enjoygreatlyyourmethod;and replying In I to yourquery,I answerhat it wouldbe morethansufficient t if it provenotmoredifficult resolve lineintopointsthanto to a divide intoa thousand it parts. SALv. willnowsaysomething hich I w mayperhaps astonish you; it refersto the possibility dividing lineinto its inof a finitely smallelements y followinghe sameorderwhichone b t employsn dividinghesamelineintoforty,sixty,ora hundred i t parts,that is,bydividing intotwo,four,etc. Hewhothinks it that, byfollowing thismethod, ecanreachan infinite h number ofpointsisgreatly mistaken;orif thisprocess erefollowed f w to etemiw
/

[821

FIRST DAY 37 eternity there wouldstill remain finiteparts whichwere undivided. Indeedby such a methodone is very far from reachingthe goal of indivisibility;on the contrary he recedesfrom it and whilehe thinksthat, by continuingthis division by multiand plying the multitude of parts, he will approachinfinity,he is, in my opinion,getting farther and farther awayfromit. My reasonis this. In the precedingdiscussion concluded we that, in an infinitenumber,it is necessarythat the squaresand cubes shouldbe as numerousas the totality of the natural numbers [tuttii numerz],becauseboth of these are as numerousas their roots which constitute the totality of the natural numbers. Nextwe sawthat thelargerthe numberstakenthemoresparsely distributed werethe squares,and stillmore sparselythe cubes; thereforeit is clearthat the largerthe numbersto whichwepass the farther we recedefromthe infinitenumber;henceit follows that, sincethis processcarriesus fartherandfartherfromthe endsought, if on turningback we shallfindthat any number can be said to be infinite,it mustbe unity. Here indeedare satisfiedall those conditions hichare requisite w foran infinite number; meanthat unity contains itselfasmanysquares s I in a there arecubesandnatural numbers[tuttii numen]. SIMP.I donot quitegraspthemeaningofthis. SALV. There is no difficulty the matter becauseunity is at in once a square, a cube, a squareof a squareand all the other powers[dignity]; oris there any essentialpeculiarityin squares n or cubes whichdoesnot belongto unity; as, for example,the propertyof twosquarenumbersthat they havebetweenthema mean proportional;take any squarenumberyou pleaseas the first term and unity for the other,then youwill alwaysfind a numberwhichis a meanproportional. Consider twosquare the numbers, 9 and 4; then 3 is the mean proportionalbetween 9 and I ;while2is ameanproportional etween4and I; between b 9 and 4 we have6 as a mean proportional. A propertyof cubes is that they must have betweenthem two mean proportional numbers; take 8 and 27; betweenthem lie IZ and 18; while between

[8S]

38 THE TWO NEW SCIENCESOF-GALILEO between and8wehave2 and4 intervening; I andbetween and I 27therelie3 and 9. Therefore weconclude unity is the that onlyinfinite number.Thesearesome ofthemarvels which our imaginationannotgraspandwhichshould c warnus againsthe t serious error of thosewhoattemptto discussthe infiniteby assigning it the sameproperties hichwe employfor the to w finite,henatures fthetwohaving othingncommon. t o n i Withregard thissubjecCt to I musttellyouof a remarkable property hichjustnowoccurso me andwhich w t willexplain thevastalteration andchange fcharacCter a finitequano which tity wouldundergo passing infinity.Let us drawthe in to straight ineABofarbitrary l length andletthe pointC divide it into twounequal parts; thenI saythat,if pairsoflinesbe drawn, nefrom o eachof theterminal ointsA andB, andif p theratiobetween thelengths ftheselinesisthe same that o as between ACandCB,theirpoints ofinterse&ion willalllieupon the circumferenceoneandthe samecircle. Thus,for exof ample, ALandBLdrawnrom f AandB,meeting t thepointL, a beating oneanotherhe sameratioas ACto BC, andthe to t meetingat K also beatingto one another thesame ratio, and AK and BK likewise thepairs pair E M, BI,AH,BH,AG, BG, AF, BF, AE, BE,havetheirpoints ofintersecdon K, L,

[84]

_c

B-ii"-"""_

Fig.7 inguponthecircumferencefoneandthesamecircle. Accordingly imagine o ifwe thepointCtomovecontinuously insuch amanner thatthelines drawn fromittothefixed terminal oints, p Aand B,always ainm tainthesame ratiobetween theirlengths asexistsbetween the original parts, CandCB,thenthepoint will, s I shall resA C a p entlyprove, escribe d acircle.Andthecirclehusdescribed t will _crcase

I,H, ,F,E,all yG l

FIRST DAY 39 increasein sizewithoutlimitas the pointC approacheshemidt dlepoint whichwe may callO; but it willdiminishin sizeas C approachesthe endB. So'thatthe infinitenumberof pointslocatedinthe lineOBwill,ifthe motionbeasexplainedabove,describecirclesofeverysize,somesmallerthan thepupilofthe eye ofa flea,otherslargerthanthe celestial quator. Nowifwemove e anyof the pointslying betweenthe two endsO and B they will all describecircles,thosenearestO, immensecircles;but if we move the point O itself,and continueto moveit accordingto the aforesaidlaw, namely,that the linesdrawn from O to the terminalpoints,A andB, maintainthesameratioasthe original linesAOandOB,what kindof a linewillbe produced A circle ? willbe drawnlargerthan the largestof the others,a circlewhich isthereforeinfinite. But fromthepointOa straightlinewillalso be drawn perpendicularto BA and extendingto infinitywithout ever turning, as did the others, to join its last end with its first; for the point C, with its limitedmotion,havingdescribed the upper semi-circle,CHE, proceedsto describe the lower semicircle EMC, thus returning to the startingpoint. But the point O havingstarted to describeits circle,as did allthe other points in the lineAB, (for the points in the other portionOA describetheir circlesalso, the largest beingthose nearest the point O) is unable to return to its startingpoint becausethe circleit describes,beingthe largestof all, is infinite;in fact, it describesan infinitestraightlineas circumferencef its infinite o circle. Think nowwhat a differencehere isbetweena finiteand t an infinite circlesince the latter changescharacter in such a mannerthat it losesnot only its existence alsoits possibility but of existence;indeed,we alreadyclearlyunderstand that there can be nosuch thingas an infinitecircle;similarlythere can be no infinitesphere, no infinitebody, and no infinitesurfaceof anyshape. Nowwhat shallwesayconcerning thismetamorphosisin the transition fromfiniteto infinite? Andwhy shouldwe feel greater repugnance,seeing that, in our search after the infiniteamongnumberswe foundit in unity? Having broken up a solidinto many parts, havingreducedit to the finestof powder

[85]

4o THE TWO NEW SCIENCESOF GALILEO powder ndhaving a resolved intoitsinfinitelymallindivisible it s atomswhymaywenotsaythat thissolid hasbeenreduced a to single continuum [unsolo continuo] perhaps fluidlikewateror a mercury evena liquifled or metal? Anddowe notseestones meltintoglass andtheglass itselfunderstrong heatbecome ore m fluidthanwater? SAtin. rewethento believe substances A that becomeluidin f virtueof beingresolved their infinitely into smallindivisible components ? SALv. amnot ableto findanybettermeansofaccounting I for certainphenomena fwhichthe following o isone. WhenI takea hardsubstanceuchasstoneormetalandwhenI reduce s it by meansof a hammeror finefileto the mostminuteand impalpableowder, is clearthat itsfinestparticles, p it although whentakenoneby oneare,onaccount ftheirsmallness, o imperceptibleo our sightand touch,are neverthelessinitein t f size,possesshape,and capability s of*beingounted. It is also c truethatwhenonceheaped uptheyremain a heap;andif an in excavation emadewithinlimitsthecavitywillremain the b and surrounding articles p willnot rushin to fill it; if shakenthe particles ometo restimmediatelyftertheexternal isturbing c a d agentis removed;he sameeffecCts t areobserved allpilesof in larger ndlarger articles, a p ofanyshape, venif spherical, e asis thecasewithpilesofmillet,wheat, eadshot,andevery l other material. But if we attempt to discoveruchpropertiesn s i water wedonotfindthem; forwhenonce heaped it immeup diately flattens outunless eldupbysome h vessel rotherextero nalretainingody; b when hollowed quicklyushes outit r intofill the cavity;and whendisturbed flu&uates it fora longtime andsends utitswaves o throughreat istances. g d Seeingthat waterhas less firmnessconsistenza] the [ than finest fpowder, o infa_ hasnoconsistence whatever, wemay, it seemsto me, very reasonablyonclude the smallest c that particlesntowhlch canbe resolvedrequitedifferent i it a from finiteanddivisible articles;ndeedthe onlydifferenceam p i I abletodiscoversthattheformer i areindivisible. Theexquisite transparency

[86]

FIRST DAY 4I transparencyfwateralsofavorsthisview;forthemosttranso parentcrystal henbroken ground w and andreduced powder to loses itstransparency; thefinerthegrinding thegreater theloss; but in the caseof waterwherethe attritionis of the highest degreewehaveextremetransparency.Goldand silverwhen pulverized ithacids[acque w fortz]morefinelythanis possible with any filestill remainpowders,*nd donotbecome a fluids untilthe finest articles indivisibih] p [gl' offireorofthe raysof thesundissolve them,asI think,intotheirultimate,ndivisible, i andinfinitely small omponents. c SACR. Thisphenomenon lightwhichyoumentionis one of which havemanyti.mesemarked ithastonishment. have, I r w I for instance, eenleadmeltedinstantlybymeansofa concave s mirroronlythreehands[palmz]n diameter.HenceI think i that if themirrorwereverylarge,well-polished andofa parabolicfigure,t would as readily i just andquickly meltanyother metal,seeing thesmall irror,which that m wasnotwellpolished andhadonlya spherical shape, asablesoenergetically melt w to lead and burn everycombustibleubstance.Sucheffe&s s as theserendercredible me the marvels to accomplished the by mirrors ofArchimedes. SALV. Speaking the effe&sproduced the mirrorsof of by Archimedes,washisownbooks(which hadalready it I readand studied withinfinite astonishment) rendered redibleome that c t allthemiracles escribed yvarious riters.Andif anydoubt d b w hadremained bookwhichFatherBuonaventura the Cavalierit [87] has recentlypublishedon the subjecc the burningglass of [speccMo ustorfo] andwhich havereadwithadmiration ould I w haveremovedhelastdifficulty. t SAGR. alsohave seenthis treatiseand havereadit with I born atMilan died I598; atBologna a Jesuitather, tointro_647; f first duce theuse oflogarithms intoItalyandfirst toderive theexpression for thefocal lengthfalens aving o h unequal ofcurvature. "method radii His ofindivisibles"to bereckoned precursor theinfinitesimal is asa of
calculus. [Trans.] * It is not clear what Galileohere meansby sayingthat gold and silverwhentreated with acidsstillremainpowders. [Trans.] Oneofthe mostactiveinvestigators mong a Galileo's ontemporaries; c

42 THE TWO NEW SCIENCES OF GALILEO pleasureand astonishment;and knowingthe author I was confirmed in the opinionwhichI had alreadyformed of him that he was destinedto becomeone of the leadingmathematicians of our age. But now, with regard to the surprisingeffect of solarrays in meltingmetals,must webelievethat sucha furious adtionis devoidofmotionor that it is accompanied bythe most rapid ofmotions? SALv. e observethat other combustions resolutionsare W and accompanied motion, and that, the most rapid; note the acby tion of lightningand of powderas used in mines and petards; note alsohowthe charcoalflame,mixedas it is with heavy and impurevapors, increasesits power to liquify metalswhenever quickened a pair ofbellows. HenceI donot understandhow by the action oflight, althoughvery pure, can be devoidof motion and that ofthe swiftesttype. SAGR. of what kind and howgreat must we considerthis But speedof lightto be? Is it instantaneousor momentaryor does it likeother motionsrequiretime? Can we not decidethis by experiment? Sn_P.Everyday experienceshows that the propagation of light is instantaneous;for whenwe see a pieceof artilleryfired, at great distance,the flash reachesour eyes without lapse of time; but the sound reaches the ear only after a noticeable interval. SAGR. Well, Simplicio,he only thing I am able to inferfrom t this familiarbit of experienceis that sound, in reachingour ear,travelsmoreslowlythanlight;it doesnot informmewhether the comingof the light is instantaneousor whether, although extremelyrapid, it still occupiestime. An observationof this kind tells us nothingmorethan one in whichit is claimedthat "As soon as the sun reachesthe horizonits light reachesour eyes"; but whowill assureme that these rays had not reached this limitearlierthan they reachedour vision? SAT.v. small conclusiveness f these and other similar The o observations onceledmeto devisea methodbywhichonemight accuratelyascertainwhetherillumination,i. e.,the propagation of light, is really instantaneous. The fac2that the speed of sound

FIRST DAY

soundis as high as it is, assures that the motionof light us cannotailto beextraordinarily Theexperiment I f swift. which devised wasasfollows: Leteach oftwopersonsakea lightcontained lantern, t ina or otherreceptacle, thatbythe interposition such ofthehand, he t onecanshutoffor admitthelightto the vision theother. of Nextlet themstandoppositeach e otherat a distancefa few o cubits andpracdce untiltheyacquireuch s skillinuncovering and occulting theirlights thattheinstant onesees thelightofhis companion willuncover isown. Aftera fewtrialsthe he h response illbe soprompt hat without ensiblerror[svario] w t s e theuncovering ofonelightis immediately followedy theunb covering oftheother,sothatassoon asoneexposes hislighthe willinstantly seethatoftheother. Having acquired skillatthis shortdistanceet the two experimenters, l equippeds before, a takeuppositions separated a distanceftwoor threemiles by o andletthemperform thesame experiment atnight, oting n carefullywhetherheexposures t andoccultations occurinthe same manner sat short istances; a d iftheydo,wemaysafely conclude that the propagation lightis instantaneous; timeis of butif required a distance threemileswhich,considering at of the goingofonelightandthe comingftheother,reallyamounts o to six,thenthe delayoughtto be easilyobservable. f the I experiment to be madeat stillgreater is distances, sayeight or ten miles,telescopes maybe employed, eachobserver da justingoneforhimself t the place a whereheis to makethe experiment night; at thenalthough thelightsarenotlargeand aretherefore invisibleo the naked t eyeat sogreata distance, theycanreadily becovered anduncovered byaidofthe since telescopes, onceadjusted andfixed,they willbecome asily e visible. SAGa. Thisexperiment strikes measa clever ndreliablena i vention.But tellus whatyouconcluderomtheresults. f SALV. fadt I havetriedtheexperiment In onlyat a short distance,essthan a mile,fromwhich havenotbeenableto l I ascertain with certaintywhetherhe appearance the o19_ t of posite

[88]

43

44 THE TWO NEW SCIENCES OF GALILEO ix)sitelight was instantaneousor not; but if not instantaneous it is extraordinarilyrapid--I shoiald it momentary;and for call the present I shouldcompareit to motionwhichwe see in the lightningflashbetweencloudseightor ten milesdistant fromus. We see the beginningof this light--I might say its head and [89] source--locatedat a particular placeamongthe clouds;but it immediatelyspreadsto the surrounding ones,whichseemsto be an argumentthat at least sometime is requiredforpropagation; for if the illuminationwere instantaneousand not gradual,we should not be able to distinguishits orlgin--its center, so to speak--from its outlyingportions. What a sea we are gradually slippinginto without knowingit! With vacua and infinities and indivisiblesand instantaneousmotions, shall we everbe able,evenby meansof a thousanddiscussions,o reach t dry land? SAGR. Really these matters lie far beyondour grasp. Just think; whenwe seek the infiniteamongnumberswe find it in unity; that whichis ever divisibleis derived from indivisibles; the vacuum is found inseparablyconnectedwith the plenum; indeedthe viewscommonly held concerning nature of these the matters are so reversedthat even the circumference f a circle o turns out to be an infinite straight line, a fact which, if my memory servesme correctly,you, Salviati,were intending to demonstrate geometrically. Please thereforeproceedwithout furtherdigression. SAr.V. am at your service;but for the sakeof greater clearI nesslet mefirstdemonstratethe following problem: Given a straightlinedividedinto unequalparts whichbear to eachother any ratio whatever,to describea circlesuch that two straight lines drawn from the ends of the given line to any point on the circumference bear to each will other the sameratio asthe twoparts of the givenline,thus makingthoselineswhichare drawnfromthe sameterminal pointshomologous. LetABrepresentthe givenstraight linedividedinto any two unequalparts by the point C; the problemis to describea circle such

FIRST DAY 45 suchthat twostraightlinesdrawn fromthe terminal points, A andB, to anypointon thecircumference willbearto each other thesameratioasthepartACbears BC,sothat lines to drawn thesameterminal ointsarehomologous. from p About C as center describe circle a havinghe shorter artCB ofthe t p given asradius.Through draw straightineADwhich line, A a l [90] shallbe tangentto the circleat D and indefinitely prolonged towardE. Drawthe radiusCDwhichwillbe perpendicular to AE. At ]3erecta perpendicular AB;this perpendicular to willintersect E at A somepointsincethe angleat A is acute; callthis pointof intersecCtion a nd E, fromit drawa per(I pendicular to AE which willintersecCt AB prolonged F. in NowI say the two straightlinesFE and FC are equal. For _I""_L if we join E and C, we shall have two Fig.8 triangles, ECand BEC,in whichthe twosidesof the one, D DE and EC, are equal to the two sidesof the other,BE and EC, both DE and EB beingtangentsto the circleDB whilethe basesDC and CB are likewise equal;hencethe two angles,DEC and BEC, willbe equal. Nowsincethe angleBCEdiffersroma rightanglebytheangle f CEB,andthe angleCEFalsod_ffers froma rightangle bytheangle CED,and sincethesedifferences areequal,it followshat the angleFCE t is equalto CEF;consequently thesidesFE and FCare equal. If wedescribe circle a withF as centerandFE asradius will it

passhrough t thepoint C;letCEG besuch circle. hissthe a T i

circle sought, orifwedrawlines f fromtheterminal pointsAand B to anypointonitscircumference theywillbearto eachother the

46 THE TWO NEW SCIENCESOF GALILEO thesameratioasthetwoportions CandBCwhich A meetat the pointC. Thisismanifestn the caseof the twolinesAE and i BE,meeting thepointE, becauseheangleE ofthe triangle at t AEBisbise_edbythelineCE,andtherefore C:CB=AE:BE. A Thesame maybeproved ofthetwolines AGandBGterminatingin thepointG. For sincethe triangles FE andEFBare A similar, e haveAF:FE=EF:FB, or AF:FC=CF:FB, and w dividendo C:CF=CB:BF, or AC:FG=CB:BF; also comA ponendo wehavebothAB:BG=CB:BFandAG:GB=CF:FB =AE:EB=AC:BC. Q.E.D. [9_] Take now any other point in the circumference, H, say wherethe twolinesAH and BH intersect;n likemanner e i w shallhave AC:CB=AH:I-lB.Prolong untilit meetsthe HB circumference I and join IF; and sincewe have already at foundthat AB:BG=CB:BF it followsthat the recCtangle AB.BF equalto therecCtangle is CB.BG orIB.BH. Hence AB: BH--IB:BF. But the anglesat B are equaland therefore AH:HB=IF:FB=EF:FB=AE:EB. Besides, mayadd,that it isimpossible I forlines which maintain this sameratioand whichare drawnfromthe terminal points, andB,to meetat anypointeitherinside routside A o the circle,CEG. Forsupposehiswere t possible;etALandBLbe l two suchlinesintersecdng the pointL outsidethe circle: at prolong LBtill it meetsthe circumference M andjoinMF. at If AL:BL=AC:BC=MF:FB,then we shallhave two trianglesAI.R and MFB whichhave the sidesaboutthe two anglesproportional, anglesat the vertex, ,equal,andthe the B two remaining angles,FMB and LAB,lessthanright angles (because therightangle atM hasforitsbasetheentirediameter CGandnotmerely partBF: andtheotherangle thepoint a at AisacutebecausehelineAL,thehomologue t ofAC,isgreater thanBL,the homologue ofBC). Fromthisit followshat the t triangles ABLandMBF are similarand therefore AB:BL= MB:BF, makingthe recCtangle AB.BF=MB.BL; ut it has b beendemonstrated therecCtangle isequaltoCB.BG; that AB.BF whence would it follow thereeCtangle that MB.BL isequalto the rectangle

FIRST DAY 47 recCtangle CB.BG whichisimpossible; therefore theintersecCtion cannot alloutsidethecircle.Andinlikemanner ecanshow f w thatit cannot allinside; ence f h alltheseintersections fallonthe circumference. But nowit istimeforus to gobackandgranttherequest f o Simplicio showingimthat it is notonlynotimpossible by h to resolve lineintoan infinite a number fpointsbut that thisis o quiteaseasyasto divideit intoitsfinite parts. ThisI willdo underthe followingondition hichI am sure,Simplicio, c w you willnot denyme,namely,hat youwillnot require t meto separatethe points,one fromthe other,and show themto you, one by one,on this paper;for I should contentthat you, be withoutseparating fouror sixpartsof a linefromoneanthe other,should show methemarkeddivisions mostthatyou orat should foldthemat angles forming square a hexagon: a or for, then, I am certainyou wouldconsiderhe division istin_ly t d andactually accomplished. S_P. I certainlyhould. s SALv. nowthechange If whichtakesplace whenyoubenda lineat angles soasto formnowa square, nowanoctagon, nowa polygon forty,a hundredor a thousand of angles,s sufficient i to bring into acCtuality four,eight, forty,hundred,and the thousand partswhich,according you,existedat firstonly to potentially thestraightline,mayI notsay,withequal ight, in r that,whenI havebentthestraightlineintoa polygonaving h an infinite number f sides,i. e., intoa circle,I havereduced o to actualitythat infinite number fpartswhich o youclaimed, while itwasstraight,werecontainedn it onlypotentially Norcan i ? one denythat thedivision intoan infinite number fpointsis o justastruly accomplished theoneintofourpartswhenthe as squareisformed intoa thousand artswhenthemillagons or p i formed;orinsucha divisionhesame f t conditions aresatisfieds a in the caseof a polygon thousand a hundred ofa or thousand sides. Sucha polygon laidupona straightlinetouches itwith oneof its sides,i. e.,withoneof its hundredthousand parts; while thecircle whichisa polygon f aninfinite o number fsides o touches

[92]

48 THE TWO NEW SCIENCES OF GALILEO touches the same straight linewith one of its sideswhich is a singlepoint differentfrom all its neighborsand thereforeseparate and distincCt no lessdegreethan is onesideof a polygon in fromthe other sides. Andjust as a polygon,whenrolledalong a plane,marks out upon this plane,by the successive contac% of its sides,a straight lineequal to its perimeter,so the circle rolleduponsuch a plane alsotracesby its infinitesuccession of contadtsa straightlineequal in lengthto its owncircumference. I amwilling,Sirnplicio, t the outset,to grantto thePeripatetics a the truth of their opinionthat a continuousquantity [if continuo]is divisibleonlyinto parts whichare stillfurther divisible so that howeverfar the divisionand subdivision e continuedno b end will be reached; but I am not so certain that they will concedeto me that noneof these divisionsof theirs can be a finalone, as is surely the facCt, ecausethere alwaysremains b "another"; the final and ultimate divisionis rather one which resolvesa continuousquantity into an infinitenumber of indivisiblequantities,a resultwhichI grant can neverbe reached by successive divisioninto an ever-increasing umber of parts. n But if they employthe methodwhichI proposefor separating [931 and resolvingthe wholeof infinity[tuttala infini_], at a single stroke (an artificewhich surely ought not to be deniedme), I think that theywouldbe contentedto admitthat a continuous quantity is built up out of absolutelyindivisibleatoms, especially since this method, perhaps better than any other, enablesus to avoidmany intricate labyrinths,suchas cohesion in solids,alreadymentioned,and the questionof expansionand contra_ion,withoutforcinguponus the objectionable admission ofemptyspaces[insolids] hichcarrieswith it the penetrability w of bodies. BothoftheseobjecCfions,appearsto me, areavoided it if we accept the above-mentionedview of indivisibleconstituents. Stop. I hardly knowwhat the Peripateticswouldsay since the viewsadvancedby youwouldstrikethemasmostlynew,and as suchwemust considerthem. It ishowevernot unlikelythat they wouldfindanswersand solutionsfor theseproblemswhich

I,

FIRST DAY 49 I, forwantoftimeandcritical bility, mat presentunable a a to solve. Leaving to onesideforthemoment, should this I like to hearhowthe introduction these indivisibleuantities of q helpsus to understand contraction nd expansionvoiding a a at thesametimethevacuum andthepenetrability ofbodies. Sacg.I alsoshalllisten withkeeninterest othissame t matter whichis farfromclearin mymind;provided amallowed I to hearwhat,a moment go,Simplicio a suggested eomit,namely, w the reasons whichAristotle offers againstthe existence f the o vacuum andthearguments hich w youmustadvance inrebuttal. SALv. willdoboth. Andfirst,just as,forthe production I of expansion, employ linedescribed y the smallcircle we the b duringone rotationof the largeone--alinegreaterthan the circumference the smallcircle--so,n orderto explain of i contraction, epointout that,duringeachrotation w ofthe smaller circle,the largeronedescribes straightlinewhichis shorter a thanitscircumference. For thebetterunderstanding thiswe proceed the conof to siderationfwhathappensn thecaseofpolygons.Employing o i [941 a figure similar the earlierone,construcCt to thetwohexagons, ABCandHIK, aboutthe common centerL, andlet themroll along theparallelines l HOMandABc. Now holding thevertex I fixed,allowthe smaller polygon rotateuntilthe sideIK to liesuponthe parallel,duringwhichmotionthe pointK will describe arcKM, and the sideKI willcoincide ithIM. the w Let us seewhat,in the meantime, sideCB of the larger the PheOlygon hasbeendoing. Since therotation isaboutthepointI, terminalpointB, of the lineIB, moving backwards, will describe thearcBbunderneathheparallel Asothat whenthe t c sideKI coincides withthelineMI,theside BCwillcoincide with bc,havingadvanced onlythrough distance the Bc,but having retreatedthrough portionofthe lineBAwhichsubtendshe a t arcBb.Ifweallow therotation ofthesmaller olygono goon p t it willtraverse describelong parallel lineequalto its and a its a perimeter; whilethe larger newilltraverse describe line o and a lessthanitsperimeter yasmanytimesthelength b bBasthere are

5o THE TWO NEW SCIENCES OF GALILEO are sides lessone; this line is approximatelyequal to that describedby the smallerpolygonexceeding onlyby the distance it bB. Here now we see, without any difficulty,why the larger polygon, hencarried by w the smaller, d oe s not measureoffwith its sides a line longer than that traversed by the smaller one;this isbecausea portion ofeach sideis superposed upon its immediately precedingneighbor. ......" " ' Let us next consider twocircles,havinga common center at A, and lyparallels,the smallerbeing tangentto its parallel 2k at the point B; the larger, at thepoint C.Herewhen the smallcirclecommences upontheir respedtive I_ I ing to roll the point B does not remain at rest {:: "for a whileso as to allow Fig.9 BC to move backward and carry with it the point C, as happenedin the caseof the polygons,where the point I remainedfixeduntil the sideKI coincidedwith MI and the lineIB carriedthe terminalpoint B backwardas far asb, so that the sideBCfelluponbc,thus superposinguponthe lineBA, the portion Bb,and advancingby an amount Bc, equal to MI, that is, to one sideof the smaller polygon. On account of these superpositions,which are the excesses f the sidesof the largerover the smallerpolygon,each o net advanceis equalto onesideof the smallerpolygonand, during one completerotation, theseamountto a straightline equal in lengthto the perimeterof the smallerpolygon. But

FIRST DAY 5I But nowreasoningn the samewayconcerning circles, i the wemustobserve whereashenumber fsidesinany polythat t o gonis comprised withina certain limit,thenumber fsidesina o circle infinite;he formerare finiteanddivisible; latter is t the infinite andindivisible. n thecaseofthepolygon, I thevertices remainat rest duringan intervalof timewhichbearsto the periodofone completeotation sameratiowhich r the oneside bearsto the perimeter;ikewise, the caseof the circles, l in the delayof eachof the infinite numberof verticess merely i instantaneous, becausean instantis sucha fracCtion a finite of interval sa pointisofa linewhichcontainsninfinite umber a a n ofpoints. Theretrogression ofthesidesofthelarger olygons p i not equalto the lengthofone of its sidesbut merely the to excess f sucha sideoverone sideof the smaller olygon, o p the netadvance eing b equalo thissmalleride; utinthecircle, t s b the pointorsideC,during theinstantaneous restofB,recedes yan b amount qualto itsexcess verthesideB,making netprogress e o a equalto B itself. In shortthe infinite numberof indivisible sides ofthegreater circle withtheirinfinite umber findivisible n o retrogressions, madeduring theinfinite umber finstantaneous n o delaysof the infinitenumber f vertices f the smaller o o circle, togetherwiththe infinite number f progressions, o equalto the infinite numberof sidesin the smaller circleall these, say, I add up to.a lineequalto that described y the smaller ircle, b c a line whichcontainsan infinitenumberof infinitely small superpositions, thusbringing abouta thickeningr contracCtion o withoutany overlapping r interpenetration finiteparts. o of Thisresultcouldnot beobtained thecaseofa linedivided in [9 6] intofinitepartssuchas istheperimeter fanypolygon, o which whenlaldoutin a straightlinecannot e shortenedxceptby b e theoverlappingndinterpenetration a ofitssides.Thiscontractionofan infinite numberofinfinitelymallpartswithoutthe s interpenetration oroverlapping offinite partsandthepreviously mentioned [p.7o,Nat. Ed.]expansion an infinite of number f o indivisibleartsby the interposition indivisible p of vacuais,in myopinion, themost hat canbesaidconcerning t thecontracdon and

52 THE TWO NEW SCIENCES OF GALILEO and rarefacCtion bodies,unlesswe giveup the impenetrability of ofmatter and introduceemptyspacesof finitesize. If youfind anything herethat you considerworthwhile,pray use it; ifnot regard it, together with my remarks, as idle talk; but this remember,we are dealingwith the infiniteand the indivisible. SAGR. frankly confessthat your idea is subtle and that it I impressesme as new and strange; but whether, as a matter of facet,nature actually behaves accordingto such a law I am unable to determine;however,until I find a more satisfacCtory explanationI shall hold fast to this one. Perhaps Sirnplicio can tell us somethingwhichI have not yet heard, namely,how to explainthe explanationwhich the philosophers have given of this abstruse matter; for, indeed,all that I have hitherto read concerning contracCtion so denseand that concerningexis pansionso thin that my poor brain can neither penetrate the formernorgraspthe latter. SIMP.I am all at sea and finddiflicultlesin followingeither path, especiallythis newone; becauseaccordingto this theory an ounce of gold might be rarefiedand expandeduntil its size wouldexceedthat of theearth, whilethe earth, in turn, mightbe condensedand reduceduntil it wouldbecomesmallerthan a walnut, somethingwhichI donot believe;nor doI believethat youbelieve it. The argumentsand demonstrations whichyou have advanced are mathematical,abstracCt, nd far removed a fromconcretematter; and I do not believethat whenappliedto the physicaland natural worldtheselawswillhold. SALV. am not able to render the invisiblevisible,nor do I I think that youwillask this. But nowthat youmentiongold, donot our sensestell us that that metal can be immenselyexpanded? I donot knowwhetheryouhaveobservedthe method [97] employed thosewhoare skilledin drawinggoldwire,ofwhich by reallyonly the surfaceis gold,the insidematerialbeing silver. The way they draw it is as follows:they take a cylinderor, if you please,a rod of silver,about half a cubitlong and three or four times as wide as one's _daumb;his rod they coverwith t gold-leaf whichis so thin that it almostfloatsin air, puttingon not

FIRST DAY 53 notmorethaneightorten thicknesses. gilded Once theybegin to pullit, withgreatforce, hroughtheholesofa draw-plate; t againandagainit ismadeto passthrough smaller nd smaller a holes,until, aftervery many passages, is reducedto the it fineness fa lady's o hair,or perhaps evenfiner;yet the surface remains ilded.Imagine owhowthesubstancefthisgoldhas g n o beenexpandedndtowhatfineness hasbeenreduced. a it Sire,.I do not seethat this process wouldproduce,as a consequence, marvelloushinning the substance fthe that t of o goldwhichyousuggest: irst,becauseheoriginal ilding f t g consisting oftenlayers fgold-leaf sensible o hasa thickness; secondly, because drawing in outthe silver grows length at the it in but sametime diminishes roportionally thickness; nd, since p in a onedimension thuscompensates theother,theareawillnotbe soincreasedsto makeitnecessaryuring a d theprocess fgilding o to reducethe thinness the goldbeyond oftheoriginal of that leaves. SALV. are greatlymistaken, implicio, You S becausehe surt faceincreases irectly the square d as rootof the length, facCt a which candemonstrate I geometrically. SAGR. Pleasegiveus the demonstration notonlyformyown sakebut alsofor Simpliciorovided thinkwe canunderp you standit. SALV. seeif I can recallit on the spurof the moment. I'll .Attheoutset,it is clear that theoriginal thickrodofsilver nd a the wiredrawn to an enormous out lengtharetwocylinders of the samevolume, incetheyare the samebodyof silver. So s [981 that,if I determineheratiobetween t thesurfaces fcylindersf o o the samevolume,heproblem t willbesolved.I saythen, The areasof cylinders f equalvolumes, eglectinghe o n t bases,bearto eachothera ratiowhichisthesquareroot oftheratiooftheirlengths. Taketwo cylinders f equalvolume o havingthe altitudesAB andCD,between which thelineE isa meanproportional. Then I claimthat,omitting thebasesofeachcylinder, surface the of the cylinder Bis to that ofthecylinder Dasthelength is A C AB

54 THE TWO NEW SCIENCES OF GALILEO is to the lineE, that is,as the squarerootofAB isto the square root of CD. Now cut off the cylinderAB at F so that the altitude AF is equal to CD. Then sincethe basesof cylindersof equal volume bear to one another the inverse ratio of their heights, it followsthat the area of the circular base of the cylinderCD willbe to the area of the circularbaseof AB as the altitude AistoDC:moreover,sincecirclesare to one another B as the squares of their diameters,the said squares will be to eachother as BA is to CD. But BA is to CD as the squareof E BA is to the squareof E: and, therefore,these _-_ 1 four their sides; form lineAB is to E as the I wise squares willso the a proportion;and likediameter of circle C is to the diameter of the circleA. But the diameters are proportional to the circumferences and the circumferences are proportionalto the areas of cylinders of "-F equalheight; hencethe lineAB is to E as the surfaceof the cylinder CD is to the surfaceof the cylinderAF. Nowsincethe heightAF is to AB asthe surfaceofAF is to the surfaceof AB; and sincethe heightAB is to the line E as the surface CD is to AF, it follows,ex _equaliin Fig.io proportione perturbata,*hat the height AF is t to E as the surfaceCD is to the surfaceAB, and con$ertendo, the surface of the cylinderAB is to the surface of the cylinder CD as the line E is to AF, i. e., to CD, or as AB is to E which is the squareroot of the ratio of AB to CD. q.E.D. If nowwe apply these resultsto the casein hand, and assume that the silver cylinder at the time of gildinghad a length of only half a cubit and a thicknessthree or four times that of [99] one's thumb,we shallfindthat, whenthe wirehasbeen reduced to the finenessof a hair and has been drawnout to a length of twenty thousand cubits (and perhaps more), the area of its surface will have been increased not less than two hundred times. Consequentlythe ten leavesof goldwhichwerelaid on *See Euclid, Book V,Def.2o., odhunter's T Ed.,p.I37(London, I877.) [Trans.]

FIRST DAY 55 have been extendedover a surfacetwohundred timesgreater, assuringus that the thicknessofthe goldwhichnowcoversthe surfaceof so many cubits of wire cannot be greater than one twentieth that of an ordinary leaf of beaten gold. Consider nowwhat degreeof fineness musthave andwhetherone could it conceiveit to happen in any other way than by enormousexpansionof parts; consideralsowhetherthisexperimentdoesnot suggestthat physicalbodies [rnaterie fisiche]are composedof infinitelysmall indivisible particles,a viewwhichis supported byothermorestrikingand conclusive xamples. e SACR. This demonstrationis so beautifulthat, evenif it does not have the cogencyoriginally intended,--althoughto my mind, it is very forceful--the short time devoted to it has nevertheless beenmosthappilyspent. SAT.v. Sinceyou are so fondof thesegeometrical emonstrad tions, which carry with them distinctgain, I will give you a companiontheorem which answersan extremelyinteresting query. We have seen abovewhat relationshold betweenequal cylindersof differentheightor length;let us nowseewhat holds when the cylindersare equal in area but unequal in height, understandingarea to includethe curvedsurface,but not the upper andlowerbases. The theoremis: The volumesof right cylindershavingequal curvedsurfacesareinverselyproportionalto theiraltitudes. Let the surfacesof the twocylinders, lEand CF, be equalbut A let the height of the latter, CD, be greater than that of the former,AB: then I say that the volumeof the cylinderAE is to that of the cylinderCF as the heightCD is to AB. Now since the surface of CF is equal to the surfaceof AE, it followsthat the volumeof CF is lessthan that of_ALE; they for,if wereequal,the surfaceof CF would,by the preceding proposition, exceedthat of AE, and the excesswouldbe so muchthe greaterif the volumeof the cylinderCF weregreaterthan that ofAE. Let us nowtake a cylinderID havinga volumeequalto that o_?_; then, accordingto the precedingtheorem,the surface of the cylinder'ID is to the surfaceof _ as the altitude

[ ool

IF

56 THE TWO NEW SCIENCESOF GALILEO IF isto themeanproportional between andAB. But since IF one datumof the problemis that the surface AE is equal of to that of CF, and sincethe surfaceID is to the surfaceCF asthe altitudeIF isto thealtitudeCD,it followshat CDisa t mean proportional betweenIF and AB. ""i Not onlyso,but sincethe volumeof the cylinderID is equalto that ofAE, each willbearthe sameratioto the volume of the cylinder F;but the volume is to C ID the volume asthealtitudeIF is to the CF altitudeCD; hencethe volumeof AE is c to the volume CF as the lengthIF is of to thelengthCD,thatis,asthelengthCD A is to thelengthAB. q.E.D. Thisexplains phenomenon a uponwhich Ithe commonpeople alwayslook with Iwonder, amely,ifwehavea piece n ofstuff !which hasone sidelongerthanthe other, !wecanmakefromit a cornsack, usingthe E B ]customary woodenbase,whichwilthold I . . m Fimorewhenthe shortsldeof the clothis usedfortheheightofthesackandthelong Fig.n sideis wrappedaroundthe wooden base, thanwith the alternative arrangement.Sothat, forinstance, froma piece ofclothwhichissixcubitsononesideand twelve ontheother,a sackcanbemadewhich willholdmorewhenthe sideoftwelve cubitsis wrapped aroundthewooden base,leaving the sacksix cubitshighthan whenthe six cubitside is putaroundthebasemaking sacktwelve the cubits high. From whathasbeenprovenabove welearnnotonlythegenerala_ f that onesackholds morethantheother,butwealsogetspecific and particularinformation to how muchmore,namely, as just in proportion the altitudeof the sackdiminisheshe as t contentsincreaseand viceversa. Thus if we use the figures givenwhich makethe clothtwiceaslongaswideandif weuse the longsideforthe seam,the volume the sackwillbejust of one-half sgreatas withthe opposite a arrangement.Likewise if

FIRST DAY

if we have a pieceof mattingwhichmeasures x 25cubits and 7 makefromit a basket,the contentsof thebasketwill,whenthe seamislengthwise, e sevenas compared b with twenty-five when the seamrunsendwise. SAGI_. is with great pleasurethat we continuethus to acIt quire new and usefulinformation. But as regardsthe subjecCt just discussed,I really believethat, amongthosewho are not alreadyfamiliarwith geometry, ouwouldscarcely y findfourpersonsin a hundredwhowouldnot,at firstsight,makethemistake ofbelievingthat bodieshavingequalsurfaceswouldbe equalin other respecCts. peaking areas,the sameerror ismadewhen S of oneattempts,as oftenhappens,to determinethe sizesofvarious cities by measuringtheir boundary lines, forgettingthat the circuit of one may be equalto the circuitof another whilethe area of the one is much greater than that of the other. And this is true not only in the caseof irregular,but alsoof regular surfaces,,wherethe polygonhavingthe greaternumberof sides alwayscontainsa largerarea than the onewith the lessnumber of sides,so that finallythe circlewhich is a polygonof an infinitenumberof sidescontainsthe largestarea ofallpolygons of equal perimeter. I rememberwith particularpleasurehaving seen this demonstrationwhen I was studying the sphere of Sacrobosco *with the aidofa learnedcommentary. SALV. Very true! I too cameacrossthe samepassagewhich suggestedto me a method of showinghow, by a singleshort demonstration,one can prove that the circlehas the largest content of all regular isoperimetric figures;and that, of other figures,the onewhichhasthe largernumberof sidescontainsa greater areathan that whichhasthe smallernumber. SAGR. Beingexceedinglyondofchoice f anduncommon propositions,I beseechyouto letus haveyourdemonstration. SALV. I can do this in a few wordsby provingthe following theorem: The area of a circleis a mean proportionalbetweenany Seeinterestingiographical b noteon Sacrobosco [JohnHolywood] inEncy.grit.,Ilth Ed. [Trans.] ,

[ioI]

57

[io ]

58

THE TWO NEW SCIENCES OF GALILEO two regularand similarpolygons f whichone circumo scribes andtheotherisisoperimetric it withit. Inaddition, the areaofthecircle islessthanthat ofanycircumscribed polygon andgreater thanthat ofanyisoperimetric polygon. Andfurther,ofthesecircumscribed polygons, theonewhich hasthegreater numberofidesissmallerhantheonewhich s t has a lessnumber; ut, on the otherhand,that isoperib metricpolygon whichhas the greaternumberof sidesis the larger. LetA andB betwosimilar olygons p ofwhich circumscribes A thegivencircle B isisoperimetric and withit. Theareaofthe circle willthenbea meanproportional between areasofthe the polygons.For ifweindicate theradiusofthecircle byAC and ifwe rememberhat the areaof thecircleisequaltothat ofa t right-angled triangle whichoneof the sidesaboutthe right in angleis equalto the radius,AC, and the otherto the circumference; ndif likewise erememberhat theareaofthe polya w t gonA is equalto the area of a right-angledriangle of t one whose sidesabouttherightanglehasthesamelength asACand theotherisequalto theperimeter ofthepolygon itself;it isthen
1

[io31

o Fig.IZ manifesthat the circumscribed t polygon bearsto the circlethe sameratiowhichitsperimeter earsto the circumference b ofthe circle, r to theperimeter fthepolygon which o o B is,byhypothesis,equalto the circumference the circle. But sincethe of polygons andB aresimilarheirareasareto eachotherasthe A t squaresoftheirperimeters; hencethe areaofthe circle is a A mean

FIRST DAY 59 meanproportional etweenthe areasofthe twopolygonsA and b B. And sincethe area of the polygonA is greaterthan that of the circleA, it is clearthat the area of the circleA is greater than that of the isoperimetric polygonB, and is thereforethe greatest of all regular polygonshavingthe same perimeteras the circle. We now demonstratethe remainingportion of the theorem, whichis to prove that, in the caseof polygonscircumscribing a givencircle,the one havingthe smallernumberof sideshas a larger area than one having a greater numberof sides; but that on the other hand, in the caseof isoperimetric polygons, the one having the more sides has a larger area than the one with lesssides. To the circlewhich has O for center and OA for radius draw the tangent AD; and on this tangent lay off, say, AD whichshallrepresentone-halfof the sideof a circumscribedpentagonand AC whichshall representone-halfof the sideofa heptagon;draw the straightlinesOC_ and OFD; then with O as a center and OC as radius draw the arcECI. Now sincethe triangleDOCis greaterthan the sectorEOCand since the sector COI is greaterthan the triangleCOAL, followsthat it the triangleDOCbearsto the triangleCO.A_greaterratiothan a the sectorEOCbearsto the secCtor COI,that is, than thesector - FOG bears to the secCtor GOA. Hence, componendo t pere mutando,the triangleDOA bears to the secCtor FOAa greater ratio than that which the triangle COA bears to the sector GOA,and also IOsuch trianglesDOA bear to IOsuch secCtors FOAa greaterratio than 14suchtrianglesCOAbear to 14such sectorsGOA,that is to say, the circumscribed pentagonbears to the circlea greater ratio than doesthe heptagon. Hencethe pentagonexceeds heptagonin area. the But nowlet us assumethat both the heptagoilandthe pentagonhave the sameperimeteras that of a givencircle. Then I say the heptagonwillcontaina largerarea than the pentagon. For sincethe area of the circleis a mean proportionalbetween areas of the circumscribed of the isoperimetric and pentagons, and since likewiseit is a mean proportionalbetween the circumscribed

[ o4]

60 THE TWO NEW SCIENCESOF GALILEO cumscribedndisoperimetric a heptagons, andsince alsowehave provedthat the circumscribedentagonis largerthan the p circumscribed heptagon,t followshat this circumscribed i t pentagonbearsto the circlea largerratiothandoesthe heptagon, that is, the circlewill bear to its isoperimetric pentagona greaterratio than to its isoperimetric heptagon. Hencethe pentagon smallerhan its isoperimetric is t heptagon.Q._. D. SAGR. Averycleverandelegant emonstration! howdid d But wecome plunge to intogeometry hile w discussing theobjections urgedbySimplicio, objecCtions ofgreatmoment, specially e that onereferringodensity t which strikes easparticularlyifficult m d ? SALV. contracCtion expansion If and [condensazione e rarefa_ione]consist n contrary i motions, oneoughtto findforeach greatexpansion correspondingly a largecontracCtion. our But surprise increased hen,everyday,weseeenormous is w expansions taking pLce almostinstantaneously.Think what a tremendous xpansion ccurswhena smallquantityof gune o powderflaresup intoa vastvolumeoffire! Thinktooofthe almost limitlessexpansion the light whichit produces! of ImaginethecontracCtion whichwould takeplaceif thisfireand this lightwereto reunite, hich, ndeed,snotimpossible w i i since onlya lktle whileagotheywerelocatedtogetherin this small space. Youwillfind, ponobservation,thousand u a suchexpansionsfor theyare moreobvious than contracCtions sincedense matterismorepalpable accessible our senses.Wecan and to takewoodand seeit goup infireand light,but wedonotsee themrecombine formwood;weseefruitsandflowers nd a to a thousand othersolidbodiesdissolveargelyintoodors,but we l do not observe thesefragrantatomscoming togetherto form fragrantsolids. But wherethe sensesfailus reason muststep in; forit willenable usto understand themotion involvednthe i condensation extremelyarefied tenuous of r and substances just as clearlyas that involved the expansionnd dissolution_ in a o solids.Moreover earetryingto findouthowit ispossibleo w t produce expansionnd contracCtionbodies a in whicharecapable of suchchanges ithoutintroducingacuaandwithoutgiving w v up

[IoSl

FIRST DAY 6I up the impenetrability ofmatter;but thisdoesnotexcludehe t possibility ftherebeingmaterials hichpossess osuchpropo w n erties and do not, therefore, arry with them consequences c which you call inconvenient nd impossible.And finally, a Simplicio,have,forthe sakeofyouphilosophers, I takenpains to findan explanationf howexpansionnd contra_ioncan o a takeplace without uradmitting o thepenetrability ofmatterand introducing acua,properties v whichyou denyand dislike; f i youwere admitthem,I should to notoppose ousovigorously. y NoweitheradmitthesediflSculties acceptmyviews sugor or gestsomethingetter. b SACR. quite agree with the peripatetic I philosophersn k denying penetrabilityfmatter. Asto thevacuaI should the o liketo heara thorough discussion Aristotle's emonstration of d inwhich heopposes them,andwhatyou,Salviati, aveto sayin h reply. I beg of you, Simplicio, you giveus the precise that proofof the Philosophernd that you, Salviati,giveus the a reply. SnaP.So far as I remember, Aristotle inveighs againstthe ancientviewthat a vacuumis a necessary prerequisiteor f motionand that the lattercouldnotoccurwithouttheformer. In oppositiono this viewAristotle t showsthat it is precisely the phenomenon f motion,as we shall see, whichrenders o untenable ideaof a vacuum. Hismethodisto dividethe the argumentntotwoparts. He firstsupposes i bodies ofdifferent weightsomove t inthesame medium; thensupposes, oneandthe samebodyto movein different edia. In the first case,he m [_o61 supposes bodies ofdifferent weighto move t inoneandthesame medium withdifferent peeds hich s w standto oneanothern the i sameratioastheweights; othat,forexample, bodywhichis s a tentimesasheavyas another illmove w tentimesasrapidlyas theother. In thesecond caseheassumeshatthe speeds fone t o and the samebodymoving different ediaare in inverse in m ratioto the densities f thesemedia;thus,for instance,f the o i densityofwaterwere tentimes thatofair,thespeedinairwould be ten timesgreater thanin water. Fromthissecond supposition,

62 THE TWO NEW SCIENCESOF GALILEO ti0n, he shows that, sincethe tenuityof a vacuumdiffers infinitelyfromthat of any mediumfilled with matterhowever rare, any bodywhichmovesin a plenumthrougha certain spacein a certaintimeoughtto movethrougha vacuuminstantaneously; instantaneous otionis an impossibility; but m it istherefore impossible a vacuum that should beproduced by motion. SALV. Theargument s,asyousee,adhorninem, is,it is i that directedagainstthosewhothoughtthe vacuum prerequisite a formotion.Nowif I admittheargumento be conclusive t and concede alsothat motioncannottakeplacein a vacuum, he t assumption a vacuumconsideredbsolutely not with of a and referenceomotion,isriottherebyinvalidated.But to tellyou t whatthe ancientsmightpossibly havereplied in orderto and betterunderstand just howconclusive ristotle's A demonstrationis,we may,in my opinion, enybothofhis assumptions. d Andasto thefirst,I greatlydoubtthatAristotle evertestedby experiment whether betruethat twostones, neweighing it o ten timesas muchas the other,ff allowed fall,at the sameinto stant,froma heightof,say,Ioocubits, ouldsodiffern speed w i that whenthe heavier ad reached h theground,heotherwould t nothavefallen morethanIO cubits. SIMI'. Hislanguage ould w seemto indicate hehad tried that the experiment, because says:Weseetheheavier; he nowthe word seeshowshat hehadmadethe experiment. t SACR. I, Simplicio, But whohavemadethe test canassure [lO7] youthat a cannon ballweighingneortwohundred o pounds, r o evenmore, illnotreachtheground w byasmuch spanahead asa ofa musketballweighingnlyhalfa pound, rovided o p bothare dropped froma heightof2oocubits. SALV. evenwithoutfurtherexperiment, ispossibleo But, it t proveclearly,by meansof a shortand conclusive argument, that a heavier bodydoesnotmovemorerapidlythana lighter oneprovided bothbodies areof thesamematerialandin short suchas thosementioned y Aristotle. But tellme,Simplicio, b whether admitthat eachfallingbodyacquires definite you a speed

FIRST DAY 63 speed fixedby nature, a velocitywhichcannot be increasedor diminishedexcept by the use of force[violenza] resistance. or SIMP.Therecan be nodoubt but that oneand the samebody movingin a singlemediumhas a fixedvelocitywhichis determinedby nature and whichcannot be increasedexceptby the addition of momentum [impeto] r diminishedexceptby some o resistancewhichretardsit. SALV.then we take two bodieswhosenaturalspeedsare If different,it is clearthat onunitingthe two,the morerapidone will be partly retarded by the slower,and the slowerwill be somewhathastenedby the swifter. Do you not agreewith me in this opinion? SIMP. ou are unquestionablyright. Y SALV. if this is true, and if a large stone moveswith a But speedof, say, eightwhilea smallermoveswith a speedof four, then when they are united,the systemwillmovewith a speed lessthan eight; but the two stones when tied togethermake a stonelargerthan that whichbeforemovedwith a speedofeight. Hencethe heavierbodymoveswith lessspeedthan the lighter; an effe_ whichis contraryto yoursupposition. Thus you see

[ oS]

how,from your assumptionthat the heavierbody movesmore rapidlythan the lighterone,I inferthat the heavierbody moves moreslowly. SIM1,. am allat seabecauseit appearsto methat the smaller I stonewhenaddedtothe largerincreases itsweightandby adding weightI do not see how it can fail to increaseits speedor, at least, not to diminishit. SaLv.Here again you are in error, Simplicio,becauseit is nottrue that the smallerstoneaddsweightto the larger. SIM1,. This is, indeed,quitebeyondmy comprehension. SALV. willnot be beyondyouwhenI have onceshownyou It the mistake under which you are laboring. Note that it is necessaryto distinguishbetweenheavybodiesin motionand the samebodiesat rest. A largestoneplacedin a balancenot only acquiresadditionalweightby havinganotherstoneplacedupon it, but even by the additionof a handfulof hemp its weightis augmented

64 THE TWO NEW SCIENCESOF GALILEO augmented sixto tenounces ccording thequantityofhemp. a to But if you tie the hempto the stoneand allowthemto fall freelyfromsomeheight,do you believe that the hempwill pressdownuponthe stoneandthusacceleratetsmotionor do i you think the motionwillbe retardedby a partialupward pressure? Onealwaysfeelsthe pressureuponhis shoulders whenheprevents themotionofa loadresting uponhim;but if onedescendsust asrapidlyasthe loadwouldfallhowcan it j gravitateorpress uponhim? Doyounotseethat thiswould be thesameastryingto strikea manwitha lance whenheisrunningawayfromyou witha speedwhichis equalto, or even greater,hanthatwithwhich t youarefollowing Youmust him? therefore concludehat, duringfreeandnaturalfall,the small t stonedoesnotpressuponthe largerandconsequently doesnot increasetsweight sit doeswhenatrest. i a Sn_P.But what if we shouldplacethe largerstoneupon thesmaller?
[109]

SALV. Itsweight ould w beincreased ifthelarger tone s moved morerapidly;but we have alreadyconcludedhat whenthe t smallstonemovesmoreslowly retardsto someextentthe it speedofthelarger, othat thecombination s ofthe two,which is a heavier odythanthelarger fthetwostones, b o would move less rapidly,a conclusion hichis contraryto your hypothesis. w Weinfertherefore largeand smallbodies that movewiththe samespeed provided theyareofthesamespecific gravity. SIMP. Yourdiscussion isreallyadmirable; I donotfindit yet easyto believe a bird-shotallsas swiftly cannon that f asa ball. SAT.v. Whynotsaya grainofsandasrapidlyasa grindstone ? But,Simplicio,trustyouwillnotfollow I theexamplefmany o otherswhodivertthe discussion fromitsmainintentandfasten uponsome statement ofminewhich lacks hair's-breadth a ofthe truth and,underthishair,hidethe faultofanother hichis as w bigas a ship'scable. Aristotle saysthat "an ironballof one hundredpoundsfallingfroma heightof one hundredcubits reaches groundbeforea one-pound allhas fallena single the b cubit." I saythat theyarriveat thesametime. Youfind,on making

FIRST DAY 65 makingthe experiment,that the largeroutstripsthe smallerby two finger-breadths,hat is, when the larger has reached the t ground,the other is shortof it by twofinger-breadths; owyou n wouldnot hidebehindthesetwofingersthe ninety-ninecubitsof Aristotle,norwouldyoumentionmy smallerrorandat the same time passover in silencehis very large one. Aristotledeclares that bodiesof differentweights, in the same medium,travel (in so far as their motion dependsupon gravity) with speeds whichareproportionalto theirweights;thishe illustratesbyuse of bodiesin which it is possibleto perceivethe pure and unadulterated effect of gravity, eliminatingother considerations, for example,figure as being of small importance [minimirnoznentz],nfluences hicharegreatly dependentuponthe medium i w whichmodifiesthe singleeffeccf gravity alone. Thus weobo servethat gold,the densestof all substances,when beatenout into a very thin leaf, goesfloatingthrough the air; the same thing happenswith stonewhengroundinto a very finepowder. But if you wish to maintain the generalpropositionyou will have to showthat the sameratio of speedsis preservedin the

[ Io]

caseof all heavy bodies, and that a stone of twenty pounds movesten times as rapidlyas one of two; but I claimthat this is falseand that, ifthey fallfrom a heightof fiftyor a hundred cubits,theywillreachthe earth at the samemoment. SnuP.Perhaps the result wouldbe differentif the fall took placenot from a fewcubits but fromsomethousandsof cubits. S_v. If this were what Aristotlemeant you wouldburden him with another error which wouldamount to a falsehood; because,sincethere is no suchsheerheightavailable earth, it on is clearthat Aristotle couldnot have made the experiment;yet he wishesto giveus the impressionof his havingperformedit whenhe speaksof suchan effecCt asonewhichwe see. S_rP.In fact, Aristotle doesnot employthis principle,but usesthe other onewhichis not, I believe,subjecCt thesesame to difficulties. SALv. ut the oneis as falseas the other;and I amsurprised B that you yourselfdo not see the fallacyand that you do not perceive

66 THE TWO NEW SCIENCESOF GALILEO perceivehat ifit weretruethat,in mediaofdifferent ensities t d and different esistances, r suchas waterand air, one and the samebodymoved air morerapidlythaninwater,inproporin tionas the densityofwateris greaterthanthat ofair,thenit wouldfollow any bodywhichfallsthrough oughtalso that air to fallthroughwater. But this conclusion isfalseinasmuchs a manybodies whichdescend air not onlydo notdescend in in water,but agtually rise. Sn_P. do not understand necessity yourinference; I the of and in additionI willsay that Aristotle discusses nlythose o bodies whichfallin bothmedia,notthosewhich fallin air but riseinwater. SAT.V. arguments The whichyou advancefor the Philosopheraresuchashehimself ould w havecertainly avoided soas notto aggravate hisfirstmistake.Buttellmenowwhetherhe t density[corpulenza]the water,or whatever maybe that of it [III] retardsthe motion,bearsa definite ratioto the densityof air whichis lessretardative;and if so fixa valuefor it at your pleasure. SIv_P. Sucha ratiodoesexist;letus assume beten;then, itto fora bodywhich fallsinboththesemedia,hespeed t inwater will betentimesslowerhanin air. t SALV. shallnowtakeone ofthosebodies I whichfallin air butnotinwater,saya wooden all,andI shallaskyouto assign b toit anyspeedyouplease foritsdescenthrough t air. SIMP. Letus suppose moves itha speed it w oftwenty. SAT.V. Verywell. Then it is clearthat this speedbearsto some smaller peedthe sameratioasthedensity s ofwaterbears to that of air; and the valueof this smaller speedis two. So that reallyif wefollow exa_lythe assumptionfAristotle o we oughtto inferthat the wooden whichfallsin air, a subball stance tentimesless-resisting thanwater,witha speedoftwenty would fallinwaterwitha speedoftwo,instead ofcomingo the t surface fromthe bottomasit does;unless perhaps ou wishto y reply, hich donotbelieve w I youwill,that therising ofthewood through thewateristhesameasits falling witha speedoftwo. But

FIRST DAY 67 But since thewooden alldoesnotgotothebottom,I thinkyou b willagreewithme that wecanfinda ballofanothermaterial, notwood, which doesfallinwaterwitha speed oftwo. SIMP. Undoubtedly e can;but it mustbe of a substance w considerably heavierhanwood. t SAT.V. Thatisit exa_ly. But ifthissecond ballfallsinwater witha speedof two,whatwillbe ks speedofdescentin air? If youholdto the ruleofAristotle youmustreplythat it will moveat the rateoftwenty;buttwentyisthe speed whichyou yourself avealreadyassigned the wooden h to ball;hencethis and theotherheavier ballwilleachmovethrough airwiththe samespeed. But nowhowdoesthe Philosopher armonize h thisresult ithhisother,namely, bodies fdifferent w that o weight movethrough samemedium the withdifferent peeds--speeds s whichareproportional theirweights?But withoutgoing to into the mattermoredeeply,howhave thesecommon and
[II2]

obvious roperties p escaped ournotice?Haveyounotobserved y that twobodies hichfallin water,onewitha speed hundred w a timesasgreatasthatOftheother,willfallinairwithspeeds o s nearlyequalthat onewillnotsurpassheotherby asmuchas t onehundredth art Thus,forexample, p ? aneggmadeofmarble willdescend wateronehundredtimesmorerapidlythana in hen'segg,while airfalling in froma height ftwentycubits o the onewillfallshort f theotherbylessthanfour o finger-breadths. In short, heavybodywhichsinksthroughencubits a t ofwater in three hourswilltraverse tencubits ofair inoneortwopulsebeats;and if the heavybodybe a ball of leadit willeasily traversehe ten cubitsof water lessthan doublethe time t in required tencubitsofair. Andhere,I amsure,Simplicio, for youfindno groundfordifference objecCtion. or Weconclude, therefore, theargument oesnotbearagainstheexistence that d t ofa vacuum; ut if it did,itwould b onlydoaway withvacuaof considerable which size neither nor, nmyopinion, I i theancients everbelievedo exist t innature, lthoughheymightpossibly e a t b produced yforce_olenza] b [ asmaybegatheredromvarious xf e periments hose w description would hereoccupy toomuchtime. Sagr.

68 THE TWO NEW SCIENCES OF GALILEO SAGR. Seeing Simplicio that issilent,I willtaketheopportunityofsaying something.Since haveclearly you' demonstrated that bodies fdifferent o weights onotmove d inoneandthesame medium withvelocities roportional theirweights, ut that p to b they allmovewith the samespeed,understandingf course o that theyare of the samesubstance at leastof the same or specificravity; ertainly g c notofdifferent specificravities, g forI hardlythink youwouldhaveus believe ballof corkmoves a [II3l withthe samespeedas oneoflead;andagainsinceyouhave clearlydemonstrated that one and the same body moving through differentlyesisting ediadoes r m notacquirepeeds hich s w are inversely proportional the resistances, am curious to I to learnwhataretheratiosa_uallyobserved thesecases. in SAT.V. Theseare interesting questions I have thought and muchconcerning them. I willgiveyouthemethodofapproach andthe result hichI finally w reached.Having onceestablished thefalsity oftheproposition thatoneandthesame bodymoving throughdifferentlyesisting r mediaacquires speeds whichare inversely proportional the resistances f thesemedia,and to o havingalsodisproved statementthat in the samemedium the bodies fdifferent o weight cquire a velocities proportional their to weights (understanding this appliesalsoto bodieswhich that differ merely inspecificravity), thenbeganto combinehese g I t twofa_s andto consider whatwould happen bodies fdifferif o entweightwereplacedinmediaofdifferentesistances; I r and foundthat the differences speedweregreaterin thosemedia in which moreresistant, hatis,lessyielding.Thisdifference were t wassuchthat twobodies whichdiffered scarcely allin their at speedthrough would,nwater,falltheonewitha speedten air i timesas greatas that ofthe other. Further,thereare bodies which willfallrapidlyin air,whereasf placed i inwaternotonly willnotsinkbut willremain restor willevenriseto the top: at for it ispossibleo findsome t kindsofwood,suchasknotsand roots,whichremainat restin waterbut fallrapidlyin air. SAcg.I haveoftentried with the utmostpatienceto add grainsofsandto a ballofwaxuntilit should acquire same the specific

FIRST DAY 69 specificgravityas water and wouldthereforeremain at rest in this medium. But with allmy care I was neverable to accomplish this. Indeed, I do not knowwhetherthere is any solid substancewhosespecific gravity is, by nature, so nearly equal to that ofwater that ifplacedanywherein water it willremain at rest. SALV. this, as in a thousand other operations,men are In surpassedby animals. In this problemof yoursonemay learn much fromthe fishwhich are very skillfulin maintainingtheir equilibriumnot only in one kind of water, but also in waters whichare notably differenteither by their own nature or by someaccidentalmuddinessor through salinity, each of which producesa marked change. So perfectlyindeedcan fishkeep their equilibriumthat they are ableto remainmotionless any in position. This they accomplish,I believe, by means of an apparatus especiallyprovided by nature, namely, a bladder located in the body and communicatingwith the mouth by meansof a narrowtube throughwhichthey are able,at will,to expela portionof the air containedin the bladder:by risingto the surfacetheycan take in moreair; thus theymakethemselves heavieror lighterthan water at will and maintain equilibrium. SAOl<. meansofanother deviceI was ableto deceivesome By Mendsto whomI had boasted that I couldmake up a ball of wax that wouldbe in equilibriumin water. In the bottom of a vesselI placedsomesalt water and uponthis somefreshwater; then I showedthem that the ball stoppedin the middleof the water, and that, when pushed to the bottom or lifted to the top, wouldnot remainin eitherof theseplacesbut wouldreturn to the middle. SALV. This experimentis not withoutusefulness. For when physiciansare testing the variousqualitiesof waters,especially their specificgravities,they employa ball of this kind so adjusted that, in certainwater, it willneitherrise nor fall. Then in testing another water, differingever so slightlyin specific gravity [peso],he ballwill sink ifthis waterbe lighterand rise t if it be heavier. And so exact is this experimentthat the addition

[II41

70 THE NEW TWO SCIENCESOF GALILEO tionoftwograinsof saltto sixpounds ofwaterissufficient to maketheballriseto thesurface thebottomtowhich had from it fallen. Toillustrate theprecisionfthisexperiment o andalsoto clearlydemonstrate non-resistance water to division, the of I wishto add that thisnotable difference specific in gravitycan be produced onlyby solution someheaviersubstance, not of but alsobymerely heating cooling; sosensitives water or and i to thisprocesshat bysimply t adding fourdrops fanother ater o w whichis slightly warmer r coolerhanthe sixpoundsonecan o t cause theballto sinkorrise;itwillsinkwhenthewarmwateris poured inandwillriseupontheaddition ofcold water. Nowyou canseehowmistaken arethose philosophers whoascribeo water t viscosity rsome o othercoherence ofpartswhich offersesistance r to separationfpartsandto penetration. o SAoR. Withregardto thisquestion havefoundmanyconI vincing argumentsna treatisebyourAcademician; i butthereis onegreatdifficulty fwhichI havenotbeenableto ridmyself, o namely, iftherebenotenacity orcoherence betweenheparticles t ofwaterhowisit possibleorthose f largedrops ofwaterto stand outinrelief uponcabbageeaves ithout cattering rspreading l w s o out? SALV. Although thosewhoarein possessionfthe truth are o ableto solve allobje_ions raised, would I notarrogateo myself t suchpower; evertheless n myinability should notbe allowedo t becloud truth. To beginwithletme confesshat I donot the t understand howtheselargeglobules ofwaterstandoutandhold themselves although knowfor a certainty,hat it is not up, I t owing any internaltenacityactingbetween particles f to the o water;whenceit must follow that the causeof this effe_ is external.Besidetheexperiments alreadyshown provethat to the cause notinternal, canofferanotherwhichisveryconis I vincing.If theparticles fwaterwhichsustain o themselves ina heap,whilesurrounded y air, didso in virtueof an internal b cause thentheywould sustain themselves moreeasily much when surrounded medium which bya in theyexhibit esstendency l to fall than they do in air; sucha medium wouldbe any fluid heavier
&

[II5]

FIRST DAY 71 heavierthanair, as,for instance, ine:and therefore some w if winebepoured aboutsucha dropofwater,thewine mightrise until the dropwasentirelycovered, ithoutthe particles w of water,heldtogetherby this internalcoherence, everparting company. But this is not the fa_; for as soonas the wine touchesthe water,the latter withoutwaitingto be covered scattersandspreads outunderneathhe wineifit be red. The t causeof this effed'ts therefore i external nd is possibly be a to foundin the surrounding Indeedthereappearsto be a air. considerable antagonism between and wateras I haveobair servedin thefollowing experiment. aving H takena glassglobe whichhad a mouthof aboutthe samediameter a straw,I as filled withwaterand turnedit mouthdownwards; it nevertheless,thewater,although uiteheavy q andproneto descend, and the air, whichis very lightand disposedo risethroughthe t water,refused,the one to descend the other to ascend and throughtheopening, both remained but stubborn anddefiant. On the otherhand,as soonas I applyto this opening glass a of redwine,whichis almost nappreciably i lighterthanwater, redstreaksareimmediatelybserved ascend o to slowly through thewaterwhile thewaterwithequalslowness descends through the winewithoutmixing, untilfinallythe globeis completely filled withwineandthewaterhasallgonedownintothevessel below.Whatthencanwesayexcept hat thereexists, etween t b waterand air,a certainincompatibility whichI donotunderstand,but perhaps .... S_P. I feelalmostikelaughingt thegreatantipathy l a which Salviati xhibits gainsttheuseof thewordantipathy; ndyet e a a it isexcellentlydapted explain difficulty. a to the 8_v. Alright, f it please i Simplido, thiswordantipathy let bethesolution fourdifficulty. eturning thisdigressmn, o R from let usagaintakeupourproblem. ehavealready W seenthat the difference speedbetween of bodiesof different pecific s gravitiess mostmarked thosemedia i in whicharethe most resistant:hus, in a medium quicksilver, t of goldnot merely sinksto the bottommorerapidlyhanleadbutit is the only t substance

[II6l

72 THE TWO NEW SCIENCESOF GALILEO substance willdescend that atall;allothermetals andstonesise r to the surfaceand float. On the otherhandthe variationof speedin airbetween ballsof gold,lead,copper, orphyry, nd p a otherheavymaterialss so slightthat in a fallof IOO i cubitsa ballofgoldwould surelynotoutstriponeofcopper y asmuch b asfourfingers.Having observed I cameto the conclusion this that in a mediumtotallydevoidofresistance llbodieswould a fallwiththesame speed. Smm.Thisis a remarkabletatement,Salvlati.But I shall s neverbelieve evenin a vacuum, f motionin sucha place that i werepossible, lockofwoolanda bit ofleadcanfallwith the a samevelocity. SAT.v. Alittlemoreslowly, Simplicio. ourdifficultys not Y i so reconditenor am I so imprudent s to warrant you in a believing that I have not alreadyconsideredhis matterand t foundthe propersolution. Hencefor my justification and foryourenlightenment hearwhatI haveto say. Ourproblem isto findoutwhathappens bodies fdifferent to o weight oving m in a medium devoid ofresistance, that theonlydifference so in speedisthat whicharisesfrominequality ofweight. Since no medium except neentirely o freefromairandotherbodies, e it b everso tenuousand yielding, an furnishour senseswith the c evidence are looking we for,and sincesucha mediumis not available, e shallobservewhat happensin the rarest and w leastresistantmediaas compared withwhathappensin denser andmoreresistant edia. Because m ifwefindasa facet the that variationofspeedamongbodies ofdifferent pecificravities s g is lessand lessaccording asthe medium becomes oreandmore m yielding, nd if finally a medium extreme a in of tenuity,though nota perfect acuum, efindthat, inspiteofgreatdiversity v w of specific ravity[peso], difference speedisverysmalland g the in almostinappreciable, thenwearejustified believingt highly in i probable in a vacuumallbodies that wouldfallwith the same speed. Letus, inviewofthis,consider hattakesplacein air, w where forthesakeofa definite figure andlightmaterialmagine i aninflated bladder. Theairinthisbladder hensurroundedy w b
air

[117]

FIRST DAY 73 airwillweigh littleornothing, it canbeonlyslightly since compressed;its weightthenis smallbeingmerely of the skin that whichdoesnot amountto the thousandth of a massof part leadhavingthe samesizeas the inflated bladder. Now,Simplicio,f weallow i thesetwobodies tofallfroma heightoffour or six cubits,by what distancedoyou imagine lead will the anticipatehe bladder?Youmaybe surethat theleadwillnot t travel three times,or eventwice,as swiftlyas the bladder, although wouldhave madeit movea thousand you timesas rapidly. Sire,.It maybe asyousayduringthefirstfourorsixcubits ofthe fall;but afterthe motionhas continued longwhile,I a believe the leadwillhaveleftthe bladder ehindnotonly that b six out of twelveparts of the distancebut even eightor ten. SAzv.I quiteagreewith you and doubtnot that,in very longdistances, theleadmightcover nehundred o miles while the bladder astraversing ne;but,mydearSimpliclo, w o thisphenomenonwhichyouadduceagainstmyproposition preciselyhe is t one whichconfirmsit. Let me oncemoreexplainthat the variationofspeedobserved inbodies ofdifferent pecific s gravities is not causedby the difference specificravitybut deof g pendsuponexternal ircumstances c and,in particular, uponthe resistance f the medium, o that if this is removed o s allbodies wouldfall with the samevelocity;and this resultI deduce mainlyfromthefacet which youhavejustadmitted andwhich is very true, namely,that, in the caseof bodieswhichdiffer widelyin weight,theirvelocities iffer d moreand moreas the spacestraversed increase, omething hichwouldnotoccurif s w the effe_ dependedupondifferences specific ravity. For of g sincethesespecificravities g remain constant,he ratiobetween t the distancesraversedoughtto remainconstant hereasthe t w facet s that this ratiokeepson increasings the motionconi a tinues. Thusa veryheavybodyin a fallofonecubitwillnot anticipate verylightonebysomuchasthetenthpartofthis a space;but in a falloftwelve cubitstheheavybodywouldoutstrip

74 THE TWO NEW SCIENCES GALILEO OF striptheother byone-third, andin.afallofonehundredubits c by9o/Ioo, etc. Sn_. Verywell: ut,following b yourownlineofargument, ifdifferences ofweightn bodies fdifferent i o specificravities g cannot roduce change the ratiooftheirspeeds, the p a in on ground theirspecificravities o not change, that g d howis it possible forthemedium, which alsowesuppose remain to constant,tobring about nychangentheratioofthese a i velocities ? S_v. Thisobjedtion withwhich youoppose ystatement m isclever; andI mustmeetit. I begin bysaying thata heavy body hasaninherentendency move t to witha constantly and uniformlyccelerated a motiontowardthe commonenterof c gravity, thatis,toward thecenter fourearth,sothatduring o equalntervals i oftimeitreceives increments equal ofmomentum andvelocity.This,youmustunderstand, whenever holds all external andaccidental hindrances havebeenremoved; butof thesethereis onewhichwecanneverremove, namely,he t medium whichmustbe penetratedndthrustasideby the a fallingody.Thisquiet, ielding, medium b y fluid opposes motion [II9] throughtwitharesistance isproportional i which totherapidity withwhich medium the mustgivewayto thepassage f the o body;which body,as I havesaid,isby naturecontinuously accelerated sothat it meets withmoreandmoreresistance in themedium andhence diminution itsrateofgain a in ofspeed untilfinallyhespeed t reaches a point ndtheresistance such a of themedium becomes sogreat hat,balancing t eachother, hey t prevent nyfurther cceleration a a andreduce themotion fthe o body toonewhich isuniform andwhich ill hereafter w t maintain a constant value.There is,therefore, anincrease intheresistance ofthemedium, notonaccount ofanychange initsessential properties, butonaccount ofthechange inrapidity ithwhich w it mustyield andgive waylaterally tothepassage ofthefalling body whichsbeing i constantly accelerated. Nowseeingow h greatistheresistance theairofferso which t theslight omentum m [momento] ofthebladderndhowsmall a that whichit offerso the largeweightpeso] the lead,I t [ of am

FIRST DAY 75 am convinced that, if the medium wereentirelyremoved, the advantage received the bladderwould so greatand that by be coming theleadsosmallthat theirspeeds ould to w beequalized. Assuminghis principle,hat all fallingbodiesacquireequal t t speedsna medium i which, naccount fa vacuum o o orsomething else,offers noresistanceo the speedofthe motion, eshallbe t w ableaccordingly determine ratiosof the speeds both to the of similar anddissimilar bodies moving eitherthrough oneandthe samemediumor throughdifferent pace-filling, therefore s and resistant, edia. Thisresultwemayobtainbyobserving m how muchtheweightofthemedilml etracts d fromtheweight fthe o moving body,which weightisthemeans employedythefalling b bodytoopena pathforitselfand to pushasidethepartsofthe medium, omething hichdoesnothappenina vacuum s w where, therefore,no difference speed]is to be expectedfroma [of difference f specific ravity. Andsinceit is knownthat the o g effecCt ofthemedium istodiminishheweight f.the t o bodybythe weight fthemedium o displaced, wemayaccomplish ourpurpose bydiminishing just this proportion speeds f the falling in the o bodies, hichina non-resisting w medium wehaveassumed be to equal. Thus,forexample, imagine leadto beten thousand timesas heavyas air whileebonyisonlyone thousand timesasheavy. [12o] Herewehavetwosubstances hosespeeds ffallin a medium w o devoidofresistancereequal:but, whenairis themedium,t a i willsubtra_from speed the oftheleadonepartintenthousand, and from speedofthe ebonyonepart inonethousand,. e. the i ten parts in ten thousand. Whiletherefore leadand ebony would fromany givenheightin the sameinterval f time, fall o provided retarding the effe_of the airwereremoved, lead the will,inair,loseinspeedonepartintenthousand; ndtheebony, a tenpartsintenthousand.Inotherwords, iftheelevation from whichthe bodiesstartbe dividedintoten thousand parts,the leadwillreach groundleaving the theebonybehind asmuch by asten,oratleastnine,oftheseparts. Isit notclearthenthat a leadenball allowedto fall froma towertwo hundred cubits high

76 THE TWO NEW SCIENCES OF GALILEO highwilloutstrip 'anebonyballby,less thanfourinches Now ? ebonyweighs thousand a timesasmuchasair but thisinflated bladderonlyfourtimesas much;therefore diminisheshe air t inherent nd naturalspeedofebonybyonepart in a thousand; a whilethat of the bladderwhich, f freefromhindrance, ould i w be the same,experiences diminution air amounting one a in to part in four. So that whenthe ebonyball,fallingfrom the tower,has reachedthe earth, the bladderwillhave traversed onlythree-quarters f this distance. Leadis twelvetimesas o heavyaswater;but ivoryisonlytwice asheavy. Thespeeds f o thesetwosubstances hich, henentirely w w unhindered, areequal willbe diminished water,that of leadbyonepart in twelve, in that of ivoryby half. Accordingly henthe lead has fallen w through eleven cubits ofwatertheivory willhavefallenthrough onlysix. Employing principle shall,I believe, inda this we f muchcloser greement a ofexperiment withourcomputation than withthat ofAristotle. In a similar anner emayfindtheratioofthespeeds m w ofone andthesame bodyindifferent media, otbycomparinghe fluid n t differentesistances r ofthemedia, ut byconsidering excess b the of the specific gravityof the bodyabovethoseof the media. Thus, for example, is one thousand fin timesheavierthan air and ten timesheavierthanwater;hence,if we divideits unhindered speedinto IOOOarts,air willrob it of oneof these p partsso that it willfallwitha speedof999,whilein waterits speedwillbe9oo,seeing waterdiminishes that itsweight yone b partintenwhile airbyonlyonepartin a thousand. Againtakea solida littleheavierthanwater,suchasoak,a ball ofwhichwillweighlet us say IOOO drachms; uppose s an equalvolume ofwaterto weigh 95o,andanequalvolume ofair, 2;thenit isclearthat if theunhinderedpeedoftheballisIooo, s its speedin airwillbe998,but inwateronly5o,seeing the that waterremoves 950of the IOOOarts whichthe bodyweighs, p leaving only50. Sucha solidwouldtherefore movealmosttwentytimesas fast in air asin water,sinceits specific ravityexceedshat of g t water

[i2i]

FIRST DAY 77 waterby onepart in twenty. Andherewe mustconsiderhe t fa_ that onlythosesubstances hichhave a specific ravity w g greaterthanwatercanfallthroughit--substances whichmust, therefore, behundreds ftimesheavier o thanair;hence whenwe try to obtainthe ratioofthe speedin air to that in water,we may,withoutappreciable error,assume thatairdoesnot,to any considerablextent,diminish freeweight[assoluta e the gravita], and consequently theunhinderedpeed[assoluta s velocitY] ofsuch substances.Having thuseasily foundtheexcess ftheweight f o o thesesubstances verthat ofwater,wecansaythat theirspeed o inairisto theirspeedinwaterastheirfreeweight[totale gravitY] isto the excess fthisweightoverthat ofwater. Forexample, o a ball of ivoryweighs20 ounces;an equalvolumeof water weighs17ounces; encethe speedof ivoryin air bearsto its h speedinwatertheapproximate ratioof20:3 . SAGR.havemadea greatstepforward I inthistrulyinterestingsubjecCt uponwhichI have longlaboredin vain. In order to put these theoriesinto practicewe need onlydiscover a methodof determininghe specificravityofair withreference t g to waterand hencewith referenceo otherheavysubstances. t Sr_P.But if we find that air has levityinsteadof gravity what then shallwe say of the foregoing iscussion hich,in d w otherrespecCts, isveryclever? SAT.v. shouldsay that it was empty,vain, and trifling. I But canyoudoubtthat airhasweight henyouhavetheclear w testimony ofAristotle affirming thatalltheelements aveweight h including and excepting nlyfire? As evidence f this he air, o o citesthe fa_ that a leatherbottleweighsmorewheninflated than whencollapsed. Snvn,. am inclinedto believethat the increase weight I of observed intheinflatedleather ottleorbladder rises, otfrom b a n the gravityofthe air,but fromthemanythickvapors mingled with it in theselowerregions.To this I wouldattributethe increase fweightintheleather ottle. o b S_mv. would haveyousaythis,andmuchlessattribute I not it to Aristotle; because,fspeaking ftheelements, i o hewished to persuade

78 THE TWO NEW SCIENCES OF GALILEO persuade eby experiment air hasweightand wereto say m that to me:"Takea leatherbottle,filliiwith heavyaporsandobv servehowits weightincreases," wouldreplythat the bottle I would weighstillmoreif filledwithbran;andwouldthenadd that this merelyprovesthat branand thickvaporsareheavy, but in regardto air I should stillremainin the samedoubtas before. However,he experiment t ofAristotle goodand the is propositionstrue. But I cannot ayasmuchofa certainother i s consideration, taken at facevalue;this consideration asofw feredbya philosopher whose nameslipsme;but I know have I readhisargument hichisthat airexhibits reater w g gravitythan levity,becauseit carriesheavybodiesdownward oreeasily m thanit doeslightonesupward. SAcR. Fineindeed! Soaccording thistheoryair ismuch to heavier thanwater,since allheavy bodies arecarried downward moreeasilythrough thanthrough air water,andalllightbodies buoyed moreeasilythrough up waterthanthrough air;further thereis an infinitenumber f heavybodies o whichfallthrough air but ascendinwaterand thereis an infinite number f subo stanceswhichrisein waterand fallin air. But, Simplicio, the question asto whether weight fthe leatherbottleis owing the o to thickvapors r to pureair doesnotaffecCt o ourproblem which isto discover owbodies h movethrough thisvapor-ladentmosa phere ofours. Returning nowto thequestion whichinterests e m more, should I like,forthesakeofmorecompletend thorough a knowledge this matter, notonly to be strengthened my of in beliefhat airhasweight ut alsoto learn,ifpossible, t b howgreat .itsspecificravityis. Therefore, g Salviati, fyoucansatisfymy i curiosity nthispointpraydoso. o SALv. experiment ith the inflatedleatherbottle of The w Aristotle provesconclusively air possesses that positive gravity and not, as somehavebelieved,levity,a propertypossessed possibly nosubstance by whatever; or if air didpossessthis f quality ofabsolute andpositive levity,it should oncompression exhibitgreaterlevityand, hence,a greatertendencyto rise; butexperiment shows preciselyheopposite. t

[I 31

As

FIRST DAY 79 As to the otherquestion, namely, owto determinehe h t specificravity g ofair,I haveemployed thefollowing method. I tooka ratherlarge glassbottlewitha narrow neckandattached aleather over, toit c bindingt tightly i about heneck t of thebottle: nthetopofthiscover insertedndfirmlyastened i I a f the valve leather ottle,through ofa b which forced I intothe glassbottle,by means f a syringe, largequantity air. o a of Andsince iseasily air condensed onecanpumpintothebottle twoorthreetimes itsownvolumefair. AfterthisI tookan o accuratealancendweighed b a thisbottle ofcompressed airwith the utmost recision, p adjusting weight ithfinesand. I the w nextopenedhevalve t andallowed thecompressed airtoescape; then replaced flaskuponthe balance the andfoundit perceptibly lighter: thesand from which adbeen asa counterh used weight now I removedndlaidaside a asmuch aswasnecessary to again securealance. nderhese b U t conditions therecan beno doubtbutthattheweight ofthesandthuslaidaside represents theweight oftheairwhich hadbeen forcedntotheflask i andhad afterwardsscaped.But afterall this experiment e tellsme merelyhattheweight t ofthecompressed airisthesame asthat ofthesand'removed thebalance; howevercomes from when it to knowingertainlynddefinitely c a theweight ofairascompared withthatofwater oranyotherheavy substance cannot thisI hopeto do without irstmeasuring volume f the [quantitY] of compressed air;forthismeasurement I havedevisedhe two t following methods. According tothefirstmethodnetakes o abottle withanarrow neck similar totheprevious overthemouth one; ofthisbottle is slipped leatherubewhichsbound a t i tightly about heneck t of theflask; theother endofthistubeembraces thevalve attached to the firstflaskandis tightly boundaboutit. Thissecond flaskis provided itha holeinthebottom w through which an ironrodcanbeplaced soasto open,at will, hevalveabove t mentioned andthuspermithesurplusirofthefirsttoescape t a afterit hasoncebeen weighed: buthissecond ottlemustbe b filled withwater. Having preparedverything themanner e in

[I241

above

8o THE TWO NEW SCIENCES OF GALILEO abovedescribed, openthe valvewiththe rod; the air willrush intothe flaskcontaininghewaterandwilldriveit through t the holeat thebottom,it beingclearthat the volume[quanti_] f o waterthusdisplaced equalto the volume[mole quanti_] is e of air escaped fromthe othervessel. Havingset asidethis displacedwater,weighthe vesselfromwhichthe air hasescaped (whichis supposedto have been weighedpreviously while containinghe compressed t air),and remove surplus the ofsand as describedbove;it is thenmanifestthat the weightof this a sandispreciselyhe weight fa volume t o [mole] equalto the ofair volume ofwaterdisplaced ndsetaside; hiswaterwecanweigh a t and findhowmanytimesits weightcontainshe weightof the t removedsand,thus determining definitelyhowmany times heavierwater is than air; and we shallfind,contraryto the opinion ofAristotle,hat thisisnotIOtimes,but, asourexperit mentshows, morenearly 40otimes. The secondmethodis moreexpeditiousnd canbe carried a out witha single vesselfittedup asthe firstwas. Herenoair is added that whichthevessel aturally to n contains ut wateris b forcedintoit withoutallowingnyair to escape;hewaterthus a t introduced necessarilyompresses air. Havingforcedinto c the thevessel smuchwateraspossible, a filling say,three-fourths it, Mll,whichdoesnot requireany extraordinary effort,placeit uponthe balance andweighit accurately; extholdthe vessel n mouthup, openthe valve,and allowthe air to escape;the volume theair thusescapingspreciselyqualto the volume of i e of watercontained the flask. Againweighthevesselwhich in willhavediminished weighton accountofthe escapedair; in this lossinweightrepresentsheweight fa volume t o ofairequal to thevolume ofwatercontained inthevessel. Sm_,.Noone candenythe clevernessndingenuity your a of devices;but whilethey appearto givecomplete intellecCtual satisfacCtion theyconfuse ein anotherdirecCtion. m Forsinceit is undoubtedlyruethat theelements t whenintheirproperplaces have neitherweightnor levity,I cannotunderstand howit is possiblefor that portionofair, whichappeared weigh,say, to 4 drachms fsand,should o reallyhavesucha weightn airasthe i sand

FIRST DAY 81 sandwhichcounterbalances It seems me,therefore, it. to that theexperiment should becarried out,notinair,butina medium
[I25]

in whichthe air could exhibititsproperty ofweightif suchit really has. SALV. TheobjectionfSimplicio o iscertainly thepointand to musttherefore itherbe unanswerable demand equally e or an clear solution.It isperfectlyvidenthatthatairwhich, e t under compression, weighed muchas the sand,losesthis weight as when allowedo escapentoitsownelement, once t i while,ndeed, i the sandretains weight. Hence this experiment beits for it comes necessary selecta place to where airaswellas sand can gravitate; because, s has beenoftenremarked, medium a the diminishes theweightof anysubstancemmersed it by an i in amount qualto theweight f the displaced edium; that e o m so airinairlosesallitsweight. Ifthereforehisexperiment t istobe made withaccuracyt should i beperformed vacuum ina where everyheavybodyexhibitsts momentum i withouttheslightest diminution. f then,Simplicio, I wewereto weigh portionof a airina vacuum would youthenbesatisfied andassuredfthe o fact? Stop.Yes truly:but this is to wishor ask the impossible. Sm_v. our obligation Y willthen be very great if, for your sake,I accomplish theimpossible. ut I donotwantto sell ou B y something hichI havealready w givenyou;for in the previous experiment weweighed theairin vacuum andnotinairorother medium. The fact that any fluid mediumdiminisheshe t weight f a massimmersed it, is due,Simplicio, theresisto in to ancewhichthis mediumoffersto its beingopenedup, driven aside,andfinallyliftedup. Theevidenceorthis isseenin the f readiness ith whichthe fluidrushesto fillup any spaceforw merlyoccupied ythe mass;if themedium not affected y b were b suchan immersion thenitwould notreactagainsttheimmersed body. Tellmenow,whenyouhavea flask,in air,filled withits naturalamountof air andthenproceed pumpintothevessel to moreair,doesthis extracharge anywayseparate divide in or or changethecircumambient Doesthevessel erhaps air? p expand
SO

82 THE TWO NEW SCIENCESOF GALILEO so that the surrounding ediumi_ displaced orderto give m in moreroom?Certainlynot. Therefore oneis ableto say that [126] this extrachargeof air is not immersedn the surrounding i mediumforit occupies o spacein it, but is,as it were,in a n vacuum. Indeed,itis really vacuum; orit diffusesntothe ina f i vacuitieswhicharenot completelyilledby the original nd f a uncondensedir. In facet donot seeany difference a I between the enclosed the surrounding edia:forthe surrounding and m medium doesnotpress upontheenclosed edium m and,vice versa, theenclosed edium m exertsnopressuregainstthesurrounding a one;thissamerelationshipxistsin thecaseofanymatterin a e vacuum,as wellas in the caseof the extracharge aircomof pressedinto the flask. The weightof this condensed is air thereforehe sameas that whichit wouldhaveif setfreein a t vacuum. It istrueofcourse thattheweight fthesandusedas o a counterpoise would a littlegreatern vacuohaninfreeair. be i t Wemust,then,saythatthe airis slightly lighter hanthe sand t requiredocounterbalance thatisto say,by anamountequal t it, totheweightin vacuo fa volume fairequalto thevolume o o of thesand.
Atthis pointin anannotatedcopyoftheoriginal ditionthefollowing e note by Galileois found. [SAcraAverycleverdiscussion, olvinga wonderful roblem, ecause s p b it demonstratesbrieflyand conciselythe manner in which one may find the weightof a body in vacuoby simply weighingit in air. The explanation as follows: hena heavybodyis immersed airit losesin is w in weightan amountequalto theweightofa volume[mole] fairequivalent o to the volume[mole]of the body itself. Hence if one adds to a body, withoutexpandingit, a quantity of air equalto that whichit displaces and weighsit, hewill obtainits absoluteweightin vacuo,since,without increasing in size,hehas increased it itsweightby just theamountwhich it lostthroughimmersion air. in Whentherefore forcea quantityof waterinto a vesselwhich alwe ready containsits normalamountof air, withoutallowingany of this air to escape is clearthat thisnormalquantity ofairwillbecompressed it and condensed intoa smallerspacein orderto make room the water for which is forcedin: it is also clear that the volume of air thus compressedis equalto the volumeof wateradded. If now the vesselbe weighed

FIRST DAY
_n _(ZCUO.

83

weighed{n air in this condition,it is manifestthat the weightof the waterwillbeincreased that ofanequalvolumeofair; the total weight by of waterand air thus obtainedis equalto the weightofthe water alone Now recordthe weightof the entire vesselandthen allowthe compressedair to escape;weighthe remainder;the difference f thesetwo o weightswill be the weightof the compressedair which,in volume,is equalto that ofthe water. Next findthe weightof the water aloneand add to it that of the compressed weshallthen have the wateralone air; in vacuo. To find the weightof the water we shall have to removeit fromthe vessel and weigh the vesselalone; subtra&this weightfrom that ofthe vesseland watertogether. It is clearthat the remainderwill bethe weightof the wateralonein air.]

'

SzMv. previousexperiments, my opinion,eft some_ The in l thingtobedesired: utnowI amfullysatisfied. b SAT.v. facts set forth by me up to this pohntand, in The particular,he onewhichshows t that difference ofweight, ven e whenverygreat,iswithout ffeCt e inchanging thespeed offalling bodies, o that as far asweightis concernedheyallfallwith s t equalspeed:this ideais, I say, so new,and at firstglanceso remotefromfaCt, hat ifwedonothavethemeans t ofmaking it just as clearas sunlight, t had betternot be mentioned; ut i b havingonceallowed passmylipsI mustnegleCt itto noexperimentorargumento establisht. t i SAGR. onlythis but alsomanyotherof yourviewsare Not so far removedfrom the commonly acceptedopinionsand doCtrineshat if youwereto publishthemyouwouldstirup t a largenumberof antagonists; humannatureis suchthat for mendonot lookwithfavorupondiscoveries--either oftruth or fallacy--intheir ownfield,whenmadeby othersthan themselves. Theycallhiman innovator f doCtrine, unpleasant o an title, bywhichtheyhopeto cutthoseknotswhichtheycannot untie, and by subterranean minesthey seekto destroystructureswhichpatientartisanshavebuiltwith customary tools. But as for ourselves hohave no suchthoughts,the experiw mentsand arguments whichyou have thus far adducedare fullysatisfaCtory; howeverf youhaveany experiments i which are

[I27]

[Iz8]

84 THE TWO NEW SCIENCES OF GALILEO aremoredirector any arguments hichare moreconvincing w wewillhearthemwithpleasure. SALV. Theexperiment madeto ascertain whetherwobodies, t differing reatlyin weight illfallfroma givenheightwiththe g w samespeedofferssomedifficulty; because,f the heightis coni siderable,he retardingeffecCt the medium, hichmustbe t of w penetrated thrustasidebythe falling and body,willbe greater in the caseof the smallmomentum ofthe verylightbodythan in the caseof the greatforce[violenza] the heavybody;so of that,in a longdistance,helightbodywillbe leftbehind;if the t heightbe small,one maywelldoubt whetherthere is any difference; if therebe a difference willbe inappreciable. and it It occurredto me thereforeto repeat manytimesthe fall througha smallheightin sucha waythat I mightaccumulate allthosesmallintervals oftimethat elapse between arrival the of the heavy and lightbodiesrespecCtively their common at terminus, o that this summakesan intervalof timewhichis s notonlyobservable, easily but observable.In orderto employ theslowestpeeds ossiblendthusreduce s p a thechange which the resisting medium produces uponthe simple effecCt ofgravityit occurred meto allow to thebodies fallalong planeslightly to a inclinedo thehorizontal.Forinsucha plane,justaswellasin t a vertical lane,onemaydiscover p howbodies ofdifferent weight behave:and besidesthis, I alsowishedto rid myselfof the resistance hichmightarisefromcontactof the moving w body with the aforesaidinclinedplane. Accordingly took two I balls,oneofleadandoneofcork,the former orethana hunm dredtimes heavier thanthelatter,andsuspended thembymeans oftwoequalfinethreads,eachfourorfivecubits long. Pulling eachballaside fromtheperpendicular,letthemgoat thesame I instant,and they, fallingalongthe circumferences circles of having theseequalstrings forseml-diameters, beyond passed the perpendicular returnedalongthe samepath. This free and vibration[perlor rnedesime and.ate le tornate]epeateda le e r hundred timesshowedlearlythat theheavybodymaintains o c s nearlythe periodof the lightbodythat neitherin a hundred

[Iz9]

swings

FIRST DAY 85 swings evenin a thousandwillthe formeranticipate nor the latter by as muchas a singlemoment[minirnonomento], r so perfecdy dotheykeepstep. Wecanalsoobserve etTecCt the of themedium which,bythe resistance hichit offers motion, w to diminishes thevibration ofthecorkmorethanthat ofthe lead, butwithoutaltering frequency feither;evenwhenthe arc the o traversed bythecorkdidnotexceed fiveorsixdegrees hile w that ofthe leadwasfiftyor sixty,theswings ereperformed w inequal times. S_rP.If this be so,whyis notthe speedofthe leadgreater thanthat of thecork,seeing theformerraverses that t sixtydegrees thesame in interval nwhich i thelattercovers carcely s six? S_mv. ut whatwouldyou say, Simplicio, both covered B if theirpathsinthesametimewhenthecork,drawn aside through thirty degrees,raversesan arc of sixty,whilethe leadpulled t asideonlytwo degrees traversesan arc of four? Wouldnot thenthe corkbe proportionately swifter?Andyet suchis the experimental fa_. But observe this:havingpulledasidethe pendulum oflead,saythroughanarcoffiftydegrees, andsetit free, it swingsbeyondthe perpendicular almostfiftydegrees, thus describingn arc of nearlyone hundreddegrees; n the a o returnswing describes littlesmaller it a arc;and aftera large number fsuchvibrations finallycomes rest. Eachvibrao it to tion, whetherof ninety, fifty, twenty,ten, or four degrees occupies sametime:accordinglyhe speedof the moving the t bodykeepson diminishingincein equalintervals time,it s of traverses arcswhichgrow smaller ndsmaller. a Precisely thesamethingshappen withthependulum ofcork, suspended a stringof equallength,exceptthat a smaller by numberof vibrations requiredto bringit to rest, sinceon is account fitslightness islessableto overcome resistance o it the oftheair;nevertheless thevibrations, whetherargeorsmall, re l a allperformedn time-intervals i which arenotonlyequalamong themselves, alsoequalto theperiodof the leadpendulum. but Henceit is true that, if whilethe leadis traversing arc of an fifty degrees corkcovers of onlyten, the corkmoves the one moreslowly thanthelead;but ontheotherhandit isalsotrue that

86

THE TWO NEW SCIENCES OF GALILEO

[ 3o]
that thecorkmaycover anarcoffiftywhile theleadpasses over one ofonlyten or six;thus, at differentimes,we havenow t thecork,nowthe lead,moving morerapidly.But if thesesame bodiestraverseequalarcsin equaltimeswe mayrest assured thattheirspeeds reequal. a Snvn_.hesitate admitthe conclusiveness I to ofthisargument becauseof the confusion whicharisesfromyourmakingboth bodiesmovenowrapidly,nowslowlyand nowvery slowly, whichleavesme in doubtas to whethertheir velocities are always qual. e SAOR. Allowme, if you please,Salviati,to say just a few words. Nowtell me, Simpliclo, whetheryou admitthat one cansaywithcertainty that the speeds ofthe corkand thelead areequalwhenever oth,startingfromrestat thesamemoment b and descending sameslopes, lwaystraverseequalspaces the a in equaltimes._ S_P. Thiscanneitherbedoubted norgainsaid. SAoR. Nowithappens,nthecaseofthependulums, each i that ofthemtraverses owanarcofsixtydegrees, n nowoneoffifty,or thirtyor ten oreightor fouror two,etc.;and whentheyboth swing through arcofsixtydegreesheydoso in equalinteran t valsoftime;thesamethinghappens henthearcisfiftydegrees w or thirtyortenor anyothernumber; nd therefore econclude a w that thespeed oftheleadinanarcofsixtydegreessequalto the i speedofthe corkwhenthe latteralsoswings through arcof an sixtydegrees;n the caseof a fifty-degreerc thesespeeds i a are alsoequal to eachother; so also in the caseof other arcs. But this isnot sayingthat the speedwhichoccursin an arc of sixty is thesameas that whichoccursin an arcoffifty;nor is the speed inanarcoffiftyequalto that inoneofthirty,etc.;but the smallerhearcs,thesmallerhe speeds; hefactobserved t t t is that oneandthe samemoving bodyrequireshe sametimefor t traversing largearcof sixtydegrees a asfor a smallarcoffifty orevena verysmall rcoften;allthesearcs,indeed, recovered a a in the sameintervaloftime. It is truetherefore the lead that

[I3I]

and

FIRST DAY 87 and the corkeachdiminish theirspeed[rnoto] proportion in as theirarcsdiminish; ut this doesnot contradicCt facchat b the t theymaintain equalspeedsnequalarcs. i My reasonforsayingthesethingshasbeenratherbecause I wantedto learnwhetherI had correcCtly understood Salviati, thanbecause thoughtSimplicioadany needofa clearer xI h e planation thanthat givenbySalviafi hich w likeeverythinglse e ofhis is extremely lucid,so lucid,indeed,that whenhe solves questions hichare di_cult not merely appearance, in w in but realityand in fact, he doesso with reasons, bservations o and experiments whicharecommon ndfamiliar everyone. a to In thismannerhehas,asI havelearned fromvarious ources, s givenoccasiono a highlyesteemed t professor forundervaluing his discoveries the groundthat theyare commonplace, on and established pona meanand vulgarbasis;as if it werenot a u most admirableand praiseworthy feature of demonstrative science that it springsfromand growsout of principles ellw known, nderstoodndconcededyall. u a b But let us continue with this lightdiet;and if Simplicios i satisfiedto understandand admit that the gravityinherent [internaravith] various g in falling bodies hasnothing dowith to the difference speedobservedamongthem, and that all of bodies,in so far as their speeds dependuponit, wouldmove withthe samevelocity,praytellus, Salviati, owyouexplain h the appreciablend evidentinequality motion; lease a of p reply alsoto theobjecCtion urgedbySimplicio--anbjecCtion o inwhich I concur--namely, a cannon that ballfallsmorerapidlythana bird-shot.Frommypointofview, nemightexpecCt o thedifferenceof speedto besmallin thecaseofbodies ofthe samesubstancemoving throughany single medium, hereas larger w the oneswilldescend, uringa single d pulse-beat, distance a which thesmaller neswillnottraverse anhour,orin four,oreven o in in twentyhours;as for instancein the caseof stonesand fine sandand especiallyhat veryfinesandwhichproduces uddy t m waterandwhichinmanyhours willnotfallthrough asmuchas twocubits,a distance hich w stones notmuchlarger illtraverse w in a single pulse-beat. Salv.

88 THE TWO NEW SCIENCES OF GALILEO SALV. atdonof the mediumin producinga greater The retardation uponthosebodies whichhavea lessspecificravity g has alreadybeenexplained y showing theyexperience b that a diminution ofweight. But to explain howone and the same mediumproducessuchdifferent etardationsin bodieswhich r are madeof the samematerialand have the sameshape,but differonlyin size,requires discussion orecleverthan that a m by whichone explains owa moreexpanded h shapeor an opposing motionof the mediumretardsthe speedof the moving body. The solution ofthe presentproblem lies,I think,in the roughnessnd porositywhichare generally almostnecesa and sarily foundin thesurfaces fsolid o bodies.Whenthebodyis in motiontheseroughplacesstrikethe air or other ambientmedium. The evidence this is foundin the humming for which accompanies rapidmotionofa bodythrough the air,evenwhen that bodyisasroundaspossible.Onehears notonlyhumming, butalsohissing andwhistling, wheneverhereisanyappreciable t cavityor elevation uponthe body. We observealsothat a roundsolidbodyrotatingin a latheproduces currentofair. a Butwhatmoredoweneed? Whena topspins onthegroundat its greatestspeeddo we not hear a distinctbuzzingof high pitch? Thissibilantnotediminishesn pitchas the speedof i rotationslackens, whichis evidencehat thesesmallrugosities t onthesurface meetresistance intheair. Therecanbenodoubt, therefore,hat in the motionof fallingbodiestheserugosities t strikethe surroundingluidandretardthespeed;andthis they f dosomuchthemoreinproportion sthesurface a islarger, hich w isthecaseofsmall odies b ascompared ithgreater. w Sire,.Stopa momentplease, amgettingconfused.ForalI though understand I andadmitthat friction ofthemedium upon the surfaceof the bodyretardsits motionand that, if other thingsare the same,the largersurfacesuffers greaterretardation,I donotseeonwhatground yousaythatthe surface ofthe smaller odyis larger. Besides b if,as yousay,thelargersurface suffersgreaterretardationthe largersolidshouldmovemore slowly, hichis notthe fac2. But this objedtion w canbe easily met

[I3 ]

FIRST DAY 89 met by sayingthat, although larger odyhas a largersurthe b face,it has alsoa greaterweight, n comparison i withwhichthe resistance fthelargersurface o isnomorethantheresistance f o the smallsurface comparison in withitssmaller eight; othat w s the speedofthe largersoliddoesnotbecome less. I therefore seenoreason forexpecdng anydifference ofspeed solongasthe drivingweight[gravit_ ovente] m diminishes the sameproporin [i33] tion as the retarding power[facol_ritardante] the surface. of SAr.v. shallanswer I allyourobje&ions once. Youwill at admit, fcourse, o Simplicio, thatifonetakes twoequal bodies, of the same material andsame figure, bodies hich w would therefore fallwithequalspeeds, andif hediminishes theweight foneof o them in the sameproportion its surface as (maintaining the similarityfshape) ewould o h notthereby diminish thespeed of thisbody. S_v. Thisinferenceeemso beinharmony s t withyourtheory which states that the weightof a bodyhasnoeffe&in either accelerating orretarding itsmotion. SALV. quiteagreewithyou in this opinion I fromwhichit appearsto follow that, if the weight f a bodyis diminished o in greaterproportion thanits surface, themotionis retardedto a certainextent;and this retardation greaterand greaterin is proportion asthe diminution ofweight xceedshat ofthe sure t face. Sma,.ThisI admitwithout esitation. h SALV. Nowyoumustknow,Simplicio, it is notpossible that to diminish surface solidbodyin the sameratioasthe the ofa weight,and at the sametime maintainsimilarity figure. of For sinceit is clearthat in the caseofa diminishing solidthe weightgrowslessin proportion the volume, nd sincethe to a volumealwaysdiminishes morerapidlythanthe surface, hen w thesameshapeismaintained, theweight usttherefore m diminish morerapidlythanthe surface. But geometry teachesus that,in the caseof similar solids,he ratioof twovolumess t i greater thanthe ratioof theirsurfaces; which, forthe sakeof betterunderstanding, I shallillustrate ya particularase. b c Take,

90 THE TWO NEW SCIENCES OF GALILEO Take,for example, cubetwoinchesona sideso that each a facehas an areaof foursquareir/ches nd the total area,i. e., a thesumofthe sixfaces, mounts twenty-fourquareinches; a to s nowknagine thiscubeto be sawed through threetimessoasto divide intoeightsmaller ubes, achoneinchontheside,each it c e faceone inchsquare,and the total surfaceof eachcube six square inchesinsteadoftwenty-four in the caseofthe larger as [I34] surface cube. It isevident hereforehat the t t ofthelittlecubeis onlyone-fourth of the larger,namely,the ratioof six to that twenty-four; thevolume but ofthe solidcubeitselfisonlyoneeighth; hevolume, ndhencealsotheweight, iminishes t a d thereforemuch morerapidly thanthesurface.If weagaindivide the littlecubeintoeightotherswe shallhave,for the total surface ofone of these,one and one-half quareinches,whichis ones sixteenth ofthe surface ofthe original ube;but ks volumeis c only one-sixty-fourth part. Thus,by twodivisions, youseethat the volumeis diminishedour timesas muchas the surface. f And,if the subdivisione continued b untilthe original olidbe s reduced a finepowder, eshallfindthat theweightofoneof to w thesesmallest particles asdiminished h hundreds ndhundreds f a o timesasmuchasits surface.AndthiswhichI haveillustrated in the caseof cubesholdsalsoin the caseof allsimilarsolids, wherethe volumes tandinsesquialteral s ratioto theirsurfaces. Observe thenhowmuchgreatertheresistance, arising fromconta_ ofthe surface ofthe moving bodywiththe medium,n the i caseof smallbodiesthan in the caseof large;and whenone considershat the rugosities nthe verysmallsurfaces f fine t o o dustparticles reperhaps a nosmaller thanthoseonthe surfaces of largersolids whichhavebeencarefully polished, ewillsee h howknportantit is that the medium shouldbeveryfluidand offer noresistanceobeing t thrustaside, asily e yielding a small to force. Yousee,therefore, hnplicio, I wasnot mAstaken S that when,not longago,I said that the surface a smallsolidis of comparatively greater thanthat ofa largeone. Snw. I amquiteconvinced; and,believe me,if I wereagain beginning mystudies,I shouldfollow adviceof Plato and the start

FIRST DAY 9I startwithmathematics,science hich a w proceeds erycautiously v andadmits nothingasestablishedntilithasbeenrigidly u demonstrated. SAG_. discussion as afforded greatpleasure;but This h me before proceeding furtherI should liketo heartheexplanation of a phraseofyourswhich isnewto me,namely,hatsimilar olids t s areto eachotherin the sesquialteral ratiooftheirsurfaces;or f although haveseenandunderstood I theproposition inwhich it is demonstrated the surfaces f similarsolidsare in the that o duplicateratio of their sidesand also the proposition hich w proves that thevolumes rein thetriplicate a ratiooftheirsides, yet I have not so muchas heardmentioned ratioof the the volume solidto itssurface. ofa S_v. Youyourself avesuggested answerto yourquesh the tion and have removed everydoubt. For if one quantityis the cubeof something fwhichanotherquantityis the square o doesitnotfollow thatthecubeisthesesquialteral ofthesquare? Surely. Now if the surfacevariesas the squareof its linear dimensions the volume while variesasthecubeofthesedimensionsmaywe not say that the volumestandsin sesquialteral ratioto thesurface? SACR. Quiteso. Andnowalthough thereare stillsomedetails, in connection with the subjeCtunderdiscussion, conceming whichI mightaskquestions et,if wekeepmaking y one digression after another,it willbe longbeforewe reachthe maintopicwhichhasto dowiththevarietyofpropertiesound f in the resistance whichsolidbodiesofferto fraCture;and, therefore, f youplease,let us returnto the subjeCt hichwe i w originallyroposed discuss. p to SaLV. Verywell;but the questions hichwe have already w consideredre so numerous nd so varied,andhavetaken up a a so muchtime that thereis not muchof this day leftto spend uponour maintopicwhichaboundsn geometrical i demonstrationscalling forcareful onsideration. c MayI, therefore, suggest that we postpone meeting ntilto-morrow, the u notonlyforthe reasonjust mentioned alsoinorderthat I maybringwith but me
[I35]

,9z THE TWO NEW SCIENCES OF GALILEO mesome papersin which havesetdowninan orderly I waythe theorems propositions and dealingwiththe variousphasesof this subjec_t, matterswhich,frommemoryalone,I couldnot presentintheproperorder. SAog. fullyconcurin youropinionand all the morewillI inglybecausethis willleavetime to-dayto take up someof my difficulties the subjeCt hichwe havejust been diswith w cussing. One question whetherwe are to consider reis the sistance ofthe mediumas sufficient destroythe acceleration to of a body of very heavy material,very largevolume,and [I36] spherical figure.I saysphericalinrderto selecCt o a volume hich w is contained withina minimum surfaceand therefore lesssubjecco retardation. t Anotherquestiondealswith the vibrationsof pendulums whichmay be regarded from severalviewpoints;he first is t whether allvibrations,arge,medium, nd small,areperformed l a in exactlyandpreciselyqualtimes:anotheristo findtheratio e ofthe timesofvibration ofpendulumsupported s bythreadsof unequalength. l SALV. These areinteresting uestions: I fearthat here,as q but in thecaseofallotherfacets, ifwetakeupfordiscussionnyone a ofthem,itwillcarryinitswakesomanyotherfactsandcurious consequences timewill that notremainto-dayforthe discussion ofall. SAOR. theseareasfullof interestasthe foregoing, would If I gladlyspendas manydaysas thereremainhoursbetween now and nightfall;and I dare say that Simplicio wouldnot be wearied bythesediscussions. Sn_P.Certainlynot; especially whenthe questions pertain to naturalscience havenot been treatedby otherphilosand ophers. SAr.V. Nowtakingup thefirstquestion, canassertwithout I hesitationthat there is no sphereso large,or composed of materialso densebut that the resistance the medium,alof thoughveryslight,wouldcheckits accelerationnd would, n a i time reduceits motionto uniformity;a statementwhich is strongly

FIRST DAY 93 stronglysupported experiment.For if a fallingbody, as by timegoeson,wereto acquire speedasgreatasyouplease, o a n suchspeed,impressedyexternal orces[motore b f esterno], canbe sogreatbutthat thebodywillfirstacquire itandthen,owing to the resisting medium,oseit. Thus,for instance,f a cannon l i ball,having fallena distance offourcubitsthrough theair and havingacquired speedof,say,ten units[gradz] a wereto strike thesurface ofthewater,and ifthe resistancefthewaterwere o not ableto checkthe momentum [irnpeto] shot,it would ofthe eitherincreasen speedormaintain uniform i a motion untilthe bottomwerereached: utsuchisnot theobserved onthe b facet; contrary,the waterwhenonlya fewcubitsdeephinders and diminisheshe motionin sucha way that the shotdelivers t to the bed ofthe riveror lakea very slightimpulse.Clearly then if a shortfall through wateris sufficient deprivea the to cannonballofits speed,his speedcannot e regained ya fall t b b ofevena thousand cubits. Howcould bodyacquire, afallof a in a thousand cubits,thatwhichit loses falloffour?Butwhat ina moreisneeded Dowenotobserve theenormous ? that momentum, deliveredo a shotbya cannon,s sodeadenedy passing t i b througha fewcubitsofwaterthat theball,sofarfrominjuring the ship,barelystrikesit? Eventhe air,although veryyielda ingmedium,canalsodiminish speedof a falling the body,as maybe easilyunderstood fromsimilarexperiments. or if a F gunbe fireddownwards fromthe topof a veryhightowerthe shot willmakea smaller impressionponthe groundthan if u the gun had beenfiredfroman elevation onlyfouror six of cubits;this is clearevidence the momentum that ofthe ball, firedfromthe top of the tower,diminishesontinually c from the instant it leavesthe barreluntil it reachesthe ground. Therefore fallfromeversogreatan altitudewillnotsufficeo a t giveto a bodythat momentum whichit hasoncelostthrough theresistanceftheair,nomatterhowitwasoriginallycquired. o a In likemanner,thedestructive ffectproduced pona wallbya e u shotfiredfroma gunat a distance oftwentycubitscannotbe duplicated the fallof the sameshotfromany altitudehowby ever

[ 371

94 THE TWO NEW SCIENCES OF GALILEO evergreat. My opinionis, therefore,hat underthe circumt stances which occurin nature,theacceleration ofanybodyfalling fromrest reachesan end and that the resistance the of medium finallyreducesits speedto a constantvaluewhichis thereafter aintained. m SAQR. These experiments re in my opinionmuchto the a purpose;he onlyquestion whetheran opponentmightnot t is makeboldto denythe factinthe caseofbodies [moh] hichare w verylargeandheavyorto assertthat a cannon ball,falling from the distance fthemoonorfromtheupperregions ftheatmoso o phere,woulddelivera heavierblowthan if just leavingthe muzzle ofthegun. Sazv.No doubtmanyobje&ions aybe raisednot all of m whichcanberefutedby experiment: howevern thisparticular i casethe followingonsideration ustbe takeninto account, c m namely, hatit is verylikelythat a heavy t bodyfalling froma heightwill,onreachingheground, t haveacquiredustasmuch j momentum aswasnecessary carry to thatheight;as may to it be clearly seenin the caseof a rather heavypendulum which, whenpulled asidefiftyor sixtydegrees romthe vertical, f will acquire preclsely that speedandforcewhichare sufficient to carryit to anequal levationaveonlythatsmall ortion e s p which it losesthrough ffi&ion onthe air. In orderto placea cannon ballat sucha heightasmightsufficeo giveit just thatmoment tum whichthe powderimpartedto it on leavingthe gun we needonlyfireit vertically upwardsromthe samegun;and we f canthenobserve hether w onfalling backit delivers blowequal a to that ofthegunfiredatcloserange; n myopinion would i it be muchweaker. The resistance the air would,therefore, of I think,preventthe muzzlevelocityfrombeingequalled a by naturalfallfromrestat anyheightwhatsoever. Wecomenowto the otherquestions, relatingto pendulums, a subjecc hichmay appearto many exceedinglyrid, esw a peciallyto thosephilosophers ho are continually w occupied withthe moreprofound questions fnature. Nevertheless, o the problemisonewhichI donot scorn. I am encouraged the by e_ample

[3s]

FIRST DAY 95 example fAristotle o whomI admire especially ecause did b he notfailto discuss verysubjecCt hethoughtinany degree e which worthyofconsideration. Impelled yourqueries maygiveyousomeof myideas by I concerningertainproblemsn music,a splendid c i subjecCt, upon whichso many eminentmenhave written:amongthese is Aristotle himself whohasdiscussed umerous n interestingcousa ticalquestions.Accordingly, onthe basisofsomeeasyand if tangible experiments, shallexplain I somestriking phenomena in the domain ofsound,I trustmyexplanations willmeetyour approval. SAGR. shallreceive I themnot onlygratefully eagerly. but For,althoughI takepleasure everykindofmusical nstruin i mentand havepaidconsiderable attentionto harmony, have I neverbeenableto fullyunderstand whysome combinations of tonesare morepleasing thanothers,or whycertaincombinations not only fail to pleasebut are even highlyoffensive. Thenthereisthe oldproblem oftwostretched strings inunison; whenoneof themis sounded,he otherbegins vibrateand t to to emit its note; nor do I understand different atiosof the r harmony [formedleconsonanze] d andsome otherdetails. SALv. Letusseewhether ecannot erive thependulum w d from a safisfafftory solution ofallthesedifficulties. ndfirst,asto A the question whetheroneand the samependulum reallyperformsits vibrations,arge,medium, nd small,all in exacdy l a thesametime,I shallrelyuponwhatI havealready heardfrom our Academician.He has clearlyshownthat the time of descentis the samealongallchords,whateverhe arcswhich t subtendthem, as wellalongan arc of I8o (i.e., the whole diameter)as alongone of IOO0, IO0, 0,_/40, 4'. It is 6o, 2 or understood, f course,that thesearcs all terminateat the o lowest ointofthecircle, hereit touches p w thehorizontal plane. If nowwe consider escentalong d arcsinsteadoftheirchords then, provided thesedonot exceed 9o,experiment shows that theyarealltraversed inequaltimes;butthesetimesaregreater for the chordthan forthe arc,an effe_whichis allthe more remarkable

[x391

96 THE TWO NEW SCIENCES OF GALILEO remarkable becauseat first glanceone wouldthink just the opposite be true. For sincethe terminalpointsof the two to motions the sameand sincethe straightline included are betweenthesetwopointsistheshortest istance d between them,it wouldseemreasonable motionalongthis lineshouldbe that executed the shortesttime;but this is not the case,for the in shortesttimc and thereforethe most rapidmotion--isthat employedlongthe arc ofwhichthis straightlineisthe chord. a Asto thetimesofvibration ofbodiessuspendedythreadsof b different lengths, heybearto eachotherthesameproportion s t a the squarerootsof the lengths the thread;or onemightsay of thelengths areto eachotherasthesquares ofthe times;sothat ifonewishes makethe vibration-time to ofonependulum twice that ofanother,hemustmakeits suspensionourtimesaslong. f In likemanner,f onependulum i hasa suspensioninetimesas n [14o] longas another,his second t pendulum willexecute threevibrationsduringeachoneofthe first;fromwhichit followshat the t lengths the suspendingordsbearto eachother the [inverse] of c ratioofthesquares fthenumber fvibrations erformedn the o o p i sametime. SAog. hen,if I understand T youcorreCtly, caneasily I measurethe lengthof a stringwhoseupperend is attachedat any heightwhatever evenif this end wereinvisible nd I couldsee a onlythelower xtremity.Forif I attachto thelowerendofthis e stringa ratherheavyweightand giveit a to-and-fro motion, andifI aska friend counta numberofits vibrations, to whileI, duringthe sametime-interval, countthe numberof vibrations ofa pendulum which isexactly onecubitinlength, henknowing t the numberof vibrationswhicheachpendulum makesin the givenintervalof time one can determinethe lengthof the string. Suppose, example, myfriendcounts2ovibrafor that tionsof the longcordduringthe sametime in whichI count 24oofmystringwhichisonecubitinlength;takingthesquares ofthetwonumbers, oand24o,namely 2 4ooand 576oo,hen,I t say,the longstringcontains576o0 units of suchlengththat mypendulum willcontain 400ofthem;and sincethe lengthof my

FIRST DAY 97 mystringisonecubit,I shalldivide 57600 y40oandthusobb tain 144. Accordingly shallcallthe lengthof the string144 I cubits. SALv. NOrwillyoumissit by asmuchasa hand'sbreadth, especially ifyouobserve largenumber fvibrations. a o SAGI_. Yougiveme frequentoccasiono admirethe wealth t and profusion naturewhen,fromsuchcommon even of and trivialphenomena, youderivefaCts hicharenotonlystriking w and newbutwhichare oftenfarremoved fromwhatwewould haveimagined.Thousands ftimesI haveobserved ibrations o v especially churches in wherelamps,suspended y longcords, b had beeninadvertently intomotion; ut the mostwhichI set b couldinferfromtheseobservations wasthat the viewofthose whothinkthat suchvibrations aremaintained ythe medium b is highlyimprobable: for,in that case,the airmustneedshave considerable judgment ndlittleelseto dobutkilltimebypusha ingto andfro a pendentweightwithperfect egularity.But I r neverdreamed learning one and the samebody,when of that suspended froma stringa hundredcubitslongandpulled aside through arcof90orevenIor i/_o, an would employ thesame timein passing through theleastasthrough thelargest fthese o arcs;and, indeed,it still strikesme as somewhat nlikely. u NowI amwaiting tohearhowthesesamesimple henomena p can furnishsolutionsorthoseacoustical f problems---solutions which willbeat leastpartlysatisfaCtory. SALV. Firstof allonemustobserve eachpendulum that has its own time of vibrationso definiteand determinate that it is not possible makeit movewith anyotherperiod[altro to periodo] thanthat whichnaturehasgivenit. Forlet any one take in his handthe cordto whichthe weightis attachedand try.,asmuchashepleases, increase rdiminishhefrequency to o t [frequenza] its vibrations;it willbe time wasted. On the of otherhand,onecanconfer otion m uponevena heavy pendulum whichisat rest bysimply blowinggainstit;by repeatinghese a t blastswitha frequency hichis thesameasthat of thependuw lumonecanimpartconsiderable motion. Supposehat bythe t first

[I4I]

98 THE TWO NEW SCIENCES OF GALILEO firstpuffwehavedisplaced thependulum fromthe vertical y, b say,halfaninch;thenif,afterthependulum hasreturned andis aboutto beginthe second vibration, eadd a second w puff,we shallimpart additionalmotion;and so on with other blasts provided theyareapplied therightinstant,andnotwhenthe at pendulum iscoming towardus since thiscasetheblastwould in impederather than aid the motion. Continuing thus with many impulses[impulsz] e impart to the pendulumsuch w momentum [impeto] a greaterimpulse that [fore.a] thanthat ofa single blastwillbeneeded stopit. to SAGR. Even as a boy,I observedthat one man aloneby givingtheseimpulsesat the right instantwasableto ring a bell so largethat whenfour,or even six,menseized rope the and tried to stop it they wereliftedfromthe ground,allof themtogetherbeingunableto counterbalance momentum the whicha single man,by properly-timed pulls,had givenit. SALV. Yourillustration akesmymeaning m clearand isquite as wellfitted,aswhatI havejust said,to explain thewonderful phenomenonfthe strings o ofthe cittem[cetera] orofthe spinet [cimbalo], namely,the fa& that a vibratingstringwill set anotherstringinmotionandcauseitto sound notonlywhenthe latterisinunison evenwhenit differsromtheformer yan but f b octaveor a fifth. A stringwhichhas been struckbeginsto vibrate and continuesthe motionas longas one hears the sound[risonanza]; thesevibrationscausethe immediatelyurs rounding airto vibrateandquiver;thentheseripples the air in expandfar intospaceand strikenotonlyall the strings the of sameinstrumentbut even thoseof neighboringnstruments. i Since stringwhichis tunedto unison that withthe oneplucked is capableof vibratingwith the samefrequency,t acquires, i at the first impulse,a slightoscillation; fter receiving a two, three,twenty,or moreimpulses, delivered properintervals, at it finallyaccumulates vibratorymotionequalto that of the a plucked string,as is clearlyshownby equality ofamplitude in theirvibrations.Thisundulation expands throughthe air and sets intovibrationnotonly strings, ut alsoany other body b

[ 42]

which

FIRST DAY 99 whichhappensto havethe sameperiodasthat ofthe plucked string. Accordingly we attachto the sideof an instrument if smallpieces ofbristleorotherflexible odies, eshallobserve b w that, whena spinetis sounded, onlythosepieces respondthat havethe sameperiodas the stringwhichhasbeenstruck;the remaining piecesdonotvibratein responseo this string,nor t dothe former ieces p respond anyothertone. to If one bowsthe base stringon a violarather smartlyand bringsnearit a gobletoffine,thin glasshavingthe sametone [tuono] asthat ofthestring,thisgobletwillvibrateandaudibly resound. That the undulations f the mediumare widely o dispersed boutthe soundingodyis evinced a b bythe facet a that glassofwatermaybemadetoemita tonemerely bythefri&ion of thefinger-tip uponthe rimofthe glass;for in this wateris produced seriesof regularwaves. The samephenomenon a is observed betteradvantage fixing baseof the goblet to by the uponthebottomofa ratherlargevessel fwaterfilled o nearlyto the edgeof the goblet;forif, asbefore, e soundthe glassby w fricCtion the finger,we shallseeripplesspreading of withthe utmostregularity andwithhighspeedto largedistances bout a the glass. I haveoften remarked,n thus sounding rather i a [1431 ' largeglassnearlyfullofwater,that at firstthewaves arespaced with greatuniformity, nd when,as sometimes appens,the a h toneof the glassjumpsan o&avehigherI havenotedthat at this momenteachof the aforesaid wavesdMdesinto two; a phenomenon hichshows learlythat the ratioinvolvedn the w c i o&ave[forma dell'ottava] istwo. SAGP. MorethanoncehaveI observed thissame thing,much to mydelightand alsoto myprofit. For a longtimeI have been perplexed about thesedifferentharmonies incethe exs planationshithertogivenby thoselearnedin musicimpress meas not sufficiently conclusive.Theytellus that the diapason,i. e. theocCtave, involvesheratiooftwo,that thediapente t whichwecall the fifthinvolves ratioof 3:2,etc.;becauseif a the openstringof a monochord e sounded afterwards b and a bridgebe placedin the middleand the halflengthbe sounded one

Ioo THE TWO NEW SCIENCESOF GALILEO onehearstheoctave;andifthe bridge beplaced 1/3thelength at of the string,thenon plucking theopenstringand afterfirst wards 2/3ofits lengththe fifthis given; orthis reason f theysay that the ocCtave dependsupon the ratio of two to one [contenuta duee l'uno]and the fifthuponthe ratioof threeto tra'l two. This explanation doesnot impress as sufficiento me t establish 2and3/2asthenaturalratios oftheoctaveandthefifth; and my reasonfor thinkingso is as follows.Thereare three different aysin whichthe toneof a stringmaybe sharpened, w namely,by shortening by stretchingit and by makingit it, thinner. If the tensionand sizeof the stringremainconstant oneobtains theoc2ave yshortening b ittoone-half, i.e.,bysoundingfirsttheopenstringandthenone-half fit; butiflengthand o sizeremainconstant ndoneattemptstoproduce a theoccave by stretchinghe willfindthat it doesnot suffice doublethe to stretching weight;it mustbe quadrupled; sothat, if the fundamentalnoteis produced a weight f onepound,fourwillbe by o required bring to outtheoccave. Andfinallyif the lengthand tensionremainconstant,while onechangeshe size* ofthe stringhewillfindthat in orderto t produce ortavethe sizemustbe reduced _ that which the to gavethe fundamental.And what I have saidconcerninghe t octave,namely,that its ratioas derivedfromthe tensionand sizeofthe stringis the squareof that derivedfromthelength, appliesequally wellto all other musicalintervals[intemalli musicz].Thusif onewishesto produce fifthby changing a the lengthhefindsthat theratioofthelengths mustbesesquialteral, in otherwordshe sounds the openstring,thentwo-thirds first ofit;butifhewishes toproduce thissameresultbystretching or thinningthe stringthen it becomes necessaryto squarethe ratio3/2that is bytaking9/4 [dupla sesguiquarta]; accordingly, if the fundamental equiresa weightof 4 pounds,the higher r notewillbe produced by 6, but by 9 pounds;the sameis not truein regardto size,the stringwhichgives fundamental the is largerthan that whichyieldsthe fifth in the ratio of 9 to 4. In viewofthesefacets, seenoreasonwhythosewisephilosI For theexact meaning of"size"seep.io3below. Trans.] [

[x44]

FIRST DAY IOI ophers shouldadopt2 ratherthan4 as the ratioof theo_ave, orwhyin thecaseofthefifththeyshould employ thesesquialteralratio, /2,ratherthanthat of9/4- Since isimpossible 3 it to countthevibrations fa sounding o stringonaccount f itshigh o frequency, shouldstillhavebeen in doubtas to whethera I string,emitting upperochre, madetwiceas manyvibra_ the tionsin the sametime as one givingthe fundamental, it had notbeenforthefollowing namely, fa_, thatat theinstantwhen thetonejumpsto the oC%ave, thewaveswhichconstantly accompany vibratingglassdivide intosmaller neswhich the up o arepreciselyalfaslongastheformer. h SALV. is a beautifulexperiment This enabling to distinus guishindividuallyhe waves t whichareproduced thevibraby tionsofa sonorous body,whichspread through theair,bringing tothetympanum oftheeara stimulus hich w themindtranslates intosound.But sincethese waves inthewaterlastonlysolong as the fri_don the fingercontinues are,eventhen, not of and constant ut arealways b forming nd disappearing, a wouldit not be a finething if onehad the abilityto produce waveswhich wouldpersist fora longwhile,evenmonthsandyears,soas to easily measure andcountthem ? SAcR. uchan invention S would,assureyou,command I my admiration. SALv. Thedevice isonewhich hituponbyaccident; I mypart consists erely theobservation andin the appreciation m in ofit ofitsvalueas a confirmation ofsomething whichI hadgiven to profound onsideration; yet the deviceis,in itself,rather c and common.As I wasscraping brassplatewitha sharp iron a .[145] chiselin orderto removesomespotsfromit and wasrunning the chiselratherrapidlyoverit, I onceor twice,duringmany strokes, eardtheplateemita ratherstrong clear h and whistling sound;onlooking the platemorecarefully, noticed long at I a rowoffinestreaksparalleland equidistant romone another. f Scraping with the chisel verandoveragain,I noticedthat it o wasonlywhenthe plate emittedthis hissingnoisethat any markswereleftuponit;whenthescraping wasnotaccompanied by

io2 THE TWO NEW SCIENCESOF GALII.F.O bythissibilant otetherewasnotthe leasttraceofsuchmarks. n Repeating trickseveraltimesaridmakingthe stroke,now the withgreaternowwithlessspeed,thewhistling followed itha w pitchwhichwascorrespondingly higherand lower. I noted also that the marksmade whenthe toneswerehigherwere closertogether;but when the toneswere deeper,they were fartherapart. I alsoobserved when,duringa single that stroke, the speedincreased towardthe end the soundbecame sharper andthestreaks grewcloserogether, ut alwaysn suchawayas t b i to remainsharplydefined equidistant.Besides henever and w the strokewasaccompanied hissing felt thechiseltremble by I in mygraspanda sortofshiver un through r myhand. In short we seeand hearin the caseof the chiselprecisely which that is seenand heardin the caseof a whisperfollowed y a loud b voice;for,whenthe breathis emittedwithoutthe produ_ion of a tone,onedoesnot feeleitherin the throat or mouthany motionto speakof in comparison iththat whichis feltin the w larynxand upperpart of the throat whenthe voiceis used, especially whenthetonesemployedrelowandstrong. a At timesI havealsoobserved amongthe strings ofthe spinet twowhich werein unison withtwoofthe tonesproduced bythe aforesaid scraping;and amongthosewhichdifferedmost in pitchI foundtwo whichwereseparatedby an intervalof a perfe_ fifth. Uponmeasuringhe distancebetween markt the ingsproduced bythe twoscrapings wasfoundthat the space k whichcontained 45ofone contained of the other,whichis 3 preciselyhe ratioassignedo thefifth. t t But nowbeforeproceedingny fartherI want to callyour a attentionto thefar that,of thethreemethods forsharpening a tone,the one whichyou referto as the fineness the string of should attributedto its weight. Solongas the materialof be

[i46]

the stringis unchanged, sizeand weightvary in the same the ratio. Thusin the caseofgut-strings, weobtainthe oc_r_ve by makingonestring4 timesas largeas the other;so alsoin the caseof brassonewiremusthave4 timesthe sizeof the other; butif nowwewishto obtainthe odtave gnt-string, ofa byuseof brass

FIRSTDAY Io3 brasswire,wemustmakeit, not fourtimesas large,but four timesas heavyas the gut-string: s regards a sizetherefore the metalstringis not fourtimesas bigbut fourtimesas heavy. Thewiremaytherefore eventhinnerthanthegut notwithbe standing thefaCtthat the lattergiveshehighernote. Henceif t twospinets arestrung,onewithgoldwiretheotherwithbrass, and if the correspondingtringseachhave the samelength, s diameter, ndtension followshat the instrument a it t strungwith goldwillhavea pitchaboutone-fifth lowerthanthe otherbecausegoldhas a densityalmosttwicethat ofbrass. Andhere it is to be notedthat it is the weightratherthanthe sizeofa moving bodywhich offers resistance change fmotion[_elocith to o delmoto] ontraryto whatonemightat firstglance c think. For itseems reasonable tobelieve abodywhich that islargeandlight should suffer reater g retardation ofmotioninthrusting asidethe mediumthan wouldone whichis thin and heavy;yet here exactly theoppositestrue. i Returning nowto the original ubje&ofdiscussion, assert s I that the ratioof a musical ntervalis not immediately i determinedeitherby the length,size,or tensionof the stringsbut ratherby theratiooftheirfrequencies, is,by the number that of pulsesof air waveswhichstrikethe tympanum the ear, of causing alsoto vibratewiththe samefrequency.This faCt it established, wemaypossibly explain whycertainpairsofnotes, differing pitch producea pleasingsensation, thers a less in o pleasanteffect,and stillothersa disagreeableensation.Such s an explanation ouldbe tantamount o an explanationf the w t o moreor lessperfectconsonances of dissonances. heunand T pleasantsensation produced the latterarises,I think,from by thediscordant ibrations ftwodifferent v o toneswhich strikethe ear out of time [sproporzionatamente]. Especially harshis the dissonance etweennotes whosefrequenciesre incommenb a surable;sucha caseoccurswhenonehas twostrings unison in and sounds neofthemopen,together itha part oftheother o w whichbearsthe sameratioto itswholelengthas the sideof a squarebearsto the diagonal;his yieldsa dissonanceimilar t s
to

[471

Io THE TWONEW SCIENCES F GALIT.F.O 4 O to the augmented fourthor diminished fifth[tr/tono semio diapente]. Agreeable consonances pairsof toneswhichstrikethe are ear witha certain regularity; regularityonsistsn the this c i fa&that thepulsesdeliveredythe twotones,in the same b interval oftime,shall ecommensurable b innumber, soasnotto keep theeardruminperpetual torment, bending intwodifferent dire&ions orderto yieldto the ever-discordant in impulses. The firstand mostpleasing consonance therefore, is, the o&ave since, oreverypulsegiven the tympanum f to bythe lower string,hesharp t stringdelivers two;accordinglyevery at othervibration f the upperstringbothpulses delivered o are simultaneously sothatone-half theentire number fpulsesre o a delivered unison.Butwhentwostrings in areinunison their vibrations alwayscoincidendthe effe& that of a single a is string; ence h wedonotrefer it asconsonance. to Thefifthis alsoa pleasingnterval inceforeverytwovibrations i s ofthe lower tring upperonegivesthree,sothatconsidering s the the entire number ofpulsesromtheupperstring f one-thirdfthem o willstrike inunison, i.e.,betweenach e pairofconcordant vibrationsthereintervenewosingle t vibrations; andwhenthe intervalisa fourth,hreesingleibrations t v intervene.ncase I the interval second hereheratiois9/8it isonlyevery isa w t ninth vibration the upperstringwhichreacheshe earsimultaof t neously ithoneofthelower; w alltheothers arediscordant and produce harsheffe& a uponthe recipient earwhich interprets them asdissonances. S_v. Won'tyoubegood enougho explain t thisargument a littlemore clearly ? SAr.v. LetAB denote lengthof a wave[lospazio la the e dilatazione d'unavibrazione] emittedby the lowerstringand CDthatofa highertringwhich s isemitting theoctave ofA.B; divide ABin the middle t E. If the twostrings egin a b their motions andC,itisclearhatwhen atA t thesharp vibrationas h reached theendD, theothervibration illhave w travellednly o asfarasE,which, notbeing aterminal point, ill mitnopulse; w e bu_thereisa blow delivered D. Accordingly theone at when wave

,: ! i i

FIRST DAY Io5 wavecomes ackfrom to C,theotherpassesonfromE to ]3; b D hencethe twopulsesfromB and C strikethe drumof the ear simultaneously.Seeingthat these vibrationsare repeated againand againin the samemanner, e conw dude that eachalternatepulsefromCD falls&, E B in unison with onefromAB. But eachofthe , [I48] " pulsations t the terminalpoints,A and ]3,is A _ o s a constantly accompanied byonewhich leaves t , , , alwaysfromC or alwaysfromD. Thisis clear; : becausefwesupposehewaves reach i t to Aand O D C at the sameinstant,then, whileonewave Fig.13 travelsfromA to B, the otherwillproceedfromC to D and back to C, so that wavesstrikeat C and B simultaneously; during thepassagefthewavefromBbackto.&thedisturbance o at C goesto D and againreturnsto C,so that oncemorethe pulsesatAandCaresimultaneous. Nextlet the vibrations ABand CD be separated y an inb tervalofa fifth,that is,bya ratioof3/2;choosehepointsE and t Osuchthat theywilldivide thewavelengthofthelowerstring intothreeequalpartsandimagine thevibrationso startat the t sameinstantfromeachoftheterminals andC. It isevident A that whenthe pulsehasbeendelivered the terminal , the at D waveinAB hastravelled onlyasfar asO;the drumofthe ear receives,herefore, nlythe pulsefromD. Then duringthe t o return oftheonevibration fromDto C,theotherwillpassfrom Oto B andthenbackto O,producingnisolated a pulse B--a at pulsewhichis out of timebut onewhichmustbe taken into consideration. Nowsincewehaveassumed the firstpulsationstarted that s fromthe terminals and C at the sameinstant,it followshat A t the second pulsation, isolated atD, occurred fteran interval f a o timeequalto that required forpassageromCto D or,whatis f the samething,fromAto O;butthe nextpulsation,heoneat t B, is separated fromthe preceding onlyhalfthis interval, by namely,thetimerequired forpassagerom0 to B. Nextwhile f theonevibration travels from OtoA,theothertravels fromCto D,

106 THE TWO NEW SCIENCES GALILEO OF D,the result ofwhichs i thatwopulsationssimultat occur neously and Cyclesthis follow after atA D. of kind one another,., solitaryof lower interposed i. one e pulsethe strkng between solitarypulsesof the upperstring. Let us now two imaginetime to be dividedinto very smallequal intervals; thenif we assume that, duringthe firsttwooftheseintervals, thedisturbances which occurred simultaneously atAandChave travelled asfarasOandD andhaveproduced pulse D; and a at if we assumethat duringthe third and fourth intervalsone disturbanceeturnsfromD to C,producing pulseat C,while r a the other,passingonfromO to B and backto O, produces a pulseat B;andif finally, uring d thefifthandsixthintervals,he t disturbancesravelfromO and C to A and D, producing t a pulseat eachofthe lattertwo,thenthe sequencen whichthe i pulsesstrikethe earwillbe suchthat, ifwebegin counttime to fromany instantwheretwo pulsesare simultaneous, ear the drumwill,afterthe lapseoftwoofthe saidintervals,receive a solitarypulse;at theend ofthe thirdinterval, nothersolitary a [I491 puIse;so alsoat the end of the fourth interval;and two intervalslater,i. e.,at theend ofthe sixthinterval, illbeheard w twopulses unison.Hereendsthe cycle--theanomaly, to in so speak--which repeatsitselfoverandoveragain. SAGR.cannolongerremain I silent;forI mustexpresso you t thegreatpleasure haveinhearing I sucha completexplanation e of phenomena with regardto whichI have so longbeen in darkness.NowI understand whyunison doesnotdifferroma f singletone;I understand whythe ocCtave the principal aris h mony,but solikeunison asoftento bemistaken it andalso for whyit occurswith the other harmonies.It resembles nison u becausehepulsations fstrings t o inunison always ccur o simultaneously, ndthoseofthe lowerstringofthe o&ave always a are accompanied those by oftheupperstring;andamongthelatter is interposed solitarypulseat equalintervalsand in sucha a mannerasto produce nodisturbance;he resultisthat sucha t harmonyis rathertoomuchsoftened lacksfire. But the and fifthischaracterized byitsdisplaced eatsandbytheinterposib tion

FIRST DAY lO7 tionof twosolitary beatsof theupperstring onesolitary and beat of the lowerstringbetweeneachpairof simultaneous pulses;thesethreesolitary ulsesareseparated intervals p by of time equalto half the intervalwhichseparates eachpair of simultaneous beatsfromthe solitarybeatsofthe upperstring. ThustheeffecCt the fifthis to produce tickling of a ofthe ear drumsuchthatits softnesssmodified ithsprightliness, i w giving at thesamemomentheimpression gentle t ofa kissandofa bite. SALv. Seeing that youhavederivedso muchpleasure from thesenovelties, mustshow I youa method bywhich theeyemay enjoythe samegameastheear. Suspend threeballsoflead,or other heavymaterial,by meansof stringsof differentength l suchthat whilethe longest akestwovibrations shortest m the willmakefourandthe medium three;thiswilltakeplace when the longest tringmeasures 6,eitherinhandbreadths r inany s 1 o otherunit,the medium andtheshortest ,allmeasured 9 4 inthe sameunit. Nowpullall thesependulums sidefromthe perpendicular a and releasethemat the sameinstant;you willseea curious interplay the threadspassingeachotherin various of manners but suchthat at the completion f everyfourthvibrationof o the longestpendulum, ll threewillarrive a simultaneously at the sameterminus, hence w theystartoveragainto repeatthe samecycle.Thiscombination ofvibrations, whenproduced on stringss precisely whichyieldsthe interval f the ocCtave i that o and the intermediate fifth. If weemploy samedisposition the

[ 5o1

of apparatusbut changethe lengthsof the threads,always however sucha waythattheirvibrations orrespond those in c to ofagreeable usical ntervals, e shallseea different rossing m i w c ofthesethreads always uchthat,aftera definite but s interval f o timeand aftera definite number fvibrations, llthe threads, o a whether threeor four,willreachthesameterminus t the same a instant,andthenbegin repetition fthecycle. a o If howeverhe vibrations t oftwoor morestringsareincommensurableo that theynevercomplete definite s a numberof vibrations t the sameinstant_ r if commensurable a o theyreturn only

Io8 THETWO NEWSCWNCES GALILVO OF

onlyaftera longintervalof timeand aftera largenumberof vibrations,hentheeyeis confused y thedisorderlyuccession t b s of crossedthreads. In like mannerthe ear is painedby an irregular equencefairwaves s o which strikethetympanum withoutanyfixed order. But, gentlemen, whitherhavewe driftedduringthesemany hoursluredonbyvarious roblems ndunexpecCted p a digressions ? The day is alreadyendedand we have scarcely touchedthe subjecCt proposed fordiscussion. ndeedwehavedeviated I sofar that I remember nlywithdifficulty urearlyintroducCtion o o and the littleprogress adeinthe wayofhypotheses principles m and foruseinlaterdemonstrations. SAng.Letus thenadjourn forto-dayin orderthat ourminds may find refreshmentn sleepand that we may return toi morrow,f sopleaseyou,andresume discussion i the ofthe main question. SALV.shallnot failto be hereto-morrow thesamehour, I at hopingnot only to ,renderyou servicebut alsoto enjoyyour company.
END OF THI_FIRST DAY.

[I5I]

SECOND DAY
._, AGR.WhileSimpliciond I wereawaiting a yourarrival ewere w tryingtorecall hatlast t consideration youadvanced sa prinwhich a ;{_iple c andbasis fortheresults ouintended y to ,_lobtain;this consideration dealt with the _[_esistance r whichallsolidsoffer fracCture to and depended upona certaincement hich w _]_eldthe partsgluedtogetherso that they h wouldyieldand separateonlyunderconsiderable [potente pull attrazzione]. ater we tried to find the explanation this L of coherence, seeking mainlyin the vacuum;hiswasthe occait t sior_ ofourmanydigressions which occupiedhe entireday and t ledus far afieldfromthe original uestion q which,as I have already stated,wastheconsideration oftheresistanceresistenza] [ that solids ffer o tofracCture. SALV. remembert allverywell. Resuming threadof I i the ourdiscourse, whateverhenatureofthisresistance hich t w solids offerto largetrafftiveforces[_olenta ttrazzione] a therecanat leastbe nodoubtof itsexistence; thoughthis resistance and is verygreatinthecaseofa direcCt it isfound, sa rule,to be pull, a lessin the caseof bending forces[nel_iolentargli traverso]. per Thus,forexample, rodofsteelorofglasswillsustain longia a tudinalpullofa thousand pounds while weight ffiftypounds a o would quitesufficient breakit if the rodwerefastened be to at right anglesintoa verticalwall. It is this second typeof resistancewhichwe must consider, seeking discovern what to i [Iszl proportion

im THE TWO NEW SCIENCESOF GALTT.V.O proportionit is foundin prismsand cylinders the same of material,whetheralikeor unlikein shape,length,and thickness. In this discussionshalltakefor grantedthewell-known I mechanical rinciplewhich has been shownto governthe p behavior f a bar,whichwe calla lever,namely,hat theforce o t bearsto the resistance inverse the ratioofthe distances hich w separatethefulcrum fromtheforceand resistanceespecCtively. r Snvw. This wasdemonstrated firstof allby Aristotle,n his i Mechanics. SALV. Yes,I amwillingto concede himpriorityin pointof time;but as regards rigorofdemonstration thefirstplace must begivento Archimedes, sinceupona single propositionroved p in his bookon Equilibrium depends onlythe law of the * not leverbutalsothoseofmostothermechanical devices. SA_R. incenowthis principleis fundamentalo all the S t demonstrations whichyoupropose setforthwouldit notbe to advisable giveus a completeand thoroughproofof this to propositionnless u possiblyt would i taketoomuchtime? SAzv.Yes, that wouldbe quiteproper,but it is better I thinkto approach sub]e& a mannersomewhat ifferent our in d fromthat employed Archimedes, by namely, y firstassuming b merely equalweights laced a balance that p in ofequalarmswill produce equilibrium--a rinciplelsoassumed yArchimedes-p a b and then proving that it is no lesstrue that unequalweights produceequilibrium when the arms of the steelyardhave lengthsinversely proportional the weightssuspended to from them;in other words,it amountsto the samethingwhether oneplacesequalweights equaldistances runequalweights at o at distances hichbearto eachotherthe inverseratioof the w weights. In orderto makethis matterclearimagine prismor solid a cylinder, B,suspendedt eachendto the rod [lineal and A a HI, supported twothreadsHA and I_B; t is evidentthat if I by i attacha thread,C,at themiddle pointofthebalance beamHI, theentireprismAB according will, totheprinciplessumed, a hang in equilibrium one-halfts weightiesonone side,and the since i l otherhalfontheotherside,ofthepointofsuspension . Now C Works ofArchlmedes. byT.L.Heath, Trans. pp.I89-22o. Trans.] [

[_ [_ _ _.; --

SECONDDAY I II supposehe prismto be dividedintounequalpartsby a plane t [ s3] the smaller: his division t havingbeenmade,imagine thread a throughthe lineD, and letthe partDAbe the largerandDB ED, attachedat the pointE and supportinghe partsAD and t relativeto lineHI: and since relativein the sameposition the position the prism of DB,in orderthat thesepartsmayremain and the beamHI remains unchanged,herecanbe no doubt t butthat theprismwillmaintain itsformer tateofequilibrium. s
H GC_B F I

Fig. 14

But circumstances wouldremainthe sameif that part of the prismwhichisnowheldup,at theends,bythe threads and AH DE weresupported the middleby a single at threadGL; and likewise otherpart DB wouldnot changeposition held the if by a threadFM placedat its middle point. Suppose owthe n threadsHA,ED, and IB to be removed,eaving l onlythe two GL and FM, thenthe sameequilibrium willbe maintained so longasthesuspension isatC. Nowletus consider thatwehave heretwoheavy bodies ADandDBhungattheendsGandF, of a balancebeamGF in equilibrium aboutthe pointC, so that the lineCG is the distancefromC to the pointof suspension ofthe heavybodyAD,whileCF is the distance whichthe at otherheavybody,DB, is supported.It remains nowonlyto showthat thesedistances earto eachotherthe inverseratio b of the weights themselves, is,the distanceGC is to the that distance CFastheprismDBistotheprismDA--aproposition which weshallproveas follows: the lineGE is thehalfof Since EH,and since isthehalfofEI, thewhole EF length GFwillbe half

iiz THE TWO NEW SCIENCESOF GALTLVD halfofthe entirelineHI, and therefore equalto CI: if nowwe subtra& thecommon artCFtheremainder Cwillbeequalto p G theremainder I, that is,to FE,and if to eachoftheseweadd F CE we shall have GE equal to CF: henceGE_F=FC:CG. But GE and EF bearthe sameratioto eachotheras do their doubles E andEI, that is,the sameratioasthe prismAD to H DB. Therefore, y equatingratioswe have,convertendo, b the distanceGCis to the distanceCF as the weightBE)is to the weight A,which D iswhatI desired toprove. If whatprecedes clear,youwillnot hesitate,I think, to is admitthat the twoprisms andDBareinequilibrium AD about the pointC sinceone-half f the wholebodyAB lieson the o rightofthe suspension and theotherhalfontheleft;inother C words, his arrangementsequivalento twoequalweights ist i t d posedatequaldistances.I donotseehowanyonecandoubt,if thetwoprisms DandDBwere A transformedntocubes, pheres, i s orintoanyotherfigure whatever ndifGandFwereretained a as pointsof suspension, they wouldremainin equilibrium that aboutthepointC,forit isonlytooevidentthat change offigure doesnotproduce change ofweightsolongasthe mass[quantit_ di material doesnotvary. Fromthiswemayderive thegeneral conclusion any two heavy bodiesare ha equilibrium that at distanceswbAch inverselyproportionalo their weights. are t This principleestablished, desire,beforepassingto any I othersubje&, o callyourattentionto thefa&that theseforces, t resistances, moments, igures, tc.,maybe considereditherin f e e the abstra&,dissociatedrommatter,or in the concrete, ssof a ciatedwith matter. Hence the propertieswhichbelongto figures that are merelygeometrical non-material and must be modified whenwe fill thesefigures with matter and therefore givethem weight. Take, for example,the leverBA which, restinguponthe supportE, is usedto lift a heavystoneD. Theprincipleust demonstrated akesit clearthat a forceapj m plied at the extremityB willjust sufficeto equilibratethe resistance offeredby the heavy bodyD providedthis force [momento] bearsto theforce[momento] the sameratioasthe atD distance

[,54]

SECOND DAY II3 distance Cbearsto thedistance A CB;andthisistruesolongas weconsider onlythemoments fthesingle o force B andofthe at resistancet D, treatingtheleverasanimmaterial a bodydevoid ofweight. But if wetake intoaccount heweightofthe lever t itselfmaninstrument hichmaybe madeeitherofwoodor of w iron--itismanifesthat,whenthisweight asbeenadded the t h to forceat B, the ratiowillbe changedand must therefore e b expressedn differenterms. Hencebefore i t goingfurther let

[Iss]

Fig. 15

us agreeto distinguishetween b thesetwopointsofview; hen w we consider n instrumentn the abstra&,i. e.,apartfromthe a i weightof its ownmaterial, e shallspeakof "takingit in an w absolute sense"[prendere assolutamente]; butifwefilloneofthese simple andabsolute figures ithmatterand thusgiveitweight, w we shall referto sucha materialfigureas a "moment"or "compoundorce"[momento f oforzacomposta]. SAoR. mustbreakmyresolution I aboutnot leading youoff into a digression; I cannotconcentrate y attentionupon for m what is to followuntil a certaindoubtis removed frommy mind,namely,you seemto compare forceat B withthe the total weightof the stoneD, a part of whichmpossibly the greaterpart--rests uponthe horizontalplane:so that SALV.understand I pede&ly: ouneedgonofurther. How" y everplease observe I havenotmentionedhetotalweight that t ofthe stone;I spoke onlyofits force[momento] the pointA, at theextremityofthe leverBA,whichforceis always lessthan the total weightof the stone,and varieswithits shapeand elevation. SAc_.Good:but thereoccurs me-another uestion to q about which

I I4 THE TWO NEW SCIENCESOF GALTI.F.O whichI am curious. For a complete understanding this of matter, I shouldlikeyou to show me,if possible, owone can h determine what part of the total weightis supportedby the underlying plane and what part by the end A of the lever. SALV. explanation The willnot delayus long and I shall therefore havepleasure granting in yourrequest. In theaccompanyingfigure,let us understand that the weighthavingits centerofgravityat A restswiththe end]3uponthehorizontal planeand with the other end uponthe leverCG. Let N be thefulcrum levertowhichtheforce ofa [potenza] isapplied G. at Let falltheperpendiculars, AOand CF,fromthe centerAand the end C. Then I say,the magnitude [momento] the entire of weightbears to the magnitudeof the force [momentoella d po_enza] G a ratiocompounded at ofthe ratiobetween two the

Fig.16 distances GN and NC and the ratio betweenFB and BO. Layoffa distance suchthat itsratioto NCisthesameasthat X ofBOto FB; then,sincethe total weight is counterbalanced A bythetwoforces t BandatC,itfollowshat theforce Bisto a t at that at C as the distanceFO is to the distanceOB. Hence, compone_o, sumofthe forcesat B and C,that is,the total the weight [momento tutto'l pesod], isto theforceat C asthe A di lineFB is to the lineBO,that is,as NCis to X: but the force [momento dellapotenza] appliedat C is to the forceappliedat G as the distanceGN is to the distanceNC; henceit follows, excequali l_roportione in perturbata,* the entireweight is that A to the forceapplied G as thedistance is to X. But the at GN ratioofGN toX iscompounded oftheratioofGNtoNC andof NC toX, that is,ofI_B B0; hencetheweight bears the to A to *Fordefinition ofperturbataTodhunter's Book ,Def. o. see Euclid. V 2 [Trans.]

[ 56]

SECONDDAY _15 equilibratingorceat G a ratiocompounded that ofGN to f of NCandofFBto BO:which wastobeproved. Let us nowreturnto our original ubjecCt; if whathas s then, hithertobeen said is clear,it willbe easilyunderstood that,

PROPOSITION I
A prismor solidcylinder f glass, teel,wood o s orotherbreakablematerialwhich iscapable ofsustaining veryheavyweight a whenapplied longitudinally as previouslyemarked, asily is, r e brokenbythe transverse applicationf a weight hichmaybe o w muchsmallernproportion sthe length i a ofthecylinder xceeds e its thickness. Let us imagine solidprismABCDfastened a intoa wallat theendAB,and supporting weight at theotherend;undera E standalsothat thewallisvertical ndthat theprismorcylinder a is fastened rightanglesto thewall. It is clearthat, if the at cylinderbreaks,fracturewilloccurat the pointB wherethe edgeofthemortiseacCts fulcrum asa forthe leverBC,to which theforce isapplied;hethickness fthesolid t o BAistheotherann oftheleveralong which islocated theresistance.Thisresistance opposeshe separation the part BD,lyingoutsidethewall, t of fromthat portionlyinginside. Fromthe preceding, follows it that themagnitude [momento] oftheforceapplied Cbearsto at themagnitudemomento] [ oftheresistance, found inthethickness ofthe prism,i. e.,in theattachment fthe baseBAto its cono tiguous parts,the sameratiowhich lengthCBbearsto half the the lengthBA;if nowwedefine absolute resistanceo fracCture t asthat offered a longitudinal to pull(inwhichcasethestretchingforceacCts the samedirecdonsthat through in a whichthe bodyismoved), henit followshat the absolute t t resistancef o the prismBDis to the breaking loadplacedat the endof the leverBCin thesameratioasthelengthBCisto thehalfofAB in the caseof a prism,or the semidiameter the caseof a in cylinder. Thisisour firstproposition.* Observe in what that *Theonefundamental whichsimplicitly error i introduced intothis
proposition and which is carried through the entire discussionof the

[i57]

II6 THE TWO NEW SCIENCESOF GALII.VO has herebeen saidthe weightof the solidBE)itselfhas been leftoutof consideration, rather,the prismhasbeenassumed or to bedevoidofweight. But if theweightofthe prismis to be takenaccount fin conjuncCtion o withtheweight ,wemustadd E to theweight one E half that of the prismBD: so that if, for example, the latter weighstwo pounds and the weightE is ten pounds we must treat the weightE as if it wereeleven D pounds. SL_P. hy not W twelve ? SALv. weight The E,my dear Simplicio, at hanging the extreme C acCts end upon lever the BC with _alloits m ment tenounds: of p so also would the Fig.17 solid BD if suspendedat the same point exertitsfilllmoment f twopounds; o but, as you know,this solidis uniformly distributedthroughSecondDay consistsin a failureto seethat, in sucha beam,there must be equilibriumbetweenthe forcesof tensionand compression over any cross-section.The correctpoint of viewseemsfirstto have beenfound by E. Mariotte in I68Oand by A. Parent in I7I3. Fortunately this error does not vitiate the conclusionsof the subsequentpropositions which deal only with proportlons--not actual strength--of beams. Following Pearson(Todhunter'sHistoryof Elasticity)one might say K. that Galileo's mistakelay in supposingthe fibresofthe strainedbeamto be inextensible.Or, confessing anachronism, nemight say that the the o errorconsistedin taking the lowestfibreof the beamas the neutral axis.

[Tra_;.]

SECOND DAY ii 7 out its entirelength, C,so that the partswhichlie nearthe B endB arelesseffeCtive thanthose moreremote. Accordinglyf we strike a balancebetweenthe two, the i weightof the entireprismmaybe considered concentrated as at its centerof gravitywhichliesmidway the leverBC. of But a weighthungat the extremityC exertsa momenttwice as greatas it wouldif suspended fromthe middle: therefore if weconsiderhe moments fbothaslocatedat theend C we t o mustaddtotheweight one-halfhatoftheprism. E t Sn_P.I understand perfeCtly; andmoreover, mistake ifI not, the forceofthe twoweights BDand E, thusdisposed, would exertthesamemomentaswouldtheentireweightBID together with twicethe weightE suspended the middle the lever at of BC. S_J_v. Precisely and a faCtworth remembering. ow so, N wecanreadily understand PROPOSITION II Howand in whatproportion rod,or rathera prism,whose a widthis greaterthan its thickness offersmoreresistance to fracCture whenthe forceis appliedin a__ the direCtionf its cV___-_r-_I] o J breadth thaninthe direCtion it s of _b-__ ___t_ For the sakeof thickness. __:_-'_ -clearness,take a ruler ad and _ _ width is acwhose whose thickness, Fig.i8 cb,is muchlessthan its width. Thequestion nowis whywill the ruler,if stoodon edge,as in the firstfigure,withstanda greatweightT, while,whenlaidflat,as in the second figure, it willnot supportthe weightX whichis lessthanT. The answeris evidentwhenwe remember hat in the one case t the

[is8]

Iz8 THE TWO NEW SCIENCESOF GAL!!.F.O the fulcrumis at the line bc, and in the other caseat ca, whilethe distanceat whichthe forceis appliedis the samein both cases,namely,the length be/:but in the first casethe distance ofthe resistanceromthe fulcrum--halfhe lineca-f t is greaterthan in the othercasewhereit is onlyhalf of bc. Therefore weightT is greaterthanX in the sameratio as the halfthe widthcaisgreaterthanhalfthe thickness c,sincethe b formerarts as a leverarmfor ca,and the latterfor cb,against the sameresistance, namely,the strength ofallthefibres the in cross-section Weconclude, ab. therefore, any givenruler, that or prism,whosewidthexceeds thickness, illoffergreater its w resistanceo fracturewhenstanding edgethanwhenlying t on flat,andthisin theratioofthewidthto thethickness. PRoPosrrmr III Considering nowthecaseofa prismorcylinder growingonger l in a horizontal irection,we mustfind out in what ratiothe d momentof its ownweightincreases comparison in with its resistanceo fracture. ThismomentI findincreasesnproport i
[I59]

tionto the squareofthe length. In orderto provethis letAD bea prismorcylinderying l horizontal withitsendA firmly fixed in a wall. Let thelengthoftheprismbe increased ythe addib tion of the portionBE. It is clearthat merelychanging the length oftheleverfrom toACwill,ifwedisregardtsweight, AB i increasehemoment fthe force[attheend]tendingto produce t o fracCture A in the ratioofCA to BA. But, besideshis, the at t weight fthe solidportionBE, addedto theweightofthe solid o ABincreaseshe moment f thetotal weightin theratioofthe t o weightof the prismAE to that of the prism_A_B, whichis the sameastheratioofthelength ACtoAB. It follows,herefore, hat, whenthe lengthand weightare t t simultaneously increased any givenproportion, moment, in the whichistheproductofthesetwo,isincreasedna ratiowhichis i the squareofthe preceding proportion.Theconclusion then is that the bendingmomentsdue to the weightof prismsand cylinders hichhavethe samethickness differentengths, w but l bear

bearto eachothera ratiowhichis the squareof the ratioof SECOND DAY 119 theirlengths, r,whatisthesamething,theratioofthesquares o oftheirlengths. We shallnextshowin what ratiothe resistanceo frac_re t

Fig. 19

[bending strength],n prismsand cylinders, i increases ith inw

[ 6o]

crease thickness hilethe lengthremains of w unchanged.Here I saythat

PIIOPOSlTION IV

In prismsand cylinders f equallength,but of unequal o thicknesses, theresistance fra_ureincreasesnthesame to i ratioas thecubeofthe diameter the thickness, e., of of i. thebase. LetAandB betwocylindersfequallengths G,FH;lettheir o D bases becircular utunequal, aving diameters DandEF. b h the C ThenI saythatthe resistance tofradture ffered ythe cylinder o b B

i2o THE TWO NEW SCIENCESOF GALII,V,O B isto that offered byAasthe cubeofthe diameter E istothe F cubeof thediameter C. For,if weconsiderhe resistanceo D t t fragtureby longitudinal pullasdependent ponthe bases, . e., u i uponthecircles andDC,noonecandoubtthat the strength EF [resistenza] the cylinderB is greaterthan that of A in the of sameproportion inwhichthe areaofthe circle exceedshat EF t of CD;becauseit is precisely this ratiothat the number f in o fibres bindingthe partsofthe solidtogetherinthe onecylinder exceedshat intheothercylinder. t But in the caseof a forceacCting transversely mustbe reit memberedhat weareemploying t twolevers inwhichthe forces C areapplied atdistances G, D FH, and the fulcrumsare locatedat the pointsD and C D F; but the resistances are 1_appliedat distances which ........... areeqiaalo the radiiofthe t the fibresdistributedover circlesDC cross-secCtions I_these entireand EF, since Fig.2o accsif concentratedt the a a centers. Remembering and rememberinglsothat the this a arms,DG and FH, throughwhichthe forces and H acCt G are equal,we can understandthat the resistance,ocatedat the l centerof the baseEF, acting againstthe forceat H, is more effecCtive [maggiore] the resistance the center of the than at base CD opposing forceG, in the ratioof the radiusFE the to the radiusDC. Accordingly resistance fracCture the to offeredby the cylinder isgreaterthan that of the cylinder B A in a ratiowhichiscompounded ofthat of theareaofthe circles EF and DCand that oftheirradii,i. e.,oftheirdiameters; ut b the areasofcircles areasthe squares oftheirdiameters.Thereforethe ratioof the resistances, beingthe producc the two of preceding ratios,isthesame asthat ofthecubes ofthediameters. ThisiswhatI setoutto prove. Alsosincethevolume fa cube o [I6I] variesas the third powerof its edgewe maysay that the resistance l-I_

SECONDDAY IZI sistance [strength] fa cylinder hoselengthremains o w constant varies asthethirdpower fitsdiameter. o Fromthepreceding eareableto conclude w that CORO_L__RY The resistance [strength] f a prismor cylinder f constant o o length variesinthesesquialteral ratioofitsvolume. Thisis evidentbecause volume prismor cylinder f the ofa o constantaltitudevariesdirecCtly areaof itsbase,i.e.,as asthe the square sideordiameter fthisbase;but, asjust demonofa o strated,the resistancestrength] [ variesasthe cubeofthis same sideordiameter.Hence theresistancearies v inthesesquialteral ratioof the volume consequentlylsoof the weight--of a the soliditself. S_v. Before proceedingurtherI should f liketo haveoneof mydifficulties removed.Up to this pointyouhavenot taken into consideration certainother kindof resistance hich,it a w appears me,diminishes thesolidgrowslonger, ndthis is to as a quiteastruein the caseofbending in pulling;t isprecisely as i thusthat in thecaseofa rope weobservehat a verylongoneis t lessableto supporta largeweightthana shortone. Whence, I believe, shortrodofwood a orironwillsupport greater a weight thanif it werelong,provided force always the be applied longitudinallyand nottransversely, provided and alsothat wetake intoaccountheweight ftheropeitselfwhich t o increases ithits w

length.

SALV. fear, Simplicio, I correcdyatchyourmeaning, I if c that inthisparticular ouaremaking y thesame mistake asmany others;that isifyoumeanto saythata long rope, neofperhaps o 4 cubits,cannotholdup sogreata weightasa shorterlength, sayoneortwocubits, fthesame o rope. SIMV. ThatiswhatI meant, ndasfarasI seetheproposition a ishighly probable. SALV. the contrary,I considert notmerelyimprobable On i but false;and I think I caneasily convince ouof yourerror. y LetABrepresentherope,fastened theupperendA: at the t at lowe.r nd attach a weightC whoseforceis just sufficient e to break

122 THE TWO NEW SCIENCESOF GALILEO breaktherope. Now,Simplicio, pointouttheexactplace where youthinkthebreakoughttooceur. [1621 Sn_. Let ussayD. S_v. Andwhyat D? SrMe. ecause this pointthe ropeisnot strongenough B at to support, ay,IOO s pounds, adeupoftheportionoftheropeDB m andthestoneC. SALV. Accordingly heneverhe ropeisstretched[violentata] w t withtheweight fIOO o pounds Dit willbreakthere. at Sna_.I thinkso. S_v. But tell me,if insteadof attaching weightat the the end of the rope,B,one fastensit at a pointnearer D, say, at E: or if, insteadof fixing upperend the ofthe ropeat A,one fastensit at some pointF, just A. aboveD, willnottherope,at thepointD, be subjeCt to thesamepullofIOO pounds ? Sn_P.It would,providedyou includewith the P stoneCtheportionofropeF__.B. IB SALV. us thereforesuppose Let that the rope is stretched atthepointD witha weight fIOO o pounds, then according yourownadmission willbreak; to it but FE isonly a smallportionofAB;howcanyou therefore maintainthat the longropeisweakerthan the shortone? Give up then this erroneous view B whichyousharewithmanyvery intelligent eople, p and let us proceed. Now havingdemonstrated that, in the caseof [uniformly loaded]prismsand cylinders f constant o thickness, he momentof forcetendingto produce t Fig.2I fracture[momentooprale proprieresistenze] s varies as the squareof the length;and havinglikewise shownthat, whenthe lengthisconstantandthe thickness aries,the resistv anceto fraCture variesas the cubeof the side,or diameter, ofthe base,letus passto theinvestigation fthe caseof solids o which simultaneously varyinbothlengthandthickness.HereI observe that,

'i

SECONDDAY PROPOSITION V

xz3

Prismsand cylinders which differin both length and thickness fferresistanceso fraffture o t [i.e.,cansupportat their endsloads]whichare directlyproportionalo the t cubesofthe diameters ftheirbasesandinversely o proportionalto theirlengths. [I65] LetABCandDEF betwosuchcylinders; thenthe resistance [bending strength] f thecylinder Cbearsto the resistancef o A o the cylinder F a ratiowhichisthe produCt fthecubeofthe D o diameter Bdivided A bythecubeofthediameter E,andofthe D lengthEF dividedby the A length BC. Make EG _--_ equal to BC:let H be a B third proportionalo the t C lines AB and DE; let I D [AB/DE =H/I]: and let - _ _: be a fourth proportional,_- ---, ".... I :S=EF:BC. G F Nowsince theresistanceA_ B of the cylinderAC is to D'. 'E that of the cylinderDG t_ as thecubeofABisto the cubeofDE, that is,asthe I _ lengthABis to the length $' I; and sincethe resistance Fig. z z ofthe cylinder G isto that of the cylinder F asthe length D D FE isto EG,that is,asI is to S, it followshat the length t AB is to S as the resistance the cylinder C is to that of the of A cylinder F. But the lineAB bearsto S a ratiowhichis the D produCtof AB/I and I/S. Hence the resistance[bending strength] fthe cylinder Cbearstothe resistancef the cylo A o inderDF a ratiowhichis the produCt f AB/I(that is,AB3/ o DE8)and of I/S (that is,EF/BC):whichis whatI meantto prove. This proposition havingbeen demonstrated, us next let consider

_z THE TWO NEW SCIENCESOF GALILEO 4 consider caseof prismsand cylinders hichare similar. the w Concerning theseweshallshow that,
PROPOSITION VI

Inthe ofsimilar case cylinders and prisms,moments the [stretching forces] whichresultfrommultiplying together theirweightand length[i.e.,fromthemoments produced by their ownweightand length], hichlatter acCts s a w a lever-arm, bearto eachother a ratiowhichis the sesquialteralofthe ratiobetween resistances the oftheirbases. In orderto provethis letusindicate twosimilar the cylinders byABandCD:thenthemagnitude ftheforce[momento] o inthe cylinder B,opposing resistancef itsbaseB, bearsto the A the o magnitude [momento] oftheforceat CD,opposing theresistance of its base D, a ratiowhichis the sesquialteral f the ratio o [I641 between resistance the ofthe baseB and the resistance f the o baseD. Andsince the A BSolids Band CD,are A effe&ivein opposing the resistances oftheir basesB andD, in proC m..... =-_D portion totheirweights and to the mechanical Fig.23 advantages[forze]of theirleverarmsrespectively, since and theadvantage [forza] f o theleverarmABisequalto the advantage [forza] fthe lever o armCD (thisis true becausein virtueof the similarity the of cylindershe lengthAB is to the radiusof the baseB asthe t length CDisto theradius ofthebaseD),it followshat thetotal t force[momento] ofthecylinder Bisto thetotalforce[mornento] A ofthe cylinder D astheweightaloneofthe cylinder Bis to C A the weightaloneof the cylinderGD,that is,as thevolumeof the cylinderAB [l'istesso cilindro AB] is to the volumeCD [all'istesso D]:but these are as the cubesof the diameters C oftheirbasesB and D; and the resistances the bases,being of
to

SECONDDAY I25 to eachother as their areas,are to eachother consequently asthesquares ftheirdiameters.Therefore o theforces [moment,] ofthecylinders reto eachotherinthesesquialteral a ratioofthe resistanceftheirbases.* o SLMP. Thispropositiontrikes s measbothnewandsurprising: at first glanceit is very different rom'anything hichI myf w self shouldhave guessed:for sincethesefiguresare similar in all other respects,I shouldhave certainlythoughtthat the forces [moment,] andtheresistances ofthesecylinders ould w havebornetoeachotherthesame ratio. SAG_. Thisistheproofoftheproposition which referred, to I at theverybeginning ofourdiscussion, oneimperfectly as understood byme. SALv. Fora while, implicio,usedto think,asyoudo,that S I the resistances ofsimilar olids eresimilar; a certaincasual s w but observationshowed that similarsolidsdo not exhibita me strength whichisproportional theirsize,thelarger nesbeing to o lessfittedto undergo roughusagejust astallmenaremoreapt than smallchildren be injuredby a fall. And, as we reto markedat the outset,a largebeamor columnfallingfroma [x651 given height illgotopieces henunderthesamecircumstances w w a smallscantling rsmall arble o m cylinder willnotbreak. It was this observation whichledme to the investigationf the fact o whichI amaboutto demonstrate toyou:itisa veryremarkable thingthat,among infinite the varietyofsolids which aresimilar oneto another,herearenotwoofwhichthe forces t [moment,], and theresistances thesesolids of arerelatedinthe sameratio. SLuP. Youremind menowofapassage inArstofle's uestions Q
* The precedingparagraph beginning with Prop.VI is of more than usualinterest asillustratingthe confusion terminology of current in the time of Galileo. The translationgiven is literal exceptin the caseof those wordsfor whichthe Italian is supplied. The facts whichGalileo has in mind are so evident that it is difficult o see howone can here t interpret "mo_nent"to mean the force "opposingthe resistance its of mechanical dvantageof the levermadeup of AB and theradius of the a baseB"; and similarlyfor "theforceof theleverarmCD."

base," nlesstheforce thelever rra betakento mean"the u " of a dg" [Trans.]

i26 THE TWO NEW SCIENCESOF GALILEO inMechanics inwhich hetriesto explain whyit isthat a wooden beambecomes eakerand canbemoreeasilybentas it grows w longer, otwithstanding faCtthat the shorterbeamis thinn the ner and the longerone thicker:and, if I remember orrectly, c heexplainst intermsofthe simple i lever. SALv. Verytrue: but, sincethis solutionseemedto leave roomfordoubt,BishopdiGuevara,* whose truly learned commentarieshave greatlyenrichedand illuminated work, this indulgesn additional leverspeculations i c withthehopeofthus overcomingll difficulties; a neverthelessvenhe is confused e as regardshis particular oint,namely,whether, t p whenthe length and thickness f thesesolidfigures o increase the sameratio, in theirstrengthand resistanceo fraCture, swellasto bending, t a remain constant. Aftermuchthoughtuponthis subjeft,I have reached following the result. FirstI shallshow that, PROPOSITION VII Amongheavyprismsand cylinders f similar o figure,there is one and only one whichunderthe stressof its own weightlies just on the limit betweenbreakingand not breaking:so that everylargerone is unableto carrythe loadofitsownweightandbreaks;while everysmaller one isable towithstand some additional tending force tobreakit. Let ABbe a heavyprism,the longestpossiblehat willjust t sustainitsownweight, othat if it belengthened leastbit it s the willbreak. Then,I say,this prismisuniqueamongallsimilar prisms--infinite number--inoccupying boundaryline in that between breakingand not breaking;so that everylargerone [i661 willbreakunderits ownweight, nd everysmaller newillnot a o break,but willbe ableto withstand some forcein addition its to ownweight. Let the prismCE be similarto, but largerthan, AB: then, I say, it willnot remaininta_ but willbreakunderits own weight. Layoffthe portionCD, equal inlengthoAB. And, t since, heresistancebendingtrength] fCDisto that ofAB as t [ s o
* BishopofTeano; b. 156I; d. I64L [Trans.]

:_

SECONDDAY 127 the cubeofthethickness fCDistothe cubeofthethickness f o o AB,that is,astheprismCEisto thesimilar rismAB,it follows p that the weight f CEis theutmostloadwhicha prismofthe o lengthCDcansustain;but the lengthofCE isgreater;thereforethe prismCE willbreak.A Nowtake anotherprismFG _B P_--"-_" _t G which is smallerthan AB. Let FH equalAB,thenit can __,___ beshownin a similarmannerC ............ E that the resistance[bending D strength]of FG is to that of Fig.24 AB as the prismFG is to the prismAB providedthe distanceABthat is FH, is equalto the distanceFG; but AB is greaterthan FG, and therefore momentof the prism the FG applied G isnot sufficient breaktheprismFG. at to SAG_. Thedemonstration isshortandclear; hile w theproposition which,at first glance,appearedimprobable nowseen is to be bothtrueandinevitable.In orderthereforeo bringthis t prisminto that limitingcondition whichseparates breaking fromnot breaking, t wouldbe necessary changethe ratio i to between thickness nd lengtheitherbyincreasinghethickness a t orby diminishing length. Aninvestigation thislimiting the of statewill,I believe, emand d equalingenuity. SALV. Nay,evenmore;forthequestion ismoredifficult; this I know because spentnosmallamount ftimein itsdiscovery I o whichI nowwishto share withyou. PRoPosmo_ VIII Givena cylinder r prismof the greatestlengthconsisto entwithitsnot breaking nderitsownweight;andhaving u givena greaterlength,to find the diameter another of cylinder r prismofthis greaterlengthwhichshallbethe o onlyandlargest necapable fwithstanding o o itsownweight. Let BCbe the largestcylinder apable c ofsustainingts own i weight;andletDE be a lengthgreater thanAC:theproblem is to findthe diameter the cylinder hich,havingthe length of w [I67] DE,

Tz8 THE TWO NEW SCIENCESOF GALILEO DE, shallbe the largestone just ableto withstandits own weight. Let I be a third proportional the lengthsDE and to AC;let the diameter FDbe to the diameter AasDE is to I; B drawthe cylinderFE; then, amongall cylinders havingthe sameproportions, is the largestand onlyonejust capable this ofsustaining itsownweight. Let M be a third proportional DE and I: alsolet Obe a to fourthproportional DE, I, and M; layoffFG equalto AC. to Nowsincethediameter isto the diameter Basthelength FD A DE isto I, and since Oisa fourthproportional DE,I andM, to it followshat FD3:BAS=DE:O. ut the resistance t B [bending _. C strength]of the cylinder G is D to the resistance f the cylinder o ]_BCasthe cubeof FD isto the cubeofBA:hence resistance the P ofthe cylinder G isto that of D :t, ' ' ! G cylinder Casthe lengthDE is B _a,J, o: : to O. And sincethe moment Fig. 5 2 of the cylinderBC is held in equilibrium [?equale lla]its resistance, by a weshallaccomplish ourend (which provethat themomentofthe cylinder E isto F is equalto the resistanceocatedat FD), if we showthat the l moment fthecylinder E isto themoment fthecylinder C o Y o B asthe resistance F isto the resistance A,that is,asthe cube D B of FD isto the cubeofBA,or as the length isto O. The DE momentof the cylinderFE is to the momentof the cylinder DG asthe squareofDE isto the squareofAC,that is, asthe length is to I; but the moment fthe cylinder G isto the DE o D moment fthe cylinder C,asthe square o B ofDFis to thesquare ofBA,that is,asthesquare ofDE isto thesquare or asthe ofI, square is to the square fM, or,asI isto O. Therefore ofI o by equating ratios,it resultsthat the moment fthe cylinder E is o F to themoment fthecylinder Casthelength isto O,that o B DE is,asthecubeofDFisto the cubeofBA,orastheresistance f o the baseDFisto the resistancefthe baseBA; hich tobe o w was proven. SAGI_. demonstration, This Salviati,is ratherlongand difficult

SECONDDAY x29 cult to keep in mind from a single hearing. Will you not, therefore, begoodenough repeatit? to SALV. Asyoulike;butI wouldsuggestnsteada moredire_ i and a shorterproof:this will,however, ecessitate different n a _o figure. _-_ [x68] ;" SAGR. favorwillbe that muchgreater:nevertheless The I _ hopeyouwilloblige meby puttingintowrittenformthe argus i mentjustgiven sothat I maystudyit atmyleisure. SALV. shallgladlydoso. LetA denote cylinder fdiamI a o ' eter DC and the largestcapable ofsustaining ownweight: its ; the problemis to determine largercylinder hichshallbe a w at oncethe maximum andtheuniqueonecapable ofsustaining itsownweight. Let E be sucha cylinder, imilarto A, havingthe assigned s length,andhaving diameter L. LetMN bea thirdpropora K tionalto thetwolengths CandKL: D n A let MN alsobe the diameter an- c_ of other cylinder, , havingthe same X lengthasE: then,I say,X isthe cylinder sought.Now sincethe resist- M anceofthe baseDC is to the resistanceofthe baseRE as the square of DCis to the square ofI{_L, is,as that the squareof KL is to the squareof Fig.26 MN,or,as the cylinder is to the cylinder , that is, as the E X momentE is to the moment ; and sincealsothe resistance X [bending strength] f the baseKL is to the resistance f the o o baseMN as the cubeof KL is to the cubeof MN,that is, asthe cubeofDC isto thecubeofKL,or,asthecylinder is A to the cylinder . that is,asthemolfient E ofAisto themoment of E; henceit follows, x cecualin proportione e i perturbata,* that themoment fAisto themoment fX astheresistancef o o o the baseDC is to the resistance the baseMN; therefore of momentand resistance re relatedto eachotherin prismX a preciselystheyareinprismA. a _
* For definition perturbataeeTodhunter'sEuclid,BookV,De{.20. of s [Trans.]

I3o THE TWO NEW SCIENCESOF GALILEO Let us now generalizehe problem;then it willread as t follows: Given a cylinderAC in whichmomentand resistance [bending strength]are relatedin any mannerwhatsoever; let DE be the lengthof anothercylinder;thendetermine whatits thickness must be in orderthat the relationbetweenits momentand resistance shallbe identicalwith that ofthecylinder AC. UsingFig.25in the samemanneras above,wemaysaythat, sincethe momentof the cylinder E is to the momentof the F portionDGas thesquareofED is to thesquareofFG,that is, as thelengthDE is to I; and sincethemomentofthe cylinder FGis to themomentofthe cylinder C asthesquareofFD is A to thesquareofA_B, or,asthesquareofED is to the square of I, or, asthe squareof I is to the squareofM, that is, as the lengthI is to O; it follows, aequali, themomentof the ex that [I69] cylinder E is to themomentofthe cylinder Cas thelength F A DE is to O, that is,as the cubeof DE isto the cubeof I, or, asthe cubeofFD isto thecubeofAB,that is,astheresistance ofthebaseFD isto theresistancefthebaseAB;which o wasto beproven. Fromwhathas a/ready beendemonstrated, youcanplainly seethe impossibility increasinghe sizeofstructuresto vast of t dimensions eitherin art or in nature;likewise theimpossibility of building ships, alaces, r temples fenormousizeinsucha p o o s waythat theiroars,yards,beams,iron-bolts, and,in short,all theirother parts willhold together;nor can natureproduce treesof extraordinary sizebecausethe branches wouldbreak downundertheirownweight;so alsoit would impossible be to buildupthebonystruCturesfmen,horses, o orotheranimals so asto holdtogetherand perform theirnormalfunCtions these if animalswereto be increased enormously height;for this in increasein heightcanbe accomplished onlybyemploy_g a material hichisharderandstrongerhanusual,orbyenlarging w t the sizeofthe bones,thus changing theirshapeuntilthe form and appearance f the animals o suggesta monstrosity.This is perhaps

SECOND DAY x3I perhapswhat our wisePoet had in mind,whenhe says,in describinghugegiant: a "Impossiblet isto reckon i hisheight "Sobeyond measure ishis size." * To illustratebriefly,I have sketched bonewhosenatural a length has been increased three times and whosethickness has beenmultiplied until,for a correspondingly largeanimal, it wouldperform same_ncCtion hichthe smallboneperthe w formsfor its smallanimal. Fromthe figures hereshownyou can see howout of proportionthe enlargedbone appears. Clearly thenif onewishes maintain greatgiantthesame to ina proportionflimbasthat o _ _t_ foundinanordinary man he must either find a harderand stronger am terial for making the [17o] bones, rhe mustadmit o a diminution strength of in comparison ithmen w ofmedium stature;forif his heightbe increased Fig. 27 inordinately hewillfall and becrushed underhisownweight. Whereas,f the sizeof a bodybe diminished, strengthof i the that bodyisnotdiminished the sameproportion; in indeed the smallerthe bodythe greaterits relativestrength. Thus a smalldogcouldprobablycarryon hisbacktwoor threedogs of hisownsize;but I believe a horsecould carryeven that not oneofhisownsize. SrM1,. Thismaybeso;butI amledto doubtit onaccount f o the enormousizereached certainfish,suchas the whale s by which,I understand, ten timesas large an elephant; et is as y theyallsupportthemselves. SALv. Your question,Simplicio, suggests anotherprinciple,
Non si pu_ compartir quantosia lungo, Si smisuratamente tuttogrosso. k Ariosto'sOrlando Furioso,XVII, o [Trans.] 3

_3z THE TWO NEW SCIENCESOF GALILEO onewhich hadhithertoescaped myattention andwhichenables giants and other animalsof vast sizeto supportthemselves and to moveaboutaswellas smaller animals This result do. maybe secured eitherby increasinghe strengthof the bones t and otherpartsintendedto carrynotonlytheirweight ut also b the superincumbent load;or, keepingthe proportions f the o bonystructureconstant, he skeleton t willholdtogetherin the samemanneror even moreeasily,providedone diminishes, in the proper proportion,the weightof the bony material, of the flesh,and of anythingelsewhichthe skeletonhas to carry. It is this second principle whichisemployed nature by in the strucCture fish,makingtheir bonesand muscles of not merely lightbutentirely devoid ofweight. SzMP. trend of your argument,Salviatl,is evident. The Since fishliveinwaterwhich onaccount ofitsdensity [corpulenza] or, as others would say, heaviness[gravitY] diminishes the weight[peso] f bodiesimmersed it, youmeanto saythat, o in for this reason,thebodiesoffishwillbe devoidofweightand willbesupported withoutinjuryto theirbones. But this isnot all;for although remainder f the bodyofthe fishmaybe the o withoutweight,therecan be noquestion that theirbones but haveweight. Takethe caseofa whale's rib,havingthe dimensions beam;whocandenyitsgreatweight ritstendency ofa o to goto thebottomwhenplacedinwater? Onewould, herefore, t [I7I] hardlyexpect hesegreatmasses sustain t to themselves. SAT.V. Averyshrewd objecCtion! Andnow,in reply,tellme whether youhaveeverseenfishstandmotionlesst willunder a water,neitherdescending the bottomnor risingto the top, to withouttheexertion offorce byswimming? Sire,.Thisis a well-knownhenomenon. p SALv. facethenthat fishare ableto remainmotionless The t underwaterisa conclusive reason forthinking thematerial that of theirbodieshas the samespecific gravityas that ofwater; accordingly, in theirmake-upthereare certainparts which if are heavier thanwatertheremustbe otherswhicharelighter, forotherwiseheywould t notproduce equilibrium. Hence

C-

SECOND DAY I33 Hence,if the bonesareheavier,it isnecessaryhat themust clesorotherconstituents fthebodyshould o belighterin order that their buoyancymay counterbalance weightof the the bones. In aquatic animalsthereforecircumstances just are reversed fromwhattheyarewithlandanimalsnasmuchs,in i a the latter,the bonessustain notonlytheirownweight ut also b that ofthe flesh,whilein the formerit is the flesh whichsupportsnot onlyits ownweightbut alsothat ofthe bones. We must thereforeceaseto wonderwhy theseenormouslyarge l animals inhabitthe waterratherthan theland,that isto say, theair. Snvn,. am convinced I onlywishto add thatwhatwe I and call landanimals oughtreallyto be calledair animals, eeing s that theylivein theair, aresurroundedy air,andbreatheair. b SAG_. have enjoyedSimplicio's I discussionncluding oth i b the questionraisedand its answer. MoreoverI can easily understand oneofthesegiantfish,if pulledashore, ould that w notperhaps sustain itselfforanygreatlengthoftime,butwould be crushedunderits ownmass as soon as the connecCtions between thebones gaveway. SALV. am inclined youropinion; nd, indeed,I almost I to a think that the samethingwould happenin the caseof a very bigshipwhichfloatson the seawithoutgoingto piecesunder [I72] its loadofmerchandisendarmament, ut whichondry land a b and in air wouldprobably apart. But let us proceed fall and show how: Givena prismor cylinder, lso its ownweightand the a maximum loadwhichit can carry,it is then possible to finda maximum lengthbeyond whichthe cylinder annot c be prolonged ithoutbreaking w underitsownweight. Let AC indicateboth the prismand its ownweight;also let D represent he maximum t loadwhichthe prismcan carry at the end C withoutfracCture; is required findthe maxit to withoutbreaking.DrawAHof sucha lengththat theweight of the prismACis to the sumofACand twicethe weightD
as

imum towhich thelengthfthesaid o prism canbeincreased

134 THE TWO NEW SCIENCESOF GALILEO asthe lengthCAisto AH;and letAGbe a meanproportional between and AH; then, I say,AG is the lengthsought. CA Since momentofthe weight[momento the gravante] attached D at thepointCisequalto themoment fa weighttwiceaslarge o asDplacedat themiddle pointAC,through whichtheweight f o ....... the prismAC a&s,it fol__ .... -x'[-_. ......... l-ttowsthat the momentof l A the resistance f the prism o k _]__ AClocatedat A isequivalentto twicethe weightD plustheweight fAC,both o acCting throughthe middle Fig.28 pointofAC. Andsince we have agreedthat the momentof the weightsthus located, namely,wiceD plusAC,bearsto the moment fACthesame t o ratiowhichthelengthHAbearsto CAand sinceAGisa mean proportional between thesetwolengths, t followshat themoi t mentoftwiceD plusACisto themoment fACas the square o ofGA is to the square CA. But the momentarisingfrom of theweight[momento premente] oftheprismC_ isto themoment ofACasthe square ofGAis to the square ofCA;thence AGis themaximum length sought,hat is,the length t upto whichthe prismACmaybeprolongedndstillsupportitself, ut beyond a b which willbreak. it Hitherto we have consideredhe moments t and resistances of prismsand solidcylinders ixedat one end with a weight f appliedat the other end;threecaseswerediscussed, namely, that in whichthe appliedforcewasthe onlyone acCting, that in whichthe weightof the prismitselfis alsotaken into consideration, that in whichthe weightof the prismaloneis and taken into consideration.Let us now consider these same [_73] prisms cylinders hensupported bothendsor at a single and w at pointplacedsomewhere betweenthe ends. In the firstplace, I remarkthat a cylinder arrying c onlyitsownweightandhaving the maximum length,beyondwhichit willbreak,will,when supported eitherin the middleor at both ends,havetwicethe length

i _: !_ ; !

SECONDDAY 135 lengthofonewhichis mortised intoa walland supported only at one end. This is very evidentbecause,if we denotethe cylinder yABCand if we assume one-half f it, AB,is b that o the greatestpossiblelength capableof supportingits own weight ithoneendfixedat B,then,forthesamereason,f the w i cylinderscarried i onthe pointG, thefirsthalfwillbe counterbalanced bytheotherhalfBC. Soalso inthecase ofthecylinder DEF,if itslengthbesuchthat it willsupport nlyone-halfhis o t B

Fig. 9 2 lengthwhenthe endD isheldfixed, rthe otherhaftwhenthe o endF isfixed,henit is evident hat whensupports, uchasH t t s and I, areplaced undertheendsD andF respecCtively themomentofany additionalorce f orweight lacedatE willproduce p fracCture thispoint. at A more intricateand difficult roblemis the following: p neglectthe weightof a solidsuchas the preceding find and whetherhe sameforce t orweight hich w producesra&ure f when applied atthe middle cylinder, ofa supported bothends,will at alsobreakthe cylinder whenapplied some at otherpointnearer oneendthantheother. Thus,forexample, onewished breaka stickby holding if to it with one hand at eachend and applying knee at the his middle, would same the force berequired tobreakit inthesame mannerifthe kneewereapplied, notat themiddle, at some but pointnearer oneend? to SAGR. problem, believe, as been toucheduponby This I h Aristotle inhisQuestions inMechanics. Salv.

I36 THE TWO NEW SCIENCESOF GALTLEO [I74] S_v. Hisinquiryhoweversnotquitethe same; orhe seeks i f merelyto discover hyit is that a stick maybe moreeasily w brokenbytakinghold,onehandat eachend ofthe stick,that is, far removedfromthe knee,than if the handswerecloser together. He givesa generalexplanation, referring to the it lengthened leverarmswhicharesecured byplacing thehandsat the endsof the stick. Ourinquirycallsfor something more: whatwewantto knowiswhether, henthehandsareretained w at the endsof the stick,the sameforceis required breakit to whereverhe kneebe placed. t SAGR. firstglancethiswouldappearto be so,because At the twoleverarmsexert,in a certainway,the same moment, eeing s that asonegrows shorter theothergrows orrespondingly c longer. SALV. Nowyouseehowreadily onefallsintoerrorandwhat caution circumspecdon and arerequired avoidit. Whatyou to have just said appearsat first glance highlyprobable, ut on b closer examination provesto be quitefar fromtrue;as will it be seenfromthe facc thatwhether theknec thefulcrum ofthe twoleverswbe placed the middle in ornotmakessucha differencethat, if fractureis to be produced anyotherpointthan at themiddle,thebreaking forceat themiddle, venwhenmultie plied ]:our,ten, a hundred,or a thousandtimes wouldnot suffice.Tobegin withweshalloffer some general onsiderations c and then passto the determination f the ratioin whichthe o breaking forcemustchange orderto produce in fraccure one at pointratherthananother. Let AB denotea woodencylinder hichis to be brokenin w the middle, verthe supporting o pointC, and letDE represent an identicalcylinder hichis to be brokenjust overthe supw portingpoint F whichis not in the middle. First of all it is clearthat, sincethe distances Cand CBare equal,the forces A applied attheextremities andAmustalsobeequal. Secondly B sincethe distance is lessthanthe distance DF ACthemoment of any forceacdng D is lessthan the momento]_he same at t forceatA, that is,applied thedistance at CA;andthemoments are lessin the ratioof thelengthDF to AC;consequently is it necessary

_ i

SECOND DAY x37 necessaryto increasethe force[momento] D in orderto overat come,or evento balance,the resistanceat F; but in comparison with the length AC the distance DF can be diminishedindefinitely:in orderthereforeto counterbalance resistanceat the F it will be necessaryto increaseindefinitely force [forza] the applied at D. On the other A C B hand, in proportionas we increasethe distance FE over 5_ that of CB,we must diminish the force at E in order to I_ P _. x:ounterbalance resistance_--_._.__i_-7-._ the at F; but the distance FE, _____ ..... measured in terms of CB, _ _ .....<--_---_<-_'_'_ cannot be increased indefiFig.3o nitelyby slidingthe fulcrumF towardthe endD; indeed,it cannot evenbe madedoublethe lengthCB. Thereforethe forcerequired at E to balancethe resistanceat F will alwaysbe more than halfthat requiredat B. It isclearthen that, asthe fulcrum F approachesthe end D, we must of necessityindefinitelyincrease the sum of the forcesapplied at E and D in order to balance,or overcome,the resistanceat F. SAGR. What shall we say, Simplicio? Must we not confess that geometry is the most powerfulof all instruments for sharpeningthe wit and training the mind to think correccly? Was not Plato perfecCtlyight when he wishedthat his pupils r should be first of all well groundedin mathematics? As for myself,I quite understoodthe property of the lever and how, by increasingor diminishingits length, one can increase or diminishthe moment of forceand of resistance;and yet, in the solutionofthe presentproblemI wasnot slightly, ut greatly, b deceived. Sire,. IndeedI beginto understandthat whilelogicis an excellentguidein discourse,it doesnot, as regardsstimulationto discovery,comparewith the power of sharp distinCtionwhich belongsto geometry. SAI_. Logic, it appears to me, teachesus how to test the conclusiveness

I38 THE TWO NEW SCIENCESOF GALILEO conclusiveness any argument r demonstrationlreadydisof o a coveredand completed; ut I donot believethat it teaches b us to discover orreCt rguments nd demonstrations. it c a a But wouldbe betterif Salviatiwereto show in just what prous portion theforces ustbeincreased m inorderto produceracture f as the fulcrumis movedfromone pointto anotheralongone andthesame wooden rod. [176] SaLv. he ratiowhichyou desireis determined T asfollows: If upon a cylinderonemarkstwo pointsat whichfracture is to be produced, hen theresistancest thesetwo t a points willbear to eachother the inverseratio of the reCtanglesormedby the distances f fromthe respeCtive pointsto the endsofthe cylinder. Let A and B denotethe leastforces whichwillbringabout fraCture the cylinderat C; likewise and F the smallest of E forces hich breakitat D. Then,I say,that thesumofthe w will forces and ]3is to the sumofthe forces and F as the area A E ofthe reCtangle AD.DBisto theareaofthe reCtangle AC.CB. Because sumoftheforces the Aand ]3bearsto the sumofthe forces andF a ratiowhich E istheproduCt ofthethreefollowing ratios,namely,(A+B)/B,]3/F,and F/(F+E); but the length BAisto thelengthCAasthesumoftheforces and]3isto the A _:force ]3; and, as the lengthDB force]3 to the forceF; also as E theto the lengthCB, so is the is lengthAD is to AB,so is the forceF to the sumof theforcesF Fig.31 andE. Henceit follows the sumof the forces and B bears that A to the sumofthe forcesE andF a ratiowhichis the produCt of the three following ratios,namely,BA/CA,BD/BC,and AD/AB.But DA/CAis the produCt f DA/BA BA/CA. o and Therefore sumoftheforces andB bearsto thesumofthe the A forcesE and F a ratiowhichis the produCt fDA:CAand o DB:CB. But the refftangle D.DBbearsto the reCtangle A AC.CBa ratiowhichis the produCt f DA/CAand DB/CB. o Accordingly AI. _D _c

SECONDDAY I39 Accordingly sumofthe forces andB isto thesumofthe the A forcesE and F as the reccangle AD.DBis to the red'tangle AC.CB, is,theresistance fracCture that to atCistotheresistance to fracCture D as the recCtangle at AD.DBis to the reccangle AC.CB. Q.E.D. [I77] Anotherratherinteresting roblem p maybe solved a conas sequence thistheorem, d namely, Giventhe maximum weight hicha cylinder prismcan w or supportat its middle-point wherethe resistancesa minii mum, and givenalsoa largerweight,find that point in the cylinder or whichthis largerweightis the maximum f loadthatcanbesupported. Let that oneof the givenweights hichis largerthan the w maximum weightsupported the middle the cylinder B at of A bearto thismaximum weightthe sameratiowhichthe length E bearsto the lengthF. The problem to findthat point is in the cylinder t whichthis largerweightbecomeshe maxa t imumthat canbe supported.Let G be a meanproportional between lengthsE and F. DrawAD and S so that they the bearto eachotherthesameratioasE to G; accordingly Swill beless thanAD. Let ADbe the diameter fa semicircle o AHD,inwhichtake AHequalto S;jointhepoints andD andlayoffDRequalto H HD. Then,I say,R is thepointsought, amely,he pointat n t whichthe givenweight, reaterthanthe maximum g supported at the middle the cylinder , wouldbecome maximum of D the load. OnABas diameter rawthesemicircle d _ANB: the pererecc pendicular N and jointhe pointsN and D. Nowsincethe R sumof the squares NR and RD is equalto the squareof on ND,that is,to thesquare ofAD,orto thesumofthe squares f o AHandHD;and,since thesquare ofliD isequalto thesquare ofDR, it followshat thesquareofNR, that is,the recCtangle t AR.RB,is equalto the squareof AH, alsotherefore the to square S;but the squareofS is to the squareofAD asthe of lengthF is to the lengthE, that is, as the maximum weight supported

I4o THE TWO NEW SCIENCESOF GALILEO supported D istothe larger fthetwogiven at o weights.Hence thelatterwillbethemaximum loadwhichcanbecarried the at pointR; which isthesolution sought. SACR. NowI understand thoroughly; amthinking andI that, since theprismABgrows onstantly c stronger ndmoreresistant a to the pressureof its load at pointswhicharemoreand more l_I removedfrom the middle,we couldin the caseof largeheavy beamscut awaya considerable A _ portion near the ends which wouldnotablylessen weight, the Z, ' ' andwhich, n the beamworkof i largerooms, ouldproveto be w It: : ofgreatutilityandconvenience. It would a finethingifonecould be discover thepropershape to givea solidin orderto makeit equallyresistantat every point, in whichcasea loadplacedat the middlewouldnot producefra&uremore easilythan if placed at any other point.* SAT.v. wasjust on the point of mentioningn interesting I a and remarkable conne&ed fa& with this veryquestion.My meaning willbe clearerif I drawa figure. Let DB represent a prism;then, aswehavealreadyshown,ts resistanceo fraci t ture[bendingtrength] t theendAD,owing loadplacedat s a toa theendB,willbelessthantheresistancet CIintheratioofthe a lengthCB to AB. Nowimagine sameprismto be cut this throughdiagonallylongthe lineFB sothat the oppositeaces a f willbe triangular;hesidefacing willbeFAB. Sucha solid t us
*The readerwillnoticethat two different roblemsare hereinvolved. p That which is suggestedin the last remark of Sagredois the following: To find a beam whosemaximumstresshas the same value when a constantloadmovesfromone end of the beamto the other. The secondproblemwtheonewhichSalviatiproceedso solve--isthe t following: To find a beam in all cross-secCtionswhich the maximumstress of is the samefora constantloadin a fixedposition. [Trans.]

Fig. 32

[_781

SECONDDAY I4I willhaveproperties different romthoseof the prism;for,if f the loadremain B, the resistance gainstfracCture at a [bending strength]at C willbe lessthan that at A in the ratioof the lengthCBto the length AB. Thisiseasily proved: orif CNO f represents cross-secCtion to AFD,thenthelengthFA a parallel bearsto the lengthCN,in the triangle FAB,the sameratio whichthe lengthAB bearsto D I weimagine and C to be the ] A .-i... the lengthCB. Therefore,fl_._.._ i placed,the leverarms in the pointsat whichthe fulcrum [ pI......... __'-'l'--,J is . twocases BA,AF andBC,CN.4, L ''_' V will be proportional[simih]. Fig.33 Hencethemoment fanyforce o applied andacCting atB through. the armBA,againsta resistance lacedat a distance p AFwill be equalto that of thesameforce B acCting at through thearm BCagainstthe sameresistance locatedat a distance CN. But now, f theforce i stillbeapplied B,the resistance overat tobe come whenthe fulcrum C,a&ingthrough armCN,is isat the lessthanthe resistance iththefulcrum A inthe sameprow at portionas the recCtangular cross-seCtion is lessthan the CO recCtangular cross-secCtion AD,that is,as the lengthCN isless thanAF,orCBthanBA. Consequently resistanceo fracCture C,offered the the t at by portion OBC, s less i thantheresistance fracCture to atA,offered by theentireblockDAB,in thesameproportion sthe length a CBissmallerhanthelength t AB. By thls diagonal aw-cut e havenowremoved s w fromthe beam,or prismDB,a portion, . e., a half,and have left the i wedge,or triangular rism,FBA. We thus have two solids p [I791 possessing oppositeproperties; ne body growsstrongeras o it is shortened whilethe othergrowsweaker. This beingso it would seemnotmerely reasonable, inevitable, there but that existsa lineofsecCtion suchthat,whenthesuperfluous material hasbeenremoved, therewillremaina solid ofsuchfigure that itwilloffer thesame resistancestrength] [ atallpoints. Simp.

I4_ THE TWO NEW SCIENCESOF GALII,FO SnaP.Evidently must,in passingfromgreaterto less, one encounter equality. SAcx.But nowthe question whatpath the sawshould is follow inmaking thecut. Sr_. It seems tomethatthisoughtnotto bea difficult task: forif by sawing theprismalongthediagonalineandremoving l halfofthematerial,the remaindercquires property the a a just oppositeothatoftheentireprism, othat at every t s pointwhere the latter gainsstrengththe formerbecomes eaker,then it w seems methat bytakinga middle to path,i.e.,byremovingalf h theformerhall orone-quarter fthewhole,hestrength o t ofthe remaining willbeconstant t allthose figure a pointswhere,nthe i twopreviousigures, f thegainin onewasequalto thelossinthe other. SALV. Youhavemissed mark,Simplicio. or,asI shall the F presentlyhow s you,theamount hich w youcanremoverom f the prismwithoutweakeningt is not a quarterbut a third. It i nowremains, s suggested Sagredo, discoverhe path a by to t along which thesawmusttravel:this,asI shall rove,mustbea p parabola.But it is firstnecessaryodemonstrate fol!owing t the lemma: If thefulcrums soplacedunder are twolevers orbalances thatthearmsthrough which theforcescCt a areto eachother inthe sameratioas thesquares fthe armsthrough o which theresistances andif theseresistances eachother acCt, areto in thesameratioasthearmsthrough whichtheyact,then the forces illbe equal. w Let AB and CD representtwo leverswhoselengthsare A,. , . , _l_divided their fulcrumsin by I_ _ -such a wayasto makethe dis_ tanceEB bearto the distance D FDa ratiowhich isequalto the Fig. 4 3 square theratiobetween of the distances EAand FC. Letthe resistances locatedat Aand C be to eachotherasEAisto FC. Then,I say, theforces which mustbe applied B andD in orderto holdin equilibrium at the resistances

[ 8o]

SECONDDAY I43 resistancest A and C areequal. Let EG bea meanpropora tionalbetweenEB and FD. Thenwe shall have BE-.EG= EG:FD=AE:CF.But this last ratiois precisely which that wehaveassumed existbetween resistances A and C. to the at AndsinceEG:FD=AE: F,it follows, C permutando, EG: that AE=FD:CF.Seeing thatthedistances CandGAaredivided D in thesameratiobythepointsF andE, it follows thesame that forcewhich, henapplied D, willequilibratehe resistance w at t at C,would ifapplied atGequilibrate tAa resistance a equalto that found atC. But onedatumoftheproblem isthatthe resistance Aisto at theresistancetCasthedistance Eisto thedistance F,oras a A C BEisto EG. Thereforeheforceapplied G, orratherat D, t at will,whenapplied atB,just balance theresis_nce located atA.
Q.E.D.

Thisbeingcleardrawtheparabola NBin theface F FBofthe prismDB. Let the prismbesawedalong parabola hose this w vertexis at B. Theportionofthe solidwhichremains illbe w included between baseAD,the recCtangular the planeAG,the straightlineBG and the surfaceDGBF,whosecurvatureis identical iththat oftheparabola w FNB. Thissolid willhave, I say,the samestrengthat everypoint. Let the solidbe cut by a planeCOparallelto D.......................... . :. the plane AD. Imaginef ., ............. 4.. ..io (_ I :...... J..I the pointsA and C to be _ _o.. ofwhich of twolevers [...fl. onewillhavethe the fulcrums .b[_ arms BAandAF;theother_ c BCandCN. Thensince in Fig.35 theparabola FBA, wehaveBA:BC=AF_: it isclearthat CN 2, thearmBAofoneleveristothearmBCoftheotherleverasthe squareof the armAF is to the squareof the otherarmCN. Sincethe resistance be balanced the leverBAis to the to by resistance be balanced to bythe leverBCin the sameratioas the recCtangle isto the recCtangle DA OC,that is as the length AFistothelengthCN,which twolengths retheotherarmsof a the levers,it 'follows, y the lemmajust demonstrated, b that the
....... ........... _ ........ _ ....

I44 THE TWO NEW SCIENCESOF GALILEO the sameforcewhich,whenapplied BGwillequilibratehe at t resistance t DA,willalsobalance resistance t CO. The a the a sameis true for any other secCfion. Therefore parabolic this solid isequally strongthroughout. It cannowbe shownthat, if the prismbe sawed alongthe lineoftheparabola FN-B, ne-third o partof it willbe removed; becausehe rectangle and the surface t FB FNBAbounded by theparabolaare the basesof twosolidsincluded between two parallelplanes,i. e.,between rectangles B andDG; conthe F sequently volumes f thesetwosolidsbearto eachother the o the sameratioas theirbases. But the areaof the red-tangle is oneand a halftimesas largeas the areaFNBAunderthe parabola; enceby cuttingtheprismalong parabola ereh the w moveone-third the volume. It is thus seenhowone can of diminish weightofa beamby asmuchas thirty-three the per centwithoutdiminishing strength; fact ofnosmallutility its a inthe construcCtion oflargevessels, especially supporting and in the decks,sincein suchstructureslightness of primeimis portance. SAcra Theadvantages derived fromthisfactaresonumerous that it wouldbe both wearisomend impossible mention a to themall;but leaving matterto one side,I should this liketo learnjust howit happensthat diminution ofweights possible i inthe ratioabove stated. I canreadily understand that,when a sectionis madealongthe diagonal, ne-half he weightis o t removed; as for the parabolic but, sectionremoving ne-third o ofthe prism, his I canonlyacceptonthewordofSalviati ho t w isalways reliable; owever prefer h I first-hand nowledge the k to wordofanother. SALe. ouwouldlikethena demonstration the factthat Y of the excess the volumeof a prismoverthe volumeofwhat of we have called.theparabolicsolidis one-third the entire of prism. This I havealready givenyouon a previous occasion; however shallnowtry to recallthe demonstration which I in I remember aving h useda certain lemmarom f Archimedes' book On Spirals,* namely,Givenany numberof lines,differing in
* For demonstrationof the theoremhere cited, see "Works of Arch-

[iSi]"

SECONDDAY 145 lengthonefromanother ya commonifference b d which asequal to the shortestoftheselines;andgivenalsoanequal umber n of lineseachof whichhasthe samelengthas thelongest f o the first-mentioned thenthe sumofthe squares the series; of linesofthis second group willbelessthanthreetimes thesum ofthesquares ofthelines inthefirstgroup.Butthesumofthe squares ofthesecond group willbegreater thanthreetimes the sumofthesquaresfallexcepting o thelongest ofthefirstgroup. Assuming this,inscribenthe re&angle i ACBPtheparabola AB. Wehavenow provethatthemixed to triangle BAPwhose sidesareBP andPA,andwhose baseistheparabola BA,is a thirdpartoftheentirere(tangle P. Ifthisisnottrueit will C beeither reater rlessthana third. Suppose beless g o itto byan area which isrepresented byX. Bydrawing parallel the lines to sidesBPandCA,we candivide re&angle intoequal the CP parts;andif the process econtinued eshallfinally b w reach a division intopartsso smallthat eachof themwillbesmaller than the areaX; let the rec-s v tangleOBrepresent neofthese o T o parts and, throughthe points _ s wherethe otherparallels utthe c "_'_ R parabola, rawlinesparallelto .... d r .'2,, _I, AP. Let usnowdescribe about ... E\ our "mixedtriangle" a figure _\ A K madeup of re&angles uchas s D, BO,IN, HM,FL,EK,andGA; ] x ] this figure willalsobe lessthan Fig. 6 3 a thirdpartofthe re&angle Pbecauseheexcess fthisfigure C t o above theareaofthe "mixed triangle" ismuchsmallerhanthe t re&angle Owhich B wehavealready madesmallerhanX. t SAGtt. Moreslowly, please; [or donotseehowtheexcess f _ I o thisfigure describedboutthe"mixed a triangle"ismuchsmaller thanthere&angle O. B SALV. Doesnotthere&angle Ohavean areawhichisequal B to thesumoftheareasofallthelittlere&angles through which imede._" translated byT.L.Heath (Camb. Press897p. Io7and Univ. I ) p. I6_, [Trans.]
[I82]

i46

THE

TWO

NEW

SCIF_CES

OF GALII.FD

the parabolapasses?I meanthe re&angles I, IH, I-IF,FE, B EG, and GAofwhichonlya partliesoutsidethe "mixedtriangle." Havewenot takenthe re&angle Osmallerhanthe B t areaX? Therefore if,as ouropponent ightsay,the triangle m plusX is equalto a third part of this re&angle the cirCP, cumscribedigure, tfichaddsto the triangle arealessthan f w an X, willstillremainsmaller thana thirdpart of the re&angle, CP. But this cannotbe, because circumscribed this figureis largerthan a thirdof the area. Hencek is not truethat our "mixedtriangle"islessthana thirdofthere&angle. to be shown thecircumscribed islargerthana third that figure partofthe re&angle P,a taskwhich C willnot, I believe, prove soeasy. S_v. Thereis nothingverydifficult boutit. Sincein the a parabolaDE'_'=DA:AZ= re&angleKE: re&angleAG, seeing thatthealtitudes fthesetwore&angles, andKL, are o _A_K equal, it followsthat E--D_:__2:_*=re&angle KE: re&angle Z. In preciselyhe samemannerit maybe shown K t that theotherre&angles LF,MH,NI,0]3,standto oneanother in the sameratioasthe squares thelinesMA,NA,OA,PA. of Let us nowconsiderhe circumscribed t figure,composedf o areaswhichbearto eachotherthesame ratioasthesquares fa o series oflineswhose common ifference lengthisequalto the d in shortestone in the series;notealsothat the re&angle is CP madeup of anequalnumber fareaseachequalto the largest o andeachequalto there&angle B. Consequently, O according to thelemma ofArchimedes, thecircumscribed islargerthan figure a thirdpartofthere&angle P;butit wasalsosmaller, C which is impossible.Hencethe "mixedtriangle"is not lessthan a thirdpartofthere&angle P. C Likewise, say,it cannot egreater. For,letus supposehat I b t itisgreater thana thirdpartofthere&angle Pandletthearea C X representhe excess fthe triangle t o overthethirdpartof the re&angle P; subdividehe rectanglentoequalre&anglesnd C t i a continue processuntiloneof thesesubdivisions smaller the is than

SAGR. haveleared You c upmydifficulty; it stillremains but

[ s3]

SECONDDAY x47 thantheareaX. LetBOrepresent sucha rec_ngle smallerhan t X. Using above the figure, ehavein the"mixedtriangle"an w inscribed figure, adeup of the red-tangles m VO,TiM, M,RL, S and QK,whichwillnot be lessthana thirdpart ofthe large rectangle P. C For the "m_ed triangle"exceeds inscribed the figure a by quantitylessthan that bywhichit exceeds thirdpart of the the recCtangle to seethat this is truewe haveonlyto reCP; memberhat theexcess fthetriangle t o overthethirdpartofthe rectangle is equalto the areaX, whichis lessthan the CP recetangle BO,whichin turn ismuchlessthanthe excess fthe o triangleover the inscribed figure. For the rectangle is BO [_84] made upof small the redt_mgles AG, GE,EF, HI, IB; FH, and and the excess f the triangleoverthe inscribed o figureis less than half the sumof theselittle reCtangles. hussincethe T triangle exceeds thethirdpartoftherectangie Pbyanamount C X, whichis moretbnn that by whichit exceeds inscribed the figure,helatterwillalsoexceed t thethirdpartofthe rectangle, CP. But, bythe lemma whichwehaveassumed, issmaller. it For the rectangle CP,beingthe sumofthe largestrectangles, bearsto the component rectangles f the inscribed o figure the sameratiowhichthe sumofallthe squaresofthe linesequal to the longestbearsto the squares the lineswhichhavea of common difference, after the squareof the longesthas been subtracCted. Therefore, s in the caseof squares,the sumtotal of the a largest eetangles,e.,the rectangle t i. CP,is greaterthanthree timesthe sumtotalof those having common ifference a d minus thelargest; ut theselastmakeupthe inscribed b figure.Hence the "mixedtriangle"is neither greaternor lessthanthe third partofrectangle P;itisthereforequalo it. C e t SACR. Afine,clever emonstration; all the moreso bed and cause gives it usthequadrature ftheparabola, o proving be itto four-thirds fthe inscribed triangle, faCt hichArchimedes o * a w demonstrates bymeansoftwodifferent, butadmirable, of series
Distinguishcarefully betweenthis triangleand the "mixed triangle" abovementioned.[Trans.]

x48 THE TWO NEW-SCIENCESOF GALII.F.O manypropositions. his sametheoremhas alsobeenrecently T established LucaValerio,* Archimedes f our age;his by the o demonstration istobefoundinhisbookdealing withthecenters ofgravityofsolids. S_v. Abookwhich, ndeed,snotto beplacedsecondo any i i t produced the mosteminent eometers itherof thepresent by g e or ofthe past;a bookwhich,assoonasit fellintothehandsof ourAcademician, ledhimto abandon hisownresearches along theselines;forhe sawhowhappilyeverything adbeentreated h anddemonstratedyValerio. b SAoR. WhenI wasinformedfthiseventbytheAcademician o himself, beggedofhimto show demonstrations I the whichhe had discovered eforeseeingValerio's ook;but in this I did b b notsucceed. SAT.V. I havea copyofthemandwillshow themto you;for youwillenjoythe diversity ofmethodemployed thesetwo by authorsin reaching proving sameconclusions; and the youwill alsofindthat some oftheseconclusions areexplained indifferent ways,although bothareinfactequally correct. SAoR. shallbe muchpleasedto seethemandwillconsider I it a greatfavorif youwillbringthemto our regularmeeting. But in themeantime, considering strength solidformed the ofa froma prismby meansofa parabolic ection, ouldit not, in s w viewofthe fa_ that this resultpromiseso bebothinteresting t andusefulinmanymechanical operations, finethingif you bea wereto givesome quickandeasyruleby whicha mechanician mightdrawa parabola pona planesurface? u SALV. Thereare manywaysof tracingthesecurves;I will o mentionmerelythe twowhichare the quickest f all. Oneof o these is really remarkable; ecauseby it I can trace thirty b or forty parabolic curveswithno lessneatness precision, and and in a shortertime than anothermancan,by the aid of a compass, eatlydrawfouror sixcircles different izes n of s upon paper. I take a perfedtIy roundbrassball aboutthe sizeof a walnutandprojectit alongthesurface metallic ofa mirrorheld
*An eminent Italian mathematician, contemporarywith Galileo.

[i8s1

[Trans.]

SECONDDAY 149 in a nearlyuprightposition, othat the ballin itsmotionwill s pressslightly uponthe mirrorandtraceout a finesharpparabolicline;this parabola willgrowlongerand narrower s the a angleof elevation increases.Theaboveexperiment furnishes clearand tangibleevidencehat the pathof a projectile a t is parabola;a factfirstobserved byourfriend demonstrated and byhiminhisbook onmotionwhich weshalltakeupat ournext meeting. In the execution this method,it is advisable of to slightly heatandmoisten theballbyrolling inthehandinorder thatitstraceuponthemirrormaybemore distincct. [I86] Theothermethod ofdrawing thedesired curveupontheface of the prismis the following: Drivetwonailsintoa wallat a convenient heightand at the samelevel;makethe distance between thesenailstwicethewidthoftherecCtangle uponwhich it is desiredto tracethe semiparabola. verthesetwonails O hanga lightchainof sucha lengththat the depthof its sag is equalto thelengthofthe prism. Thischainwillassume the formof a parabola,* that if this formbe markedbypoints so onthe wallweshallhavedescribed complete arabola hich a p w canbe dividedintotwoequalpartsby drawing verticalline a througha pointmidway between twonails. The transfer the ofthis curveto the twoopposingacesoftheprismisa matter f ofnodifficulty; anyordinary mechanic illknowhowto doit. w By use of the geometrical linesdrawnuponour friend's compass,t nemay easilylayoffthosepointswhichwilllocate o this same curveuponthe same oftheprism. face Hithertowe have demonstratedumerous onclusions n c pertainingto the resistance whichsolidsofferto fracture. As a startingpointforthisscience, weassumed theresistance that offered the solidto a stralght-away by pullwasknown;from this baseone mightproceed the discovery manyother to of resultsandtheirdemonstrations; oftheseresults thenumberto
* It is nowwellknownthat this curveis not a parabolabut a eatenary the equationof whichwas firstgiven,49 years after Galileo's death, by James Bernoulli. [Trans.] t The geometrical nd militarycompass Galileo,describedin Nat. a of Ed.Vol.2. [Trans.l

i5o THE TWO NEW SCIENCES GALtT._O OF be foundin natureisinfinite. But, in orderto bringour daily conferenceo an end,I wishto discusshe strengthofhollow t t solids, hichareemployed art--and stilloftenerinnature-w in in a thousand operations thepurposeof greatlyincreasing for strengthwithoutaddingto weight;examples ftheseare seen o in thebonesofbirdsandin manykindsofreeds whicharelight and highlyresistantboth to bending braking. For if a and stemof strawwhichcarriesa headof wheatheavierthan the entirestalkweremadeup of the samemount ofmaterialin solid formitwould offer lessresistance bending to andbreaking. Thisis anexperience hichhas beenverified w andconfirmedn i practice whereit isfoundthat a hollow lance a tubeofwood or ormetalismuchstrongerhanwould solidoneofthesame t bea lengthandweight, newhich o would necessarily t_nner;men be havediscovered, therefore,hat in orderto makelances t strong as wellas lighttheymustmakethemhollow.We shallnow show that: In the caseof two cylinders, hollowthe other solid one but havingequalvolumes nd equallengths, heirresista t ances[bending strengths] reto eachotherin the ratioof a theirdiameters. Let AE denotea hollow cylinderand IN a solidoneof the A same weight and length; (_ .-_' _-- _____2 --_ then, I say, that the resistance against fracCturexe hibitedbythetubeAEbears B ! to that of the solidcylinder -_--_-_. ' ,-_, IN the sameratioas thediameterAB to the diameter IL. lh_s _s veryewdent;/or Fig. 7 3 sincethe tubeandthe solid cylinderN havethesamevolume ndlength,heareaofthecirI a t cularbaseILwillbeequaltothat oftheannulus Bwhich A isthe baseofthetubeAE. (Byannulus heremeantthe areawhich is liesbetweentwo concentric circles different adii.) Hence of r theirresistances straight-away toa pullareequal;butinproducing
I _L _ -,r n--i-iL_n e q

[I87]

SECONDDAY I5I ingfra_ure bya transverse ullweemploy,n thecaseof the p i cylinderN, thelength asoneleverarm, he pointL asa I LN t fulcrum, andthediameter I,orits half,asthe opposing L lever arm:whilein the caseofthe tube,the length which BE plays thepartofthefirstlever isequalo LN,theopposing arm t lever armbeyond fulcrum, is thediameter B,or its half. the B, A Manifestly thenthe resistancebending [ strength] f the tube o exceeds thatofthesolidcylinderntheproportion i inwhich the diameter ABexceeds thediameter which IL" isthedesired result. Thus the strengthof a hollowtube exceedsthat of a solid cylindern the ratioof theirdiameters i whenever thetwo are made ofthesame material havethesame eight ndlength. and w a It maybewellnextto investigatehegeneral t caseoftubes andsolidcylinders ofconstantength, utwiththeweight nd l b a thehollow portionariable. irstweshall how v F s that: Givena hollowtube,a solidcylinder aybedetermined m which beequal eguale] it. will [ to Themethod isverysimple. LetABdenote theexternal nd a CD the internal diameter fthe tube. In the larger o circle lay off thelineAE equalin lengthto the di/L ameterCD; join the points E and B. f Nowsincethe semicircle, angleat E inscribed a //_/ AEB,i8a right-angle, area/ the in ofthecircle whose diameter isABisequal[j to the sumof the areasofthe twocircles|/ whoserespecCtive diameters reAE and_" a EB. ButAEi8thediameter ofthehollow\ portionof the tube. Thereforehe area t ofthe circle whose diameteris EB is the sameas the area of the annulus ACBD. Fig. 38 Hencea solidcylinder f circular asehaving diameter B o b a E willhavethe samevolume the wallsof thetube of equal as length. Byuse ofthistheorem, iseasy: it To findtheratiobetween resistance the [bending strength] of anytubeandthatofanycylinder equalength. of l Let

[ 881

I52 THE TWO NEW SCIENCESOF GALII.EO LetABEdenotea tubeandRSMa cylinder fequallength:it o is requiredto findthe ratiobetweentheirresistances.Using theprecedingroposition, p determine cylinderLNwhich a I shall have the same volumeand "_% - _ ]engthas the tube. Drawa ) :B lineV of sucha lengththat it willbe relatedto IL and RS (diameters f the bases o 1_i ofthecylindersNandR_M), I (_ ' _._ as ,ollows:V:RS--RS:IL. / Then,I say,theresistance f yrT kl/ o v, thetubeAE isto that ofthe Ik cylinderRM as the length (D )_: ofthelineAB thelength isto [189] V. For,sincethe tubeAE is Fig.39 equal both in volumeand length, o the cylinderN, theresistancefthe tubewillbearto t I o theresistancefthecylinderhesameratioasthelineABto IL; o t but the resistance f the cylinder N is to that ofthe cylinder o I RMasthe cubeofIL isto thecubeofRS,that is,asthe length IL is to lengthV:therefore, x cequali, resistancebending e the [ strength] fthe tubeAE bearsto the resistance fthe cylinder o o RM thesameratioasthe length ABto V. q. _. D.

END

OF SECOND DAY.

THIRD DAY
[ 9o] CHANGEOF POSITION.[De Motu Local ]
.................................................to set fortha very new science Y purposeis is, in nature, perhaps nothing older than motion,concerning whichthe bookswritten by philosophersare neither {ewnor small; neverthelessI have discoveredby experiment somepropertiesof it whichare worth _ dealingwith a very ancientsubjecCt.There knowingand whichhave not hitherto been either observedor temonstrated. Somesuperficial bservations o have beenmade,as, forinstance,that thefreemotion[naturalem motum]of a heavy fallingbody is continuouslyaccelerated; * but to just what extentthis acceleration occurshasnot yet been announced;for so far as I know,noonehas yet pointedout that the distancestraversed, during equal intervals of time, by a body fallingfrom rest, stand to one anotherin the sameratio as the oddnumbersbeginningwithunity/f It has been observedthat missilesand projecCtiles describe a curvedpath of somesort;howevernoone haspointedout the fagt that this path is a parabola. But this and otherfacets, ot n few in number or less worth knowing,I have succeededin proving; and what I considermoreimportant,there have been openedup to this vast and most excellentscience,of whichmy "Natural otion"oftheauthor m hashere beentranslated into"free motion"--since thisisthetermused to-day todistinguish the"natural" from the"violent"motionsftheRenaissance. o [Trans.] t Atheorem demonstrated 175 onp. below.[Trans.]

I54 THE TWO NEW SCIENCESOF GALII.F.O workis merelythebeginning, waysand meansbywhichother mindsmoreacutethar_minewillexploreits remotecomers. This discussion dividedinto three parts; the first part is deals withmotiorr hichissteadyoruniform; hesecond w t treats ofmotionaswefindit accelerated nature;thethirddeals in with theso-callediolent otions ndwithprojecCtiles. v m a [I911 UNIFORMMOTION In dealing with steadyor uniformmotion, e needa single w definition hichI giveasfollows: w
DEFINITION

By steadyor uniform motion,I meanonein whichthe distances traversedby the movingparticleduring any equal intervals time_ of arethemselvesqual. e

CAtmo_
Wemustadd to the olddefinition (which defined steadymotion simplyas one in whichequaldistances traversedin are equaltimes)theword"any," meaning this,allequalinterby vals of time; for it may happenthat the movingbodywill traverseequal distances duringsomeequalintervalsof time and yet the distances traversedduringsomesmallportionof thesetime-intervals aynot be equal,eventhoughthe time_ m intel_rals beequal. Fromtheabove definition, fouraxiomsollow, f namely:

Axio_I
In the caseofone andthe sameuniform motion,he distance t traversedduringa longerintervalof time is greaterthan the distance traversed during shorter a interval ftime. o Axao_tII In the caseof one and the sameu_formmotion,the time required traverse greaterdistance longer to a is thanthe time required lessdistance. fora

THIRD DAY I55 _oM III In oneand thesameinte_alof time,he distance traver6ed at a greater speedis largerthan the distancetraversed a at lessspeed.
A_OM IV

[i921

The speedrequired traversea longerdistanceis greater to than that required traversea shorterdis_nceduringthe to same time-interval.
THEOREM PROPOSITION I, ! If a moving particle, arried c uniformlyt a constant peed, a s traverses distanceshe tim_e-intervals two t requiredareto eachotherin the ratioofthesedistances. Let a particlemoveuniformly with constantspeedthrough twodistances B,BC,andletthetimerequired traverse B A to A berepresentedyDE;thetimerequired traverse C,byEF; b to B
. : i_.,, , :

_-D T;E TF . : I l 'LTJlk "rB ,t

TC . ..... i" ' ..... IX

Fig.4o thenI saythat the distance ABis to the distanceBC as the timeDEisto thetimeEF. Letthedistancesndtimesbeextended nbothsides a o towards G,H and I, K; letAGbe divided intoanynumber hatever f w o spaceseach equalto AB,and in like mannerlay off in DI exactlythe samenumber time-intervals d eachequalto DE. Againlay off in CH any numberwhatever distanc_each of equalto BC; and in FK exa&lythe samenumberof timeintervalseachequalto EF; thenwillthe distance BGand the timeEl he equaland arbitrarymultiples the distance of BA and the timeED; and likewisehe distance and the time t HB KE are equalandarbitrarymultiples fthe distance and o CB the timeFE. Andsince isthe timerequiredo traverse DE t AB,thewhole
time

I56 THE TWO NEW SCW.NCES GALILEO OF timeEI willbe required the w_oledistance for BG,andwhen themotionisuniform therewillbein E1asro_ny time-intervals eachequaltoDE astherearedistancesnBGeachequalto BA; i andlikewiset followshat KE representshe timerequired i t t to traverseHB. Since,however,he motionis uniform,t follows if the t i that distanceGBis equalto the distanceBH, thenmustalsothe timeIE beequalto thetimeEK; andifGBisgreaterthanBH, thenalsoIE willbegreaterthanEK; and if less,less.* There arethenfourquantities, he firstAB,the second the third t BC, DE, and the fourthEF; the timeIE and the distance are GB arbitrarymultiplesof the first and the third, namelyof the distance ABandthetimeDE. But it has beenprovedthat bothof theselatter quantities areeitherequalto, greaterthan,or lessthanthe timeEK and the spaceBH,whicharearbitrarymultiples f the second o and the fourth. Therefore first is to the second, amelythe the n distance Bisto thedistance A BC,asthe thirdis to thefourth, namelythetimeDE isto the timeEF. q.w.D.

[ 93]

Tao= II,PRoPosmo II
tervalsof time,thesedistances illbeartoeachotherthe w sameratioas the speeds. Andconversely thedistances if areas the speeds thenthe timesareequal. ReferringoFig.4o,letABandBCrepresent t thetwodistances traversed equaltime-intervals, in thedistance forinstance AB with thevelocityDE, and the distanceBC with the velocity EF. Then,I say, thedistance ABisto the distance BCas the velocity is to the velocity DE EF. Forif equalmultiples f o bothdistances nd speeds taken,asabove,namely, B and a be G IE ofAB andDE respe_ively, nd in likemannerIIB andKE a of BC and EF, then one may infer, in the samemanneras above,that themultiples B andIE areeitherlessthan,equal G
* The methodhereemployedby Galileois that of Euclid as set forth in the famous5th Definition the Fifth Bookof hisElements,orwhich of f seeart.Geometry ncy. Brit. Ixth Ed. p. 683. [Trans.] E

If a movingparticletraverseswodistancesin equal int

THIRD DAY 157 to, or greaterhanequalmultiples f BHandEK. Hencethe t o theorem isestablished.

nI,P oPosmoN nI
In the caseofunequalspeeds,the time-intervals required to traversea givenspaceareto eachotherinversely as thespeeds. Let the largerofthe twounequal peeds s beindicated byA; the smaller,by B; and let the motioncorresponding both to traverse givenspaceCD. ThenI saythetimerequiredo the t traverse thedistance Dat speedAk: C A isto thetimerequired travto ersethe samedistanceat speed_ j _: _ .: B, asthe speedB isto the speed _ D' A. For let CD be to CE as A is to B; then, fromthe preced-]_t t ing, it follows the timerethat Fig. 1 4 qi_ired tocomplete distance at speedAis the sameas the CD [I94] the time necessary complete at speedB; but the time to CE needed traverse to thedistance CEat speedB isto the timerequiredto traversethe distance at the samespeedas CE CD isto CD; therefore timein whichCD is covered speed the at A isto the timeinwhichCDis covered speedBas CEisto at CD,that is,asspeed isto speed B A. Q.E.D. with a different peed,thedistances overed s c bythemduringunequalintervals oftimebearto eachotherthe compoundratioof the speeds nd timeintervals. a Let the twoparticles hicharecarried w withuniform motion beE andFandlettheratioofthespeed ofthebodyE beto that ofthebodyF asAisto B;butlettheratioofthetimeconsumed bythe motionofE be to thetimeconsumedythe motionof b F asCistoD. Then,I say,thatthedistance overed yE, with c b speed intimeC,bearsto the space A traversed with"peed byF s B

TE_OR_M IV,Pgo_osrrioN IV Iftwoparticles arecarried uniform otion, with m buteach

I58 THE TWONEW SCIENCES F GALII,_O O Binthn.e a ratiowhichstheproduct D i oftheratioofthespeed A tothespeed BbytheratioofthetimeCtothetimeD. For ifG isthedistanceraversedyE at speed during t b A thetime1_. A----.----_ G_ L ;interval Aistothesp_dB; C,andifG istoI as C_ "thesp.e! B: : I: .... -_ andif alsothetime-interval C is is to L,time-interval D _.;..... : x.:: : asI to the thenit follows Fig.4z that I is the distance traversed F in thesame by timethatG istraversedyE since b G istoI in the same ratioasthespeed to thespeed And A B. since isto L in thesameatioasthetime-intervals I r CandD, if I isthe distanceraversedyF during t b theinterval C,then Lwill ethedistance b traversedyFduringheinterval at the b t D speed . B Butthe ratioofG to L isthe product fthe ratios to I o G andI to L, thatis,oftheratios fthespeed to thespeed o A B andofthetime-interval Ctothetime-interval D. Q.g, D. T_Eo_a V, PgoPosmo_ V Iftwoparticles aremoved a uniform at rate,butwithunequalspuds,through unequal distances, thentheratioof thetime-intervals occupied betheprodu_oftheratio will ofthedistances bytheinverseatioofthespeeds. r Letthetwomovingarticles p bedenotedyA andB, andlet b the speedof A bed]_w. : _ . _, tothespeed ofBin :._. theratioofVtoT; *".:_ ' in likemannerlet_ _: _: c. ....... the distancesray- T: t ersed beintheratio Fig. 3 4 ofS toR; thenI saythatthe ratio ofthetime-interval during which themotion fA occurs tohetime-interval o _t occupied by themotion fBistheprodu_oftheratioofthespeed to the o T speed Vbytheratio ofth_di_st_.nc Stothedistanc R. LetCbethetime-interval occupied bythemotion fA,and o let

[,9s]

! '_ _

THIRD DAY 159 letthe time-interval bearto a time-interval the sameratio C E asthespeed tothespeedV. T AndsinceC is the time-interval during whichA, withspeed V,traverses thedistance Sandsince the speed T, ofB,isto the speedV, as the time-interval is to thetime-interval then C E, E willbe the timerequired the particleB to traversethe by distanceS. If nowwe let the time-interval beto the timeE interval as thedistance to thedistance thenit follows G Sis R, that G isthetimerequired byB totraverse space Since the R. theratioofC to G isthe productoftheratiosC to E andE to G (while lsothe ratioofCto E istheinverse a ratioofthespeeds ofAandB respeCtively, i.e.,theratioofT toV);and since the ratioofE to G is the sameas that of the distances and R S respectively, thepropositionsproved. i [x961 T_sotu_M PRox-osrrlO_ Vt, VI If twoparticles carriedat a uniformrate,the ratioof are theirspeeds illbetheproduct f theratioofthe distances w o traversed bytheinverse ratioofthetime-intervals occupied. Let A and B be the twoparticles hichmoveat a uniform w rate; and let the respective istances d traversed themhaye by the ratio of V.._C , to T, but letthe time-intervals be 1_ asStoR. Then _ I say the speed B G of A will bear t_" "_ to the speedof Fig. 4 4 B a ratiowhichisthe product ftheratioof thedistance to o V thedistance andthetime-interval thetime-interval T Rto S. LetCbethespeedat which traverseshedistance during A t V the time-interval S;and letthe speedC bearthe-same ratioto anotherspeedE asV bearsto T; thenE willbe the speedat whichB traversesthe distanceT duringthe time-interval S. If nowthe speed isto another peedGasthe time-interval E s R is to thetime-interval S,thenG willbethe speedat whichthe particle

16o THE TWO NEW SCIENCESOF GALIT,EO particleB traversesthe distance duringthe time-interval T R. Thuswehave the speedC at whichthe particle coversthe A distance duringthetimeSand alsothespeedG at whichthe V particleB traversesthe distance duringthe time R. The T ratioofC to G is the producc ofthe ratioC to E and E to G; theratioofC to E isby definitionhe sameasthe ratioof the t distance to distance andthe ratioofE to G isthe sameas V T; theratioofR to S. Hencefollowsheproposition. t SAT.V. Thepreceding iswhatourAuthorhaswrittenconcerninguniform motion. Wepassnowto a newandmorediscriminatingconsiderationf naturallyaccelerated otion,suchas o m that generally experienced heavyfalling by bodies; ollowing f is the titleand introdu_ion. [197] NATURA ACCEI,ERATt_,D T,I,Y MOTION The propertiesbelongingo uniform t motionhavebeendiscussed the precedingection; ut accelerated otionremains in s b m to beconsidered. And firstofallit seems desirable findand explain definito a t_n bestfittingnaturalphenomena.Foranyone mayinventan arbitrarytype of motionand discussits properties;hus, for t instance, ome s haveimagined elices h andconchoids asdescribed bycertainmotions hicharenotmetwithin nature,and have w verycommendably established theproperties hichthesecurves w possessnvirtueoftheirdefinitions; wehavedecided coni but to siderthe phenomenaf bodies o falling Mth an accelerationuch s as ad'tually occursin nature and to make this definition of accelerated otionexhibitthe essentialfeaturesof observed m accelerated otions.Andthis,at last,afterrepeated m efforts we trustwehavesucceeded indoing. Inthisbelief eareconfirmed w mainlyby the consideration experimental that resultsare seen to agreewith and exatlycorrespond with thoseproperties which have been, one after another, demonstrated us. by Finally,in the investigationf naturallyaccelerated otionwe o m wereled,byhandasitwere,infollowing thehabitandcustom of nature

i :

THIRD DAY I6i natureherself,in all her variousother processes, employ to onlythosemeans whicharemostcommon, simple andeasy. For I thinknoone believeshat swimmingr flying t o canbe accomplished a mannersimpleror easierthanthat instincin tivelyemployedyfishes b andbirds. When,therefore,I observea stoneinitiallyat rest falling froman elevatedpositionand continually cquiring a newincrements fspeed, hyshould notbelieve suchincreases o w I that takeplacein a mannerwhichisexceedingly simple rather and obviouso everybody?If nowweexamine t themattercarefully wefindnoaddition increment or moresimple thanthat which repeatsitselfalwaysin the samemanner. This we readily understand whenweconsiderhe intimate t relationship between timeandmotion; orjust asuniformity fmotionisdefined f o by and conceivedhroughequaltimesand equalspaces(thuswe t calla motion uniform whenequaldistances retraversed a during equal time-intervals), also we may, in a similarmanner, so throughequal time-intervals, conceive dditions speedas a of takingplacewithoutcomplication; thuswemaypicmreo our t [I981 minda motion asuniformlyndcontinuously a accelerated when, duringany equalintervalsoftimewhatever, qualincrements e of speedare givento it. Thusif any equalintervals time of whateverhave elapsed, ounting c fromthe timeat whichthe moving bodyleftits position ofrest andbeganto descend, the amountof speedacquiredduringthe first two time-intervals will be doublethat acquiredduringthe first time-interval alone; o theamountaddedduring s threeofthesetime-intervals willbetreble;andthat infour,quadruplehatofthefirsttimet interval. To put the mattermoreclearly, f a bodywereto i continuets motionwiththe samespeed i whichit hadacquired duringthe firsttime-lnterval andwereto retainthis sameuniform speed,henitsmotion t would betwiceasslow asthatwhich it would haveif itsvelocity hadbeenacquired during twotimeintervals. Andthus,it seems, e shallnot be farwrongif we put the w increment f speedas proportional the increment f time; o to o hence

,6z THE TWO NEW SCW.NCES GALTT,FO OF hence thedefinitionfmotion o which weareaboutto discuss ay m bestatedasfollows: Amotion issaidtobeuniformlyccelerated, a whenstartingfromrest,it acquires, during equaltime-intervals, equalincrementsfspeed. o SAGR. AlthoughI can offerno rationalobjecCtion this or to indeed anyotherdefinition, to devised byany authorwhomsoever, sinceall definitions arbitrary,I may nevertheless are withoutoffensee allowedo doubtwhether b t sucha definitions a theabove, stablished e inanabstracCt manner, orresponds and c to describes that kind of accelerated otionwhichwe meet in m naturein thecaseoffreelyfalling bodies.AndsincetheAuthor apparently maintainsthat the motiondescribed his definiin tion is that of freelyfallingbodies,I wouldliketo clearmy mind of certaindifficultiesn order that I may later apply i myselfmore earnestlyto the propositions nd their demona strations. SALV. iswellthat youandSimplicioaisethesedifficulties. It r They are, I imagine,he samewhichoccurred me whenI t to firstsawthistreatise,andwhich wereremoved eitherbydiscussionwiththe Authorhimself, rby turningthematteroverin o myownmind. SAGI_.. WhenI thinkofa heavy bodyfalling fromrest,that is, startingwithzerospeedand gaining speedin proportiono the t ['99] timefromthe beginningf themotion; ucha motion o s aswould, forinstance,n eightbeatsofthe pulseacquire i eightdegrees f o speed;havingat the end of the fourthbeat acquired fourdegrees;at the endofthe second,wo;at the endofthefirst,one: t andsincetimeisdivisible ithoutlimit,it follows w fromallthese considerations if the earlierspeedofa bodyislessthanits that presentspeedin a constantratio,then thereis no degree of speedhowever mall(or,one maysay, no degree slowness s of however reat)withwhichwemaynotfindthis bodytravelling g afterstartingfrominfinite slowness, e.,fromrest. Sothat if i. that speedwhichit had at the endofthe fourthbeatwassuch that, if keptuniform,the bodywouldtraversetwomilesin an hour,and if keepingthe speedwhichit had at the end of the second

THIRD DAY I63 second beat,it would traverseonemileanhour,wemustinfer that, as the instantof startingis moreand morenearlyapproached, thebodymoves soslowly that,if itkeptonmoving at this rate,it would nottraversea milein anhour,or ina day,or in a yearor ina thousand years;indeed, would it nottraversea spanin an evengreatertime;a phenomenon whichbaffles the imagination, oursenses while show usthat a heavyfalling body suddenly acquires reatspeed. g SM_v. Thisisoneofthe difficulties whichI alsoat thebeginning,experienced, butwhichI shortlyafterwards removed; and theremoval aseffecCted veryexperiment w bythe whichcreates thedifficultyoryou. Yousaytheexperiment f appearso show t that immediately aftera heavybodystartsfrom itacquires rest a veryconsiderable speed:andI saythat thesameexperiment makesclearthefacet theinitial otions fa falling that m o body,no matterhowheavy,arevery slowand gentle. Placea heavy bodyupona yielding material, leaveit therewithoutany and pressure exceptthat owing its ownweight;it is clearthat if to oneliftsthisbodya cubitor twoand allowst to falluponthe i same material, will,withthisimpulse, it exerta newandgreater pressure thanthat caused byits mereweight;andthis effecc is broughtaboutbythe [weight ofthe]falling bodytogether ith w the velocityacquiredduringthe fall,an effecCt whichwillbe greaterand greateraccording the heightofthe fall,that is to according the velocity the fallingbodybecomes reater. as of g Fromthequalityandintensity ofthe blowwearethusenabled to accurately estimate thespeedofa falling body. But tellme, gentlemen, nottruethat if a block isit beallowedo fallupona t stakefroma heightoffourcubitsanddrivesit intotheearth, say, fourfinger-breadths, coming that froma heightof two cubitsitwilldrivethe stakea muchlessdistance, ndfromthe a height fonecubita stilllessdistance; ndfinally o a iftheblock be lifted onlyonefinger-breadth howmuchmore willit accomplish than if merelylaid on top of the stake withoutpercussion? Certainlyvery little. If it be liftedonlythe thickness a of leaf,the effectwillbealtogethermperceptible. i Andsincethe effecc

[zoo]

164 THE TWO NEW SCIENCESOF GALIT._'O effec_t the blowdependsuponthe velocityof this striking of body,cananyonedoubtthemotionisveryslowandthe speed morethansmallwheneverhe eiTecCt t [oftheblow] simperceptii ble? Seenowthepower oftruth;the sameexperiment whichat first glanceseemedto showone thing,whenmorecarefully examined, assures usofthecontrary. But withoutdepending uponthe aboveexperiment, whichis doubtless ery conclusive, seemsto methat it oughtnot to v it bedifficulto establish t sucha fa_ by reasoninglone. Imagine a a heavystoneheldin theairat rest;the supportsremoved i and thestone setfree;thensince isheavierhantheairit begins it t to fall,and not withuniform motion slowly thebeginning but at andwitha continuously accelerated motion.Nowsince velocity canbe increased diminished ithoutlimit,what reason and w is thereto believethat sucha moving bodystartingwithinfinite slowness, is,fromrest,immediatelycquires speedoften that a a degreesratherthan one of four,or of two,or of one,or of a half,orofa hundredth; r,indeed, fanyoftheinfinite o o number ofsmallvalues[ofspeed] Praylisten. I hardlythinkyouwill ? refuseto grantthat the gainof speedof the stonefalling from restfollowshe samesequencesthediminution t a andlossofthis samespeedwhen, y some b impelling thestoneisthrownto force, its formerelevation: ut evenif youdonotgrantthis,I donot b seehowyoucandoubtthat the ascendingtone,diminishing s in speed,mustbeforecoming rest passthrougheverypossible to degree ofslowness. SIMP. if the numberof degrees greaterand greater But of slownessslimitless,heywillnever allexhausted, i t be therefore suchan ascending heavybodywillneverreachrest,but will continue tomovewithoutlimitalwaysat a slower ate;but this r isnottheobserved fa_. SA_v. his wouldhappen,Simplicio,f the movingbody T i weretomaintainitsspeedforanylengthoftimeat eachdegree of velocity; ut it merelypasseseachpointwithoutdelaying b morethan an instant:and sinceeachtime-interval owever h [2oi] smallmay bedividedintoan infinite number f instants, hese o t will

THIRD DAY 165 willalways esu_cient[innumber]ocorrespond theinfinite b t to degrees fdiminished o velocity. That sucha heavyrising bodydoesnotremain foranylength oftimeat anygivendegree ofvelocity evidentromthe folis f lowing: because if,some time-interval having beenassigned, the bodymoves with the samespeedin the lastas in thefirstinstant ofthat time-interval, couldfromthis second it degree of elevation in like mannerraisedthroughan equalheight, be just as it wastransferred fromthe firstelevation thesecond, to and by the samereasoning ould w passfromthe second the to third and wouldfinallycontinuein uniform motionforever. Saam From theseconsiderations appears me that we it to may obtain a proper solutionof the problemdiscussed by philosophers, namely,what causesthe acceleration the in naturalmotionofheavybodies Since, s it seemsto me,the ? a force[_irt_t] impressed bytheagentprojecCting thebodyupwards diminishes continuously, thisforce, olongasit wasgreater s than the contraryforceof gravitation,mpelled i thebodyupwards; whenthe two are in equilibrium bodyceasesto rise and the passesthrough thestateofrestin which impressed the impetus [impeto] notdestroyed, onlyitsexcess vertheweight f is but o o thebodyhasbeenconsumed--the excess hich w caused thebody to rise. Thenasthediminutionftheoutsideimpetus o [impeto] continues, andgravitation gains theupperhand,the fallbegins, but slowly first on account f the opposing at o impetus[virt_z impressa],largeportion a ofwhich stillremainsn thebody;but i as this continues diminish alsocontinueso bemoreand to it t moreovercomeygravity, ence continuous b h the acceleration of motion. Sn_P.Theideaisclever, etmoresubtle y thansound;oreven f if the argument ereconclusive,wouldexplain w it onlythe case in whicha naturalmotionis preceded y a violent otion,in b m whichtherestillremainsaCtive portionofthe external orce a f [virth esterna]; butwherethereisnosuchremainingortionand p thebodystartsfroman antecedent stateofrest,the cogencyf o thewhole argumentails. f SACR. believe youaremistaken that this distincI that and tion

i66 THE TWO NEW SCIENCESOF GALILEO tionbetween cases which youmakeissuperfluous orrathernonexistent. But, tell me, cannota proje&ilereceivefromthe proje&or ithera largeor a smallforce[v/rt_]uchaswillthrow e s it to a heightof a hundredcubits,andeventwentyor fouror
one ?

S_w. Undoubtedly, yes. SAcg.So therefore impressedorce[virt_ this f impressa] may exceed resistance gravityso slightlyas to raiseitonly the of a finger-breadth; finallythe force[v/rt_] f the projector and o maybe just largeenoughto exacdy alancethe resistance b of gravityso that the body is not liftedat all but merelysustained. When oneholds stoneinhishanddoeshedoanything a but giveit a forceimpelling [v/rt_ impellente] upwards it equal tothepower [facoltd] ofgravitydrawing downwards Anddo it ? younot continuouslympresshis force[virtU] i t uponthe stone as longas youholdit in the hand? Doesit perhapsdiminish withthetimeduring which oneholdsthestone ? Andwhatdoesit matterwhether support hichprevents this w the stonefromfalling furnished one'shandor bya table is by or by a ropefromwhichit hangs? Certainlynothingat all. You must conclude,therefore,Simplicio,hat it makesno t difference whatever hetherhefallofthestoneispreceded ya w t b periodof rest whichis long,short,or instantaneous rovided p onlythe falldoesnot take placeso longas the stoneis a&ed uponby a force[v/rtfi] pposed its weightand sufficient o to to holdit at rest. SAT.v. presentdoesnot seemto be daepropertime to The investigatehe causeofthe accelerationf naturalmotioncont o cerning whichvarious opinions avebeenexpressed h byvarious philosophers, someexplaining by attra&ionto the center, k othersto repulsion betweenthe very smallpartsof the body, while stillothersattributeit to a certainstress inthesurrounding mediumwhichclosesin behindthe fallingbodyand drivesit fromoneof its positions another. Now,allthesefantasies, to and otherstoo,oughtto be examined; it isnotreallyworth but while. At presentit is the purposeof our Authormerelyto investigate

[zoz]

THIRD DAY 167 investigate nd to demonstrate a someof the properties f aco celerated motion(whateverhe cause this acceleration ay t of m be)--meaninghereby motion, uchthatthemomentum t a s ofits velocity[i momenti ellasuavelocitY] d goeson increasing after departure fromrest,in simple roportionality p tothetime,which is the sameas sayingthat in equaltime-intervals body the receives qualincrementsfvelocity; ndif wefindthepropere o a ties[ofaccelerated otion] m which willbedemonstrated laterare realized infreely falling andaccelerated bodies, wemayconclude that the assumed definition includes ucha motionof failing s bodiesandthat theirspeed[accelerazione] goesonincreasings a thetimeandtheduration ofthemotion. SAGR. far as I seeat present,the definition ighthave So m been put a little moreclearlyperhapswithoutchanging the fundamental idea,namely, niformlyccelerated u a motion issuch that its speedincreasesn proportiono thespacetraversed; i t so that, forexample, speedacquired a bodyin falling the by four cubitswouldbe doublethat acquired falling in twocubitsand thislatterspeed would bedoublehat acquiredn thefirstcubit. t i Because thereis nodoubtbut that a heavybodyfalling from the heightof six cubitshas, and strikeswith, a momentum [impeto] double it had atthe endofthreecubits,triplethat that which ithadattheendofone. SALV.isvery comforting me to havehad sucha comIt to panioninerror; ndmoreover a letmetellyouthatyourpropositionseemssohighlyprobablehat ourAuthor imself dmitted, t h a whenI advanced thisopinion him,thathehadforsome to time sharedthe samefallacy. But whatmostsurprised mewasto see two propositionso inherentlyprobable s that they commandedthe assentofeveryoneo whomtheywerepresented, t provenin a few simple wordsto be not onlyfalse,but impossible. SIM1,. am one of thosewhoacceptthe proposition, I and believe a falling that bodyacquiresorce f [vires] initsdescent,ts i velocityincreasingn proportion the space,and that the i to momentum [momento] ofthefalling bodyisdoubled hen from w it falls

[203]

I68 THE TWO NEW SCIENCESOF GALII.EO froma doubledheight;thesepropositions, appearsto me, it oughtto beconceded ithout esitation rcontroversy. w h o SALV. Andyettheyareasfalseandimpossible asthat motion should completed be instantaneously; hereis a veryclear and demonstration it. If the velocitiesrein proportion the of a to spacestraversed,or to be traversed, hen these spacesare t traversed equalintervals f time;if, therefore, velocity in o the withwhic_thefailing bodytraverses space a ofeightfeetwere double withwhichit coveredhefirstfourfeet(justas the that t one distance doublethe other)then the time-intervals is required forthesepassages ould equal. But foroneandthe w be samebody to falleightfeet and fourfeetin the sametime is possible nlyinthecaseofinstantaneous o [discontinuous] motion; but observationhows that the motionofa falling s us bodyoccupies time,andlessofit incovering distance ffour a o feetthan ofeightfeet;therefore isnottruethat itsvelocity it increases in proportion tothespace. Thefalsityofthe otherproposition aybeshown m withequal clearness. orifweconsider single F a striking bodythedifference ofmomentum its blows dependonlyupondifference in can of velocity;forif the strikingbodyfallingfroma doubleheight were deliver blowofdouble to a momentum, wouldbenecesit saryfor this bodyto strikewith a doubled velocity; ut with b this doubledspeedit wouldtraversea doubledspacein the sametime-interval; observation howevershows that the time required forfallfromthegreater eightslonger. h i SAQI_. Youpresentthese recondite atterswithtoo much m evidence ease;this greatfacilitymakesthemlessappreand ciatedthan theywouldbe had theybeenpresented a more in abstrusemanner. For, in my opinion,peopleesteemmore lightlythat knowledge hichtheyacquire w withso little labor thanthatacquiredhrough t longandobscureiscussion. d S_LV. thosewhodemonstrate ithbrevityand clearness If w the fallacy ofmanypopular beliefs ere w treated withcontempt insteadof gratitudehe injurywouldbe quitebearable; uton t b the otherhandit is veryunpleasant nd annoying seemen, a to who

[204]

THIRD DAY 169 whoclaimtobepeers ofanyone certain ina fieldofstudy,take for grantedcertainconclusions whichlater are quicklyand easilyshownbyanother be false. I donot describe to sucha feeling one of envy,whichusually as degeneratesntohatred i andanger againstthose whodiscover suchfallacies;would I call it a strongdesireto maintainolderrors,ratherthan accept newlydiscovered truths. Thisdesire timesinduceshemto at t uniteagainstthesetruths,although heartbelievingnthem, at i merely forthepurpose loweringheesteem whichcertain of t in othersareheldbytheunthinkingrowd.Indeed, haveheard c I fromour Academician any suchfallacies eld as true but m h easily refutable; oftheseI haveinmind. some SAGI_. must not withholdthemfromus, but, at the You propertime,tellus aboutthemeventhough anextrasessione b necessary.But now,continuing thethreadofourtalk,it would seemthat upto thepresent ehaveestablished w thedefinition of uniformlyaccelerated motionwhichis expressed follows: as A motionis said to be equallyor uniformly accelerated when,starting fromrest,itsmomentum (celeritatis momenta) receivesqualincrements equaltimes. e in SALv. hisdefinitionstablished, T e theAuthor makesa single assumption, namely, The speedsacquired one and the samebodymoving by downplanesof differentnclinations re equalwhenthe i a heights oftheseplanes areequal. By theheightofan inclined lane p wemeantheperpendicular letfallfromthe upperendoftheplaneuponthehorizontal line drawnthroughthe lowerend of the sameplane. Thus, to illustrate,et the lineABbe horizontal, lettheplanesCA l and andCDbeinclinedoit; thentheAuthorcallstheperpendicular t CB the "height"of theplanesCAand CD;he supposeshat t the speedsacquiredby one and the samebody,descending along theplanesCAandCDtotheterminal pointsAandD are equalsincethe heightsof theseplanesarethe same,CB; and alsoit mustbe understoodhatthis speedisthat whichwould t beacquired ythe same b body.alling C toB. f from Sagr.

[205]

17o THE TWO NEW SCIENCESOF GALILEO SAog. ourassumption Y appearsto meso reasonablehat it t oughtto beconceded ithoutquestion, w provided ofcourse there are no chanceor outsideresistances, that the planesare and _o hard and smooth,and that the figure the moving of bodyis perfelly round, o that neither s plane normoving bodyisrough.Allresistance and oppositionhaving been removed,my reason tells A me at oncethat a heavyand perFig. 5 4 fecdyoundballdescendinglong r a the linesCA,CD, CB wouldreachthe terminalpointsA, D, B,withequalmomenta [impeti eguah]. SALV. Yourwordsarevery plausible; ut I hopeby experib mentto increaseheprobability anextentwhich t to shallbelittle shortofa rigiddemonstration. Imaginethis page to represent verticalwall,with a nail a drivenintoit; and fromthe naillet therebe suspended lead a bulletof oneor twoounces meansof a finevertical hread, by t AB,sayfrom fourto sixfeetlong, nthiswalldrawa horizontal o lineDC,at rightanglesto theverticalthreadAB,whichhangs abouttwofinger-breadths frontof thewall. Nowbringthe in threadABwiththeattachedballintotheposition ACand setit free;firstit willbeobserved descend to alongthe arc CBD,to passthe pointB, and to travelalongthe arcBD,till it almost reaches thehorizontal CD,a slightshortage being caused bythe resistance f the air and the string;fromthis we mayrightly o inferthat theballin its descentthroughthe arcCBacquired a momentum [impeto] reaching whichwasjust sufficient on B, to carryit througha similar BDto the sameheight. Having arc repeated thisexperiment manytimes,letusnowdrivea nailinto the wallcloseto the perpendicular AB,say at E or F, so that it proje_sout some fiveor sixfinger-breadths orderthat the in thread,againcarrying thebulletthrough thearcCB,maystrike uponthenailE whenthebulletreaches B,andthuscompelt to i traversethe arcBG,described aboutE as center. Fromthis
we

[206]

THIRD DAY :y: we canseewhatcanbe doneby thesame momentum [impeto] whichpreviouslytarting the same s at pointB carriedhesame t bodythrough arcBD tothehorizontal the CD. Now,gentlemen,youwillobserve withpleasurehat theballswingsothe t t pointG in the horizontal, nd youwouldseethe samething a happenif theobstacle wereplacedat some lower point,sayat F,aboutwhich theballwould describe thearcBI,theriseofthe A/

Fig. 6 4 ballalways terminatingxa&ly e onthe lineCD. Butwhenthe nailis placedso lowthat the remainder thethreadbelow of it willnotreachtotheheightCD(which would happen ifthenail wereplacednearer than to the interse&ion B ofABwiththe horizontal D)thenthe threadleapsoverthe nailand twists C itselfaboutit. Thisexperiment leaves noroomfor doubtas to the truth of oursupposition; forsince thetwoarcsCBandDBareequaland similarly placed, he momentum t [momento] acquired bythe fall through thearcCBisthesameasthatgained byfallthrough the arcDB; but themomentum [momento] acquired B,owingto at fallthrough CB,isableto liftthesame body[mobile] through the arc BD; therefore, he momentum t acquired the fall BD is in equalto that whichliftsthe samebodythrough samearc the fromB to D; so,in general, verymomentum e acquired fall by through

[207]

i7z THE TWO NEW SCIENCESOF GALILEO throughan arc is equalto that whichcanlift the samebody through samearc. But allthesemomenta the [momentz] which causea risethroughthe arcsBD,BG,and BI are equal,since they are producedby the samemomentum,gainedby fall through CB,as experiment shows.Therefore llthe momenta a gained byfallthrough thearcsDB,GB,IBareequal. SAcR. he argument eemsto meso conclusivend the exT s a perimentso welladaptedto establishthe hypothesishat we t may,indeed,consider asdemonstrated. it SALV. donotwish,Sagredo,hat we troubleourselvesoo I t t much aboutthismatter,since wearegoing toapplythisprinciple mainlyin motions hichoccuronplanesurfaces, notupon w and curved,alongwhichacceleration ariesin a mannergreatly v different fromthat which wehaveassumed forplanes. So that, although aboveexperiment howsus that the the s descent fthemoving o bodythrough arcCBconfers ponit the u momentum[momento] sufficiento carry it to the same just t heightthrough anyofthearcsBE), G,BI, wearenotable,by B similarmeans,to showthat the eventwouldbe identicalin the caseofa perfedtlyoundballdescendinglong r a planes whose inclinations re respe&ivelyhe sameas the chordsof these a t arcs. It seems likely, ntheotherhand,that, since o theseplanes formanglesat the pointB,they willpresentan obstacle the to ballwhichhasdescendedlong chordCB,andstartsto rise a the along chordBD,BG,BI. the In strikingtheseplanessomeof its momentum [impeto] will belostanditwillnotbeableto riseto theheightofthelineCD; but this obstacle, hichinterferes w with the experiment, once removed,t is clearthat the momentum i [im2e[v] (which gains in strengthwithdescent) illbe ableto carrythe bodyto the w sameheight. Let us then,for the present,take this as a postulate,the absolute truth ofwhichwillbe established henwe w find that the inferencesromit correspondo and agreeperf t fe&lywithexperiment.Theauthorhavingassumed thissingle principle passesnextto the propositions whichhe clearlydemonstrates;hefirstoftheseisasfollows: t

[ o8]

THIRD DAY 173 T_EoP, I, PRo_osmoN x_ I The timeinwhichany spaceis traversed a bodystartby ingfromrestanduniformlyccelerated a isequaltothetime in whichthat samespacewouldbetraversed thesame by bodymovingat a uniform speedwhose valueisthe mean ofthe highestspeedandthe speedjustbeforeacceleration began. Let us represent y the lineABthe timein whichthe space b CD is traversed a bodywhichstartsfromrest at C and is by uniformlyaccelerated;et the finaland highestvalueof the l speedgained duringthe intervalABbe representedytheline b EB drawnat right anglesto AB;drawthe lineAE, then all linesdrawnfromequidistant ointsonABand parallelto BE p willrepresentheincreasingalues t v ofthespeed, c beginning iththe instantA. Let thepointF w G _. bisecCt lineEB; drawFG parallelto BA, the andGAparallelo FB,thusforming parallelt a ogram AGFB whichwillbeequalinareato the triangle AEB,since sideGFbisecCts the theside AEat thepointI; forif theparallelinesinthe l triangle AEBareextended GI, thenthesum x, to ofalltheparallels ontainednthequadrilateral/ c i isequalto thesumofthose contained inthetriangleAEB;for thosein the triangleIEF are equal to thosecontained the triangle in GIA, while thoseincluded inthetrapezium AIFBare i common.Sinceeachandeveryinstant ftime_ 1_ o in the time-interval ABhas its corresponding j pointon the lineAB,fromwhichpointspar,_ allels rawninandlimitedbythetriangle d AEB D represent he increasingaluesof thegrowing Fig.47 t v velocity, nd since a parallels ontained ithinthe re.angle repc w resentthevalues speedwhich ofa isnotincreasing, butconstant, it appears,nlikemanner,hatthemomentamomenta] i t [ assumed bythe moving bodymayalsoberepresented, inthe caseof the accelerated otion,by the increasingarallels f the triangle m p o

AEB;

174 THETWONEWSCIENCES FG.ATJTFO O AEB, nd,inthecase a oftheuniform motion, bytheparallels of thereccangle For, hatthemomenta GB. w maylackinthefirst partoftheaccelerated motion (thedeficiency ofthemomenta being represented bytheparallels ofthetriangle GI)ismade A upbythemomentaepresented r bytheparallels ofthetriangle IEF. Hence isclearthatequalspaces ill etraversed it w b inequal times ytwo b bodies, oneofwhich, startingrom moves ith f rest, w Art a uniform acceleration, whilethe momentum of "_ .theother,moving withuniform speed,s one-half i o D. its maximum momentum underaccelerated om _ tion. Q._.D. F._ THEOREM PROPOSITION II, II Thespaces described byabody falling rest from a witha uniformly accelerated motionreto each a otherasthe squares fthe time-intervals o employed intraversing distances. these . Y Letthetime beginning withanyinstant berepA resentedythestraightineABinwhich b l aretaken anytwotime-intervals ADandAE. LetHIrepresentthedistancehrough t which body,starting the fromrestat H, falls withuniform acceleration. If HL represents thespace traverseduring d thetimeinterval AD,andHM that covereduring ind the Bi t tervalAE,thenthe space stands the space R_I to - " LHina ratiowhich isthesquare ftheratioofthe o timeAEto thetimeAD;orwemay saysimply that thedistances HMandHI.arerelated asthesquares Fig. 8 ofAEandAD. 4 DrawthelineACmakingnyangle a whatever withthe line AB;andfromthepoints andE, drawtheparallel DO D lines andEP;ofthese twolines, DOrepresents thegreatest elocity v attained during interval the AD,while EPrepresents themaximum velocity acquired uringthe interval E. But it has d A just beenprovedthat so far as distances traversed areconcerned

[2o9]

THIRD DAY _75 cernedit is precisely samewhethera bodyfallsfromrest the with a uniform accelerationrwhether fallsduringan equal o it time-interval witha constantspeedwhichisone-halfhe maxt imumspeedattainedduringthe accelerated otion.It follows m therefore thedistances 1VI that H andHL arethesameaswould betraversed, uring time-intervals d the AEandAD,byuniform velocities equalto one-half hoserepresented y DO and EP t b respecCtively. therefore, ne can showthat the distances If, o HM and HL are in the sameratioas the squares ofthe timeintervals EandAD,ourproposition illbeproven. A w But in the fourthproposition fthe firstbook[p.157above] o it hasbeenshown thespacestraversed that bytwoparticlesn i uniform motion beartooneanothera ratiowhich isequalto the productof the ratioof the velocities y the ratioof thetimes. b But inthiscasetheratioofthevelocitiessthesameasthe ratio i ofthe time-intervalsforthe ratioofAE to AD isthe sameas ( that ofx/4 to I/4 EP DOorofEP to DO). Hencetheratioofthe spaces traversedis the sameas the squared ratioof the timeintervals. Q.v..D. Evidently thenthe ratioofthe distancessthe square the i of ratioofthefinalvelocities, is,ofthelinesEP andDO,since that thesearetoeachotherasAEtoAD. COROT.I ARY I Henceit is clearthat if we takeany equalintervals time of whatever, ounting c fromthe beginning ofthe motion,suchas AD, DE, EF, FG, in whichthe spaces HL,LM,MN,NI are traversed,hesespaces t willbearto oneanotherthesameratio astheseries ofoddnumbers,, 3,5,7;forthisistheratioofthe I differences the squaresof the lines[which of represent ime], t differences whichexceed anotherby equalamounts,this one excess eing b equalto thesmallest line[viz.heonerepresenting t a single time-interval]: orwemaysay[thatthisisthe ratio]ofthe differences ofthe squares ofthe naturalnumbers eginning ith b w unity. While,
[2IO1

176 THE TWO NEW SCIENCESOF GALTT,EO While,therefore, uring d equalintervals ftimethe velocities o increase sthe natural umbers, a n theincrements thedistances in traversed uring d theseequaltime-intervals aretooneanothers a theoddnumbers beginning ithunity. w SA6R. Pleasesuspend discussion moment incethere the fora s just occursto mean ideawhichI wantto illustrate ymeans b of a diagram order in that it maybe dearerboth to you and tome. Let thelineAI represent thelapseoftimemeasured fromthe initialinstantA throughA draw .straightineAFmaking the l ajA any.anglewhatever;., the terminal join /i "pointsI andF; dlvlde thetimeAI inhalf at C; drawCB parallelto IF. Let us consider B as the maximum C valueof "B / ] at velocitywhichincreases p-c thethe beginning, simple fromzero in proportionality to linesdrawnparallelto triangle the interceptson the BC; or ABCof whatisthesamething, etussupposehe l t

2v g _ _t z time;thenI admitwithoutquestion,n velocity increase proportiono the to in t i spacedescribed a bodyfailingin the by i aforesaidmannerwill be equalto the space traversed bythesamebodyduring thesamelength timetravelling itha of argument, w ' viewofthepreceding EC, thethat the uniformspeedequalto halfof It. il. o BC. Furtherlet us imaginethat the Fig.49 [21I] bodyhas fallenwith accelerated otionso that, at the inm stant C, it has the velocityBC. It is clearthat if the body continuedto descendwith the samespeedBC, withoutacceleration,it would in the next time-intervalCI traverse doublethe distancecoveredduringthe intervalAC,with the uniform speed ECwhichishalfofBC;but since thefalling body acquires equalincrements f speedduringequalincrements f o o time, it follows that the velocityBC, duringthe nexttimeinterval

THIRD DAY 177 interval CI willbe increasedby an amountrepresented the by parallelsofthe triangleBFG whichis equalto the triangleABC. ILthen, oneadds to the velocityGI halfof the velocityFG,the highest speed acquiredby the acceleratedmotion and determined by the parallelsof the triangleBFG, he will have the uniformvelocitywith which the same spacewouldhave been describedin the time CI; and sincethis speedIN is three times as great as EC it followsthat the spacedescribedduringthe intervalCI isthree timesasgreatasthat described duringtheinterval AC. Let us imaginethe motionextendedover anotherequal time-intervalIO, and the triangleextendedto APO; it is then evident that if the motioncontinuesduringthe intervalIO, at the constantrate IF acquiredby acceleration duringthe timeAI, the spacetraversedduringthe intervalIO willbe fourtimesthat traversedduringthe firstintervalAC, becausethe speed IF is fourtimes the speedEC. But ifwe enlargeourtriangle asto so includeFPQ whichis equalto ABC, stillassumingLhe acceleration to be constant,we shalladd to the uniformspeedan incrementRQ, equal toEC; then the valueofthe equivalentuniform speedduring the time-intervalIO willbe five timesthat during the firsttime-intervalAC; therefore spacetraversedwillbe the quintuplethat duringthe first intervalAC. It is thus evident by simple computationthat a movingbody starting fromrest and acquiringvelocityat a rate proportionalto the time, will, during equal intervals of time, traversedistances which are relatedto eachother as the oddnumbersbeginningwith unity, I, 3, 5; * or considering total spadetraversed,that covered the [212[ in doubletimewill be quadruplethat coveredduringunit time; in triple time, the space is nine timesas greatas in unit time. *Asillustrating thegreater legance e andbrevity ofmodernnalytical a methods, onemayobtainthe result f Prop.n directly the funo from damental equation where gistheacceleration ofgravity ands,thespace traversedetween b theinstants txandh. Ifnow t2-tl= I,sayonesecond, s = g]_ + ta) then (t2 where t_+tx, mustalways bean oddnumber, seeing thatit isthesumof twoconsecutive termsintheseries fnaturalnumbers. [Trans.] o

s=I/,g(t',-t l) =g/2 +t3

tl)

I78 THE TWO NEWSCIENCESOF GALILEO Andin generalhe spaces t traversed arein +.he duplicate ratioo_f thetimes,i.e.,inthe ratioofthesquares fthe times. o SnwP. truth, I findmorepleasure this simple clear In in and argument fSagredo o thanin theAuthor'sdemonstration which to me appearsrather obscure;so that I am convinced that mattersareas described, oncehavingaccepted definition f the o uniformly accelerated otion. But asto whether acceleram this tion is that whichone meetsin nature in the caseof falling bodies, amstilldoubtful; nd it seems me,notonlyformy I a to ownsakebut alsofor all thosewhothink as I do, that this wouldbe the propermomentto introduce oneof thoseexperimentswandthere are many of them, I understand--which illustrate several aystheconclusions in w reached. SALV. Therequest hichyou,asa manofscience, w make,is a veryreasonable ne;for this is the custom--and o properly so--in thosesciences where mathematical demonstrations areapplied to naturalphenomena, is seenin the caseof perspecCtive, as astronomy, mechanics, music,and otherswherethe principles, onceestablishedywell-chosen b experiments, becomehefoundat tionsofthe entiresuperstrucCture.hopetherefore willnot I it appearto be a wasteoftimeifwediscuss t considerable a length this first and most fundamental uestionupon whichhinge q numerous consequences whichwe have in this bookonlya of smallnumber,placedthereby the Author,whohas doneso muchto opena pathwayhitherto closed mindsofspeculative to turn. So far as experiments othey have notbeenneglecCted g bythe Author;andoften,inhis company, haveattemptedin I the following annerto assuremyselfthat the acceleration m acCtually experienced byfalling bodies isthatabove described. A pieceof woodenmouldingor scantling, bout I2 cubits a long,half a cubit wide,and three finger-breadths thick, was taken; on its edgewascut a channela littlemorethan one fingernbreadth;having i madethisgroove erystraight, mooth, v s and polished,and having lined it with parchment,also as smoothand polished possible, e rolledalongit a hard, as w smooth,and very round bronzeball. Having placed this

[ I3]

board

THIRD DAY 179 boardin a sloping position, y lifting b oneendsomeoneortwo cubitsabovetheother,werolledtheball,asI wasjust saying, alongthechannel, noting,n a manner resently bedescribed, i p to the timerequiredto makethe descent.We repeated exthis perimentmorethanoncein orderto measure timewithan the accuracysuch that the deviationbetweentwo observations neverexceeded one-tenthof a pulse-beat.Havingperformed this operation andhavingassured ourselves ofitsreliability, we nowrolledthe ballonlyone-quarterhelengthofthe channel; t and havingmeasured timeof its descent, e foundit prethe w cisely one-half ftheformer. Nextwe o triedotherdistances, comparingthetimeforthewhole length withthatforthehalf,orwith that for two-thirds, orthree-fourths, orindeed foranyfra_ion; in suchexperiments, repeated fullhundredtimes,wealways a foundthat thespaces traversed wereto eachotherasthesquares ofthe times,and this wastruefor allinclinations ofthe plane, i. e.,ofthe channel, longwhichwe rolledthe ball. Wealso a observed thatthetimesofdescent, orvariousnclinations f i ofthe plane,boreto oneanotherpreciselyhat ratiowhich, sweshall t a seelater,theAuthorhadpredicted demonstrated and forthem. Forthe measurement time,weemployed largevessel f of a o water placedin an elevatedposition;to the bottomof this vessel assoldered pipeofsmalldiameter iving thinjet of w a g a water,which wecolle_ed small lassduring ina g thetimeofeach descent, hether w forthewhole length ofthechannel orfora part ofits length;the waterthuscollecCted wasweighed, aftereach descent, na veryaccurate o balance;hedifferences t andratios of theseweights aveus thedifferences g andratios ofthetimes, nd a this with suchaccuracy that althoughthe operation wasrepeatedmany,manytimes, herewasnoappreciable t discrepancy in theresults. S_P. I would liketo havebeenpresentat theseexperiments; but feelingconfidencen the carewithwhichyou performed i them, and in the fidelitywith whichyou relatethem, I am satisfiedndacceptthemastrueandvalid a SM, Therrwe v. canproceed ithout iscussion. w d [2_4]

x8o THE TWO NEW SCIENCES GALILEO OF COROT,I,ARYI I Secondly, itfollowshat, startingfromanyinitialpoint,if we t takeany twodistances, traversed anytime-intervals in whatso"i_ ever,thesetime-intervals earto oneanotherthe same b ratioas oneofthe distanceso themeanproportional t of the twodistances. Forif wetaketwodistances T andSYmeasured S from the initialpointS,themeanproportional ofwhichisSX, T the timeoffall through is to thetimeoffallthrough ST .:I: YasSTistoSX; oronemaysaythetimeoffallthrough S SYis to thetimeoffallthrough STasSYisto SX. Now sinceit has beenshown the spaces that traversed in are Y"thesameratioasthesquares fthe times; ndsince, oreo a m Fig. Oover ratio of the spaceSY to the spaceST is the 5 ' the squareof the ratioSYto SX,it followshat the ratioofthe t timesof fall throughSYand ST is the ratioofthe respecdve distances YandSX. S SCHOLIUM The abovecorollary asbeenprovenfor the caseofvertical h fall;but it holdsalsoforplanesinclined anyangle;forit isto at beassumed alongtheseplanesthevelocity that increasesn the i sameratio,that is,inproportiono the time,or,ifyouprefer,as t the series ofnaturalnumbers.* SAr,v. Here,Sagredo, should I llke,if it be nottootediousto Simplicio, interruptfor a momentthe presentdiscussionn to i orderto makesomeadditions n the basisofwhathas already o beenprovedandof whatmechanical principles wehavealready learnedfromour Academician. This additionI makefor the betterestablishment logical ndexperimental on a grounds, ofthe principle whichwe have aboveconsidered; what is more and important, forthepurpose ofderiving geometrically', it afterfirst demonstratingsingle a lemma hich w isfundamental thescience in ofmotion [impet_]. *Thedialogue intervenes which between thisScholium andthefollowingtheorem waselaborated byViviani, thesuggestion at ofGalileo. See National Edition, 23. [Trans.] viii,

THIRD DAY I8i SAGR.the advance whichyou proposeto make is such as If will confirmand fully establishthesesciences motion,I will of gladly devoteto it any lengthof time. Indeed,I shall not only be gladto haveyou proceed, ut I begofyou at onceto satisfy b the curiosity whichyou have awakened me concerningour in y proposition; thinkthat Simpliciosofthe samemind. andI i SIMP. QuiterighL SALV. Sincethen I haveyourpermission, usfirstof allconlet siderthis notableacc, the momenta speeds[imomenti le that or o velocitY] ofone and the samemovingbodyvary with the inclinationof the plane. The speedreachesa maximumalonga verticaldirection,and for other dirertionsdiminishesas the plane divergesfrom the vertical. Therefore the impetus, ability, energy, [l'impeto,il talento,l'energia] r, one might say, the momentum[il momento] o of descentof the movingbody is diminishedby the plane upon whichit is supportedand alongwhichit rolls. For the sakeofgreaterclearness reCtthe lineABperpendicular e to the horizontalAC; next draw AID,AE, AF, etc., at different inclinations the horizontal. ThenI saythat allthemomentum to of the fallingbodyis alongthe verticaland is a maximum whenit fallsin that dirertlon;the momentumis less alongDA and still lessalongEA, and evenlessyet alongthe moreinclined planeFA. Finallyon the horizontalplane the moB mentum vanishesaltogether;the body J) findsitself in a conditionof indifference asto motionor rest;hasnoinherenttendIg encyto movein any dirertion, nd offers a noresistanceto beingset in motion. For just as a heavy bodyor systemof bodies cannotof itselfmoveupwards, r recede o from the commoncenter [comuncentro] (31 H towardwhichall heavythingstend, soit ! A is impossible forany bodyof its ownac- O cord to assumeany motionother than Fig.51 one which carriesit nearerto the aforesaidcommoncenter. Hence,alongthe horizontal,by whichwe understanda surface, everypointofwhichisequidistant romthissamecommoncenter, f the bodywillhave nomomentum whatever. This

[215]

I82 THE TWO NEW SCIENCES OF GALILEO [516] This changeof momentumbeingclear,it is herenecessaryfor me to explainsomethingwhichour Academician wrote when in Padua, embodyingit in a treatise on mechanics prepared solely for the use of his students, and provingit at lengthand conclusivelywhen considering originand nature of that marvellous the machine,the screw. What he provedis the mannerin whichthe momentum[impeto] arieswiththe inclinationofthe plane,asfor v instancethat oftheplaneFA,oneendofwhichiselevatedthrough a verticaldistanceFC. This dire_ionFC is that alongwhichthe momentum a heavybodybecomesa maximum;let us discover of whatratio thismomentumbearsto that of the samebodymoving alongthe inclinedplane FA. This ratio, I say, is the inverseof that of the aforesaidlengths. Suchis the lemmaprecedingthe theoremwhichI hopeto demonstratea littlelater. It is clearthat the impelling force[impeto] cCting a bodyin a on descentis equal to the resistanceor least force[resistenzaforza o minima]sufficiento holdit at rest. In orderto measurethis force t and resistance_orzae resistenza] proposeto use the weightof I another body. Let us placeupon the plane FA a body G connecCted the weightH by means of a cord passingover the to point F; then the body H will ascend or descend,along the perpendicular,the same distancewhichthe body G ascendsor descendsalongthe inclinedplane FA; but this distancewill not be equalto theriseorfallofG alongthe verticalin whichdirection aloneG, asother bodies,exertsits force[resistenza]. hisis clear. T For if weconsiderthe motionof the bodyG, fromA to F, in the triangleAFC to be madeup ofa horizontalcomponent C and a A verticalcomponentCF, and rememberthat this bodyexperiences no resistanceto motionalongthe horizontal(becauseby such a [2x71 motionthe bodyneithergainsnor losesdistancefromthecommon centerof heavy things) it followsthat resistanceis met only in consequence the body risingthroughthe verticaldistanceCF. of Sincethenthe bodyG in movingfromA to F offersresistance only in so far as it rises throughthe vertical distanceCF, whilethe other bodyH must fallverticallythroughthe entire distanceFA, and sincethis ratio is maintainedwhetherthe motionbe largeor small,the two bodiesbeinginextensiblyconnected,we are able to assertpositivelythat, in.caseof equilibrium(bodiesat rest) the momenta,

THIRD DAY 183 momenta,thevelocities,or theirtendencyto motion[propensioni al mot@i. e., the spaceswhichwouldbe traversedby them in equaltimes,mustbe in the inverseratioto theirweights. This is whathas beendemonstratedin everycaseof mechanical otion.* m So that, in orderto hold the weight(3at rest, onemustgive H a weightsmallerin the sameratio asthedistanceCFis smaller han t FA. If wedothis, FA:FC=weightG:weightH; then equilibrium willoccur,that is,the weights andG willhavethe sameimpellH ing forces [momentieguali],and the two bodieswill come to rest. And sincewe are agreedthat the impetus,energy,momentum or tendencyto motionof a movingb.ody asgreat as theforceor is least resistance [forzaoresistenza zn_ma] m sufficiento stopit,and t sincewe have foundthat the weightH is capableof preventing motionin theweightG, it followshat the lessweightHwhoseent tireforce[momento totale]s alongtheperpendicular, i FC,willbean exactmeasureof the component f force[momento o parziale] which the largerweightG exerts alongthe planeFA. But the measure of the total force[total omento] nthe bodyG is its ownweight, m o sinceto preventits fall it is only necessaryto balanceit with an equal weight,providedthis secondweightbe freeto movevertically;thereforethe component f the force[momentoarziale] n o p o G alongthe inclinedplaneFAwillbearto themaximum andtotal forceon this samebodyG alongthe perpendicular C the same F ratio as the weightH to the weightG. This ratio is, by construction, the samewhichthe height, FC, of the inclinedplane bearsto the lengthFA. We have here the lemmawhichI proposedto demonstrateand which, as you will see,has been assumedby our Authorin the secondpart of the sixthproposition ofthepresenttreatise. SAGg. rom what you have shownthus far, it appearsto me F that one might infer, arguingex aequaliconla proportioneerp turbata,hat the tendenciesmomentz] t [ ofoneandthe samebodyto movealongplanesdifferently inclined, ut havingthe samevertib cal height, as FA and FI, are to each other inverselyas the lengthsofthe planes.
[218]

SAT.V. Perfectlyright. This point established,I pass to the demonstration ofthe following theorem: *A nearapproach the principle virtualworkenunciated y to of b JohnBernoulli 17I [Trans.] in 7.

THE TWO NEW SCIENCES OF GALII.F.O If a bodyfallsfreelyalongsmoothplanesinclined anyangle at whatsoever, ut of the sameheight,the speedswith whichit b reachesthe bottomarethe same. First we must recallthe fact that on a planeof any inclination whatevera body starting from rest gains speed or momentum [laquantit& dell'impao]in directproportion the time, in agreeto mentwiththe definition ofnaturally accelerated motiongivenby theAuthor. Hence,ashe hasshownin the preceding proposition, the distancestraversedare proportionalto the squaresof the times and thereforeto the squares of the speeds. The speed relations arehere the sameas in the motionfirst studied[i. e., vertical otion], m sincein eachcasethe gainof speedis proportional to the time. LetAB be an inclinedplanewhoseheightabovethe levelBC is AC. Aswehave seenabovethe forceimpelling[l'impeto] body a A tofallalongthevertical Cis totheforce A clinedplaneAB asAB is toAC. Onthe inclineAB, lay off AD a third proportionaltoAB and AC; then the forceproducingmotionalongAC is to that along whichdrivesthe samebodyalongthe in" AB (i.e.,alongAD)as thelengthACis to B B IJthe lengthAD. Andthereforethe body Fig.52 willtraverse the spaceAD, alongthe inclineAB,in the sametimewhichit wouldoccupyin fallingthe vertical distanceAC, (sincetheforces[moment,] arein thesameratio as thesedistances);alsothe speedat C is to thespeedat D asthe distanceAC is to the distanceAD. But, accordingtothe definitionof acceleratedmotion,the speedat B is to the speedof the samebody at D asthe timerequiredto traverseAB is to the time required forAD; and, according the last corollary to ofthe second proposition,he timeof passingthroughthedistanceAB bearsto t the time of passingthroughAD the same ratio as the distance AC (a mean proportionalbetweenAB and AD)to AD. Accordinglythe two speedsat B and C eachbearto the speedat D the sameratio, namely,that of the distances andAD;hencethey AC areequal. This is the theorem whichI setoutto prove. Fromthe abovewe arebetterableto demonstrate thefollowing thirdproposition f theAuthorin whichhe employsthe following o principle,namely,the time required traverse inclinedplane to an is

I84

THIRD DAY i85 is to that requiredto fall throughthevertical heightofthe plane inthe sameratioasthelengthofthe planeto itsheight. For,according the secondcorollary to ofthesecondproposition, if BA representsthe time requiredto passover the distanceBA, the time requiredto passthe distanceADwillbe a meanproportional betweenthese two distancesand will be representedby the lineAC; but ifAC representsthe timeneededto traverseAD it willalsorepresentthe timerequiredto fallthroughthe distance AC,sincethe distancesAC and AD aretraversedin equaltimes; consequently AB representsthe time requiredforAB then AC if willrepresentthe timerequiredforAC. Hencethetimesrequired to traverseABand AC are to eachotheras the distances and AB AC. In like mannerit can be shownthat the time requiredto fall throughAC is to the time requiredfor any other inclineAE as the lengthAC isto thelengthAE; therefore, xaequali, e thetimeof fall alongthe incline is to that alongAEas the distanceABis AB to the distanceAE, etc.* Onemightby application this sametheorem,as Sagredowill of readilysee,immediatelydemonstratethe sixthpropositionof the Author; but let us here end this digression which Sagredohas perhapsfoundrathertedious,thoughI considerit quiteimportant forthe theory ofmotion. SACR. the contraryit has givenme great satisfaction,and On indeed I find it necessaryfor a completegrasp of this principle. SALv. willnowresumethereadingof thetext. I

[219]

[215]

TttEOltEM PROPOSXTXON III, III If one and the same body, starting from rest, fallsalong an inclinedplaneand alsoalonga vertical,eachhavingthe sameheight, the times of descentwill be to eachother as the lengthsof the inclinedplane and the vertical. Let AC be the inclinedplaneandAB the perpendicular, ach e having the same vertical height above the horizontal,namely, BA;hen I say, the time of descentof one and the samebody t *Putting_hisargumentn a modern i andevidentnotation, nehas o AC= 1h gt_andAD= 1[_ gt_.If now_-_2= tLB. D,itfoUows _-A at oncethatt_ = t_. [Trans.] Q" "_" _

186 THE TWO NEW SCIENCESOF GALILEO [zI6l alongthe planeACbearsa ratioto the timeoffall alongthe perpendicular AB,whichis the sameasthe ratioofthe length ACto thelengthAB. LetDG,EI and LFbe anylinesparallel to the horizontal B; then it follows C from whathaspreceded a bodystartingfrom that A willacquire thesamespeedat the pointG a Dasat D, sincein eachcasethe verticalfallis the same;in likemannerthe speeds I and at E willbe the same;soalsothoseat L andF. And ingeneralhe speeds tthe twoextremit a tiesof anyparalleldrawnfromany pointon AB to the corresponding pointonACwillbe _ _ equal. Fig.53 Thus the two distances and AB are AC traversed the samespeed. But it has alreadybeenproved at [zI7] that iftwodistances retraversed bodymoving a bya withequal speeds,thenthe ratioofthe timesofdescent illbethe ratioof w the distances themselves; therefore,he timeof descentalong t ACis to that alongAB as the lengthoftheplaneAC is to the vertical istance d AB. O.v.D. SacR.It seems methat theabovecould to havebeenproved clearly briefly and onthebasisofa propositionlready a demonstrated,namely,that the distancetraversedin the caseof accelerated otionalong m ACorABis the sameasthat covered [zr9] by a uniform speed whose valueisone-halfhemaximum t speed, CB;the twodistances C andAB havingbeentraversed the A at sameuniformspeedit is evident,fromProposition, that the I timesofdescent illbeto eachotherasthedistances. w COROLLARY Hencewemay inferthat the timesof descentalongplanes havingdifferentnclinations, thesamevertical eightstand i but h to

[181

THIRD DAY 187 to one another the sameratioas thelengthsof theplanes. in Forconsider nyplaneAMextending A tothe horizontal a from CB;thenit maybe demonstrated inthe samemannerthatthe timeof descentalongAM is to the timealong as thedisAB tanceAMistoAB; butsincethetimealong istothat along AB ACasthelength ABistothelengthAC, itfollows, excequali, that asAMistoACsois the timealong AM to thetimealong AC.

IV,PRoPosmo IV
Thetimesof descentalongplanesof the samelengthbut of differentinclinations re to eachotherin the inverse a ratio of the squareroots of their heights Froma single pointB drawthe planesBAand BC,having the samelengthbut differentnclinations; i letAE and CD be horizontalines drawnto meet the perpendicular D; and l B [22ol letBErepresenthe heightoftheplaneAB,andBDtheheight t ofBC; alsoletBIbe a meanproportional BDandBE; then to the ratioof BD to BI is equalto the B squareroot of the ratioof BD to BE. .m, Now,I say,the ratioofthe timesofde/,I/ scentalong BAandBCistheratioofBD / / toBI;so that the timeof descent long a / / BAis relatedto the heightof the other / / planeBC,namelyBD asthe timealong / / BC is relatedto the heightBI. Nowit / / _mustbe provedthat the timeof descent & _/ along BAisto thatalong BCasthelength _y BDisto thelength BI. DrawIS parallelto DC;and sinceit Fig.54 hasbeenshown the timeoffallalongBAisto that along that the verticalBE as BAisto BE;and alsothat the timealong BE isto that alongBDas BEis to BI; and likewisehat the t timealongBID that alongBCasBDisto BC,or asBI to isto BS;it follows, excequali, thetimealong that BAisto that along BC asBAto BS,or BCto BS. However, isto BSas BD BC isto BI;hence follows ur proposition. o

x88 THE 2_zO NEW SCIENCESOF GALII.EO THEO_ V, PRoPosmo_ V The timesofdescentalong planes differentength, lope of l s and heightbearto one anothera ratiowhichis equalto the product f theratioofthe lengthsbythe squareroot o oftheinverse ratiooftheirheights. Draw the planesAB and AC, havingdifferentinclinations, lengths, ndheights. My theorem a thenis that the ratioofthe Atimeof descent long a ACto that along ABis .41 equaltotheproduct f theratioofACto AB o //' [ by thesquare rootoftheinverseatiooftheir r /./ ! heights. / / I ForletADbea perpendicular which to are / / ]Gdrawnthe horizontalinesBGand CD;also l B /_ i letAL bea meanproportional theheights to / ] AG and AD;fromthe pointL drawa hori1_/ [I_zontallinemeeting inF; accordingly AC AF / ] willbea meanproportional between ACand / [ AE. Nowsince thetimeofdescentalong AC C/ 110s to that alongAE as the lengthAF is to i Fig.55 AE; and sincethe timealong is to that AE along asAE istoAB, it isclearthat thetimealong AB ACisto thatalong asAF istoAB. AB Thusit remainso beshown theratioofAF toABisequal t that to the productofthe ratioofACto ABby the ratioofAG to AL,whichis the inverse ratioof thesquarerootsoftheheights DA and GA. Nowit is evidentthat, if we considerhe line t ACin connection ithAFandAB,the ratioofAFto ACisthe w sameasthat orAL toAD,orAGto ALwhichisthesquare root of the ratioof theheightsAG andAD;but theratioofACto AB is theratioof the lengthsthemselves.Hencefollowshe t theorem. THEOm_M PROPOSITIOn VI, VI I__rom thehighestorlowestpointina verticalcircle there be drawn anyinclined lanesmeeting p thecircumference the times

[2i]

THIRD DAY I89 timesofdescentalongthesechordsare eachequalto the other. Onthe horizontal lineGH construfit verticalcircle.Fromits a lowest oint--thepointof tangency p withthehorizontal,draw the diameter Aand fromthe highest oint,A, drawinclined F p planesto B and (2,any points whatever the circumference; on thenthe timesof descentalong B theseare equal. DrawBD and CEperpendicular the diameto ter; make AI a mean propor- tionalbetween theheights fthe o planes,AE and AD; and since the refitangles A.AEand FA. F ADarerespefitively equalto the squares ofACandAB,while the rectangle FAakEis to the refit.... angleFA_AD AE is to AD, G as _ it followshat the square t ofAC Fig.56 is to the squareofAB as the lengthAE is to the lengthAD. But sincethe lengthAE isto AD as the square ofAI isto the squareofAD, it followshat the squareson the linesACand t ABareto eachotherasthesquares onthelinesAI andAD,and hencealsothe length is to the lengthABas AI is to AD. AC But it has previouslyeendemonstrated the ratioof the b that time of descentalongAC to that alongAB is equalto the produfit f the tworatiosACto AB and AD to AI; but this o lastratioisthesameasthatofABtoAC. Thereforeheratioof t the timeof descentalong ACto that alongABis the product ofthe tworatios,ACto ABandABto AC. Theratioofthese timesistherefore nity. Hencefollows u ourproposition. By useoftheprinciples fmechanics o [exmechanic#] onemay obtainthe sameresult,namely,hat a falling t bodywillrequire equaltimesto traversethe distances AandDA,indicated C in

thefollowing Layoff Aequal figure. B toDA, ndletfallthe a

[zzzl

perpendiculars and DF; it follows BE fromthe principles f o mechanics

I9o THE TWONEW SCIENCESOF GALII.FO mechanicshat the component the momentum t of [momentum ponderis] acCting alongthe inclined planeABCis to the total momentum [i.e.,the momentum the bodyfalling of freely] as BEis to BA; in likemanner the momentumalong the planeADis to its total momentum[i.e.,themomentum of the bodyfallingfreely] as DFistoDA,ortoBA.Therefore the momentumof this sameweightalongthe plane _ DA isto that along plane the ABC as the lengthDF is to _._ thelength BE;forthisreason, thissameweight illin equal w A mF timesaccording thesecond to Fig. 7 5 propositionf the firstbook, o traversespacesalong planesCAandDAwhichareto each the otherasthelengths E andDF. Butit canbeshown B thatCA isto DA asBE isto DF. Hencethefalling bodyvcill raverse t the twopathsCAand DA in equaltimes. Moreoverhe fa_ that CAiSto DA asBEisto DF maybe t demonstrated follows: as JoinC and D; through drawthe D, lineDGLparallelo AF andcuttingthelineACin I; through t B drawthe lineBH,alsoparallelto AF. Thenthe angleADI willbe equalto the angleDCA,sincetheysubtend equalarcs LAandDA,and sincethe angleDACis common, sidesof the the triangles, ADand DM, aboutthecommon nglewillbe C a proportional eachother; accordinglys CAis to DA so is to a DA to IA,that is as BAis to IA, or asHA is to GA,that is asBEisto DF. _. D. The sameproposition aybe moreeasilydemonstrated m as follows: Onthehorizontal lineABdrawa circle whose diameter DCisvertical. Fromtheupperendofthisdiameter drawany inclinedlane, F, extending p D tomeetthecircumference; then,I say, a body occupythe sametime in fallinglongthe will a planeDF as alongthe diameter DC. For drawFG parallel to

THIRD DAY 191 to ABandperpendicular DC;joinFC; andsince timeof to the fall alongDC is to that alongDG as the meanproportional betweenCD and GD is to GD itself;and sincealsoDF is a meanproportional between and DG, the angleDFC inDC scribed semicircle ina beinga rightangle,and FG beingperpendicular fallalong DCisto that along as f J DG thelength FDisto GD. But it has //,,I already beendemonstrated theF_" that timeof descentalong is to that [ DF along DGasthelength DFistoDG; _ hence thetimesofdescent timeof \ /"_ a DF to DC, it followshat the long t ____ 1_ \ andDCeachbearto thetimeof fall _ _ " /_, alongDG the same ratio; conse- _ \ quentlytheyareequal. In likemannerit maybe shown _ c that ifonedraws thechordCEfrom Fig.58 the lower ndof the diameter, lsothe lineEH parallelo the e a t horizon, andjoinsthepointsE andD, thetimeofdescent long a EC,will ethe sameasthat alongthe diameter, C. b D COROLLARY I Fromthisit followshat thetimesofdescentalong t allchords drawn through eitherCorDareequaloneto another. COROI,ARY J II It alsofollowshat, if fromany onepointtherebe drawna t verticalineandaninclined nealong l o which thetimeofdescent .isthe same,the inclined linewillbe a chordof a semicircle of which theverticallineisthediameter.

[223]

COROllARY III
Moreoverhe timesof descentalonginclined t planeswillbe equalwhenthe vertical eights h ofequallengths theseplanes of are

I92 THE TWO NEW SCIENCESOF GALILEO areto eachotherasthelengths oftheplanesthemselves; thusit is clearthat the timesof descentalongCA and DA, in the figurejust beforethe last, are equal,providedthe vertical heightof AB (ABbeingequalto AD),namely,BE, is to the vertical eight F asCAisto DA. h D SAcR. Pleaseallow meto interruptthelecturefora moment inorderthatI mayclear upanidea which justoccurs tome;one which,if it involvenofallacy, uggests t leasta freakish s a and interesting ircumstance, c suchas oftenoccurs natureand in in therealm ofnecessaryonsequences. c If, fromany pointfixedin a horizontal plane,straightlines be drawnextending indefinitelyn all dire_ions,and ff we i imagine pointto movealong a eachoftheselines withconstant speed,allstartingfromthefixedpointat the sameinstantand moving withequalspeeds, thenit isclearthat allofthesemoving pointswilllie uponthe circumference a circlewhich of grows largerandlarger,always avingtheaforesaidixed h f point asitscenter;thiscircle spreads utinpreciselyhesame o t manner as the littlewavesdo in the caseof a pebbleallowed drop to intoquietwater,where thelmpacc ofthestone startsthemotion in alldirections, whilethe pointof lmpac_c remains center the oftheseever-expanding circular aves. But imagine vertical w a planefromthe highest ointofwhichare drawn p linesinclined at everyangleand extending indefinitely; imagine alsothat heavyparticles descend longtheselineseachwitha naturally a accelerated otionand eachwith a speedappropriate the m to inclination its line. If thesemovingparticles always of are visible, hatwillbethe locus w oftheirpositions t anyinstant a ? Nowtheanswero thisquestion t surprises me,forI amledby_e preceding theoremso believethat theseparticles illalways t w lieuponthe circumference a single of circle, verincreasingn e i sizeasthe particles ecede r fartherandfartherfromthepointat whichtheirmotionbegan. To be moredefinite,et A be the l fixedpointfromwhicharedrawnthe lines AFandM-Iinclined at any anglewhatsoever.On the perpendicular/_B takeany twopointsC _ndDaboutwhich, scenters, irclesredescribed a c a passing

[2241

THIRD DAY I93 passingthrough the point A, and cutting the inclined lines at the pointsF, H, B, E, G, I. Fromthe precedingtheoremsit is clearthat, ifparticlesstart, at the sameinstant, fromA and descendalongthese lines,when one is at E another will be at G and anotherat I; at a later instant A theywillbe foundsimultaneously at F, H and B; these, and indeedan travellingalong an infinite number of differentslopeswill at successive infinite numbe_22_upon particles_ / instants always lie ther a single x ever-expandingircle.Thetwokinds c of motion occurringin nature give rise thereforeto two infinite series of circles,at once resemblingand Fig.59 differingfromeachother; the onetakes its risein the centerof an infinitenumberof concentriccircles;the other has its origin in the contacc, their highestpoints,of an infinitenumberof at eccentriccircles;the formerareproducedby motionswhichare equal and uniform; the latter by motions which are neither uniformnor equal amongthemselves,but whichvary fromone to anotheraccording theslope. to Further, if fromthe two points chosenas originsof motion, we draw lines not only along horizontal and vertical planes but in all direcCtions just as in the formercases,beginning then at a singlepoint ever-expandingirclesare produced,so in the c latter casean infinitenumberof spheresare producedabout a single point, or rather a singlesphere which expandsin size withoutlimit; and this in twoways,onewith the origin at the center,the otheron the surfaceof the spheres. S_v. The idea is really beautifuland worthy of the clever mindofSagredo. Sn_P.Asfor me, I understandin a generalway howthe two kinds of natural motionsgive rise to the circlesand spheres; and yet as to the producCtion ofcircles accelerated by motionand its proof,I am not entirely clear;but the facet hat onecan take t the

I94 THE TWONEW SCIENCES GALII.F.O OF theorigin ofmotion eitherat theinmost enterorattheverytop c ofthe sphere leadsoneto think that theremaybe somegreat mysteryhiddenin thesetrueandwonderful results, mystery a relatedto the creationof the universe(whichis said to be sphericaln shape),and relatedalso to the seatof the first i cause [prima causa]. SaJ.v.I haveno hesitation agreeing ithyou. But proin w found considerations ofthSs kindbelong a higher ciencehan to s t ours[api_taltedottrinehelenostre].Wemustbe satisfied c to belong that class lessworthyworkmen hoprocure to of w from the quarrythe marbleout ofwhich,later,the giftedsculptor produceshosemasterpieces t whichlayhidden thisroughand in shapeless exterior. Now,ifyouplease,etus proceed. l [2z6] THEOREM PROPOSITION VII, VII I{theheights ftwoinclined lanes o p aretoeachotherinthe sameratioasthe squaresof theirlengths,bodiesstarting fromrest willtraversetheseplanesin equaltimes. Taketwoplanes ofdifferentengths nddifferent l a inclinations, AEandAB,whose heightsareAF andAD:letAFbeto ADas Athe squareofAE is to the squareof _A_B; then,I say,that a body,starting fromrestatA,willtraverse theplanes AE andABinequaltimes. Fromthe verticalline,drawthehorizontal parG I_allellinesEF andDB,the latter cuttingAE at G. Since FA:DA=EA*: B--_2, sinceFA:DA=EA:GA, t and i _, followsthat EA :GA--_2 : B--_2. :FHenceBAis a meanproportional beFig. o 6 tweenEA and GA. Now sincethe timeofdescentalong ABbearsto thetimealong AGthe same ratiowhichABbearsto AGand sincealsothe timeof descent alongAGis to the timealongAE asAGis to a meanproportionalbetween GandAE,that is,to AB,it follows, equali, A ex that

THIRD DAY 195 thatthetimealong ABisto the timealong AEasABistoitself. Thereforehe timesare equal. t Q.E.D. THEOm_a P_oPosmoN VIII, VIII Thetimes ofdescent long a allinclinedlanes hich p w intersedt one and the sameverticalcircle,eitherat its highestor lowest oint,areequalto the timeoffallalong p thevertical diameter;orthoseplanes f whichfallshortofthisdiameter thetimesareshorter; orplanes f which cutthisdiameter, the timesarelonger. Let ABbe the verticaldiameter a circle of whichtouches the horizontal lane. It has already p beenproven the timesofdethat _a. scent alongplanesdrawnfrom eitherend, A or B, to the circumferencere equal. In order a to show the timeofdescent } that I alongthe planeDF whichfalls shortof the diameteris shorter wemaydrawtheplaneDBwhich isboth longer ndlesssteeply a indinedthanDF;whencet follows i ..... thatthetimealongDF isless than"' B that along and consequently DB Fig. 6I alongAB. In like manner,it is shownthat the timeof descentalong COwhichcutsthediametersgreater:forit isboth i longerand lesssteeplyinclined than CB. Hencefollowshe t theorem. THEOREM PRoeosmoNX IX, I If from point any onahorizontal planes, llne two inclined at angle,drawn, if are ut yaline any are and they c b which makesithhem w t angles alternately equal angles tothe between planes the these and horizontal, times then the requiredtraversethoseportions to oftheplaneutoffby c the aforesaid line are equal. Through

[227]

*96 THE TWO NEW SCIENCESOF GALILEO Throughthe point C bn the horizontalineX, drawtwo l planesCD and CE inclined any anglewhatever:at any at pointin the lineCD layoffthe angleCDFequalto the angle XCE;let thelineDF cut CEat F sothat the angles CDFand CFDarealternately equalto XCEand LCD;then,I say,the .... timesof descentoverCD sincethe angle CDF is equalo the angle CEby t X x and CF are equal. Now construcCtion, evident it is D _B //'_ c that the angleCFD must For if the common angle DCF be subtracCted _ _: beequalto theangle from DCL. the threeangles the triof Fig. _ 6 angleCDF,together qual e to tworightangles, towhich alsoequalalltheangles ( are which canbedescribedboutthepointConthe lowersideof theline a LX)there remainin the triangle angles,CDFand CFD, two equalto thetwoangles XCEand LCD;but,by hypothesis, the angles CDFandXCEareequal; ence h theremainingngle a CFD isequalo theremainder t DCL. TakeCEequaltoCD;fromthe pointsDandE drawDAandEBperpendicular tothehorizontal lineXL; andfromthe pointC drawCGperpendicular DF. to NowsincetheangleCDGis equalto the angle ECBand since DGCandCBEarerightangles, itfollowshatthetriangles DG t C andCBEareequiangular; consequently DC:CG =CE:EB. But DCisequalto CE,andtherefore GisequaltoEB. Since C also the angles C and atA, in the triangle at DAC,areequalto the anglesat F and (3 in the triangle CGF,we haveCD:DA= FC:CG and, permutando, DC:CF--DA:CG=DA:BE. Thus the ratio of the heightsof the equalplanesCD and CE is thesameasthe ratioofthe lengths CandCF. Therefore, D by Corollary ofProp. thetimesofdescent I VI, along theseplanes willbe equal. Q.E.D. Analternative proofs thefollowing: FSperpendicular i Draw to

[2 8]

THIRD DAY 197 to the horizontaline AS. Then,since the triangleCSF is l similarto the triangleDGC,we have SF:FC--GC:CD; and sincethe triangleCFG is similarto the triangleDCA,we have FC:CG=CD:DA. z. L c _" x: CG=CG:DA. ThereforeCG isa meanproportional between SF andDA,while DA:SF= D-A_:C-G _. Again since x thetriangle issim- _"/_ _ a [[_.---"7_ Hence,ex ACT) SF: cquali, //_x c, we have DA'_C=GC: ilartoand, I_ermutando, CGF, CF thetriangle DA:CG= DC:CF:also Fig.63 b-A2:C-G_=D--C_:_ it hasbeenshown D-A2:C-G_= 2. But that DA:SF. Therefore -C_.'CF*=DA:FS. D Hencefromthe above Prop.VII,sincetheheightsDA andFSof the planesCDand CFareto eachotherasthesquares fthelengths o ofthe planes, it followshat the timesof descent longtheseplaneswillbe t a equal.
THEOREM X, PROPOSITION X

Thetimes ofdescent inclined ofthe along planes same height, utofdifferent b slope, aretoeachotherasthelengths of theseplanes;andthis istruewhether motionstarts the from rest or whetherit is precededby a fall from a constant eight. h Letthepathsofdescent ealong b ABCandABDto thehorizontalplaneDCsothat thefallsalong BDandBCarepreceded y b thefallalong AB;then,I say,that thetimeofdescent long a BD is to the timeofdescentalong asthe lengthBDis to BC. BC Drawthe horizontal line_A_F extend untilit cuts this and DB [229] lineat F; letFE bea meanproportional between and FB; DF drawEOparallelto DC;thenAOwillbe a meanproportional between AandAB. If nowwerepresent C thetimeoffallalong AB

198 THE TWO NEW SCIENCES GALII.EO OF ABbythelength thenthe timeofdescent long A_B, a FBwillbe represented ythedistance soalsothetimeoffallthrough b FB; the entiredistance willbe represented y the meanproAC b portional O:and for the entiredistance by FE. Hence A FD the timeoffallalongthe remainder, BC,willbe representedy b A. "_!. BO,andthat alongtheremainder, BD, follows,f we allowthe bodiesto fall i firstalong and FB,or,whatis the AB samething,alongthe common stretch AB,thatthe sinceBE:BO--BD'_C, t a BD "byBE; but timesofdescent longthe and BC willbe to eachother as i But we p o lengths havepreviouslyroventhat BDandBC. thetimeofdescent, fromrestatB, along Fig. 4 6 BDis tothe timealong BCin the ratio whichthelengthBDbearsto BC. Hencethetimesofdescent along different lanes p ofconstant eightareto eachotherasthe h lengths oftheseplanes, whether themotion startsfromrestor is preceded ya fallfrom constant eight. b a h Q._. 9. THeOReM X_I,P_oPosmoN XI If a planebedivided intoanytwopartsandifmotion along it startsfromrest, thenthe timeofdescentalong first the part is to the timeof descentalongthe remainder "-_ as the lengthof this firstpart is to the excess fa o meanproportional between thisfirstpartandtheentire lengthoverthis firstpart. Let the fall take place,fromrest at A, throughtile entiredistance ABwhichis divided anypointC; also e at letAFbea meanproportional between theentirelength p BAandthefirstpartAC;thenCFwilldenote theexcess ofthemeanproportional FAoverthefirstpartAC. Now, I say,the timeof descent long a ACwillbetothe timeof B subsequent allthroughCB as the length f ACisto CF.Fig. 5 6 Thisis evident, ecausehetimealong b t ACis to the timealong the entiredistance ABasACis to themeanproportional AF. Therefore, _

THIRD DAY I99 Therefore, divldendo, time alongAC willbe to the time the alongthe remainder CBasACisto CF. If weagreeto representthe timealongACby the lengthACthenthe timealong CBwillberepresented byCF. Q._. D. In casethemotion isnotalong thestraightineACB l butalong thebroken lineACD thehorizonto A tal lineBD,andif fromF wedraw _ thehorizontal lineFE, itmayinlike mannerbe provedthat the time along ACisto thetimealong theinclined lineCDasACisto CE. For the time alongAC is to the time alongCB as AC is to CF; but it has alreadybeen shownthat the D_' timealong CB,afterthefallthrough Fig.66 thedistance C,isto thetimealong A CD,afterdescenthrough t the same distance AC, _ asCBisto CD,or,asCFisto CE;therefore, x cequali, timealongACwillbeto thetimealongCD e the asthelength ACisto thelengthCE. THEOg_XII, PRoPosmon XII If a verticalplaneand any inclined planearelimitedby twohorizontals, andifwetakemeanproportionals between the lengthsof theseplanesand thoseportionsof them which liebetween theirpointofintersecCtion andtheupper horizontal,hen the time of fall alongthe perpendicular t bearsto thetimerequiredo traverseheupperpartofthe t t perpendicular plusthe timerequired traversethe lower to part of the intersecCting planethe sameratio whichthe entirelengthof the vertical earsto a length b whichisthe sum of the mean proportional the verticalplus the on excessof the entirelengthof the inclined planeoverits meanproportional. LetAFandCDbetwohorizontal planeslimiting thevertical planeACandthe inclined lane letthe twolast-mentloned p DF; planesintersecCt B. LetARbe a meanproportional at between the

[ 3o]

2OO THE TWO NEW SCIENCESOF GALILEO the entireverticalACand its upperpart AB;andlet FSbe a meanproportional between and its upperpart FB. Then, FD I say,the timeoffallalongtheentireverticalpathACbearsto the timeoffallalongitsupperportionABplusthe timeoffall A_1_ alongthe lowerpart ofthe inclined plane,namely,BD, the sameratio whichthe lengthAC bearsto the meanproportional n the vertical, o namely, AR,plus thelength SDwhich over its meanproportional FS. Jointhe pointsR and S givinga is the excess f theentireplaneDF horizontalineRS. Nowsincethe l o D o c timeof fall throughthe entiredisFig.67 tanceAC is to the time alongthe portionAB as CAis to themeanproportional it follows AR that, if we agreeto representhe timeof fallthrough t ACby the distance AC,the timeof fallthrough distance Bwill the A berepresentedyAR;andthe timeof descenthroughthe reb t mainder, BC,willberepresentedyRC. But,if thetimealong b ACis takento be equalto the length AC,thenthe timealong FDwillbeequalto thedistance FD;andwemaylikewise infer that thetimeofdescent long whenpreceded ya fallalong a BE), b FB orAB,is numericallyqualto the distance e DS. Therefore thetime required fallalong to thepathACisequalto ARplus PC;while thetimeofdescentalong thebroken ABDwillbe line equaltoARplusSD. Q.E.D. The samethingis true if, in placeof a verticalplane,one takesanyotherplane,asfor instance O;the method N ofproof isalso thesame.
PROBLEM PROPOSITION I, XIII

Givena perpendicular lineof limitedlength, t is required i to finda planehavinga verticalheightequalto the given perpendicular so inclined and that a body,havingfallen fromrest alongthe perpendicular, willmake its descent along

THIRD DAY 2o: alongthe inclined planein the sametimewhichit occupiedinfalling through thegiven perpendicular. Let ABdenotethe givenperpendicular: prolong lineto this C makingBC equalto AB,and drawthe horizontal linesCE andAG. It isrequired drawa planefrom thehorizontal to Bto line CE suchthat aftera bodystartingfromrest at A has fallenthroughthe distance B,it willcomplete pathalong A its this planein anequaltime. LayoffCDequalto BC,and draw thelineBD. ConstructhelineBEequalto thesum t ofBDand DC;then, I say,BE is the required plane. Prolong tillit EB intersectsthe horizontal G at G. Let GF be a meanproA portional etweenGE and GB; b _ @ thenEF:FB=EG.K3F, _--_2: and -F--B2=E--O2: G--F2=EG: But GB. EG istwice GB;hence thesquare of EF istwicethe squareofFB;

the square square ofBC. Consequently soalsois the EF:FB =DB:BC, ndcomponendo a ofDBtwice_../fif/ et permutando, EB'J)B+BC= D 0 BF:BC. But EB=DB+ BC; Fig. 8 6 henceBF=BC=BA. If weagreethat thelength ABshallrepresentthetimeoffallalong thelineAB,thenGBwillrepresent the timeof descent longGB,andGFthetimealong a theentire distance GE; therefore willrepresentthe timeof descent BF along thedifferencef thesepaths,namely, E,afterfallfrom o B Got fromA. Q._. F. PRosn_-mPP.oPosiTIO_ If, XIV Given an inclinedplane and a perpendicular assing p throughit, to finda lengthonthe upperpart of the perpendicularhrough t whicha bodywillfallfromrest in the sametimewhichis required traverse inclined to the plane afterfallthrough thevertical istanceustdetermined. d j LetACbe the inclined laneandDB theperpendicular. t is p I requiredto findon Mc verticalAD a lengthwhichwillbe traversed

[23 ]

2oz THE TWO NEW SCIENCESOF GALILEO traversed body,falling bya fromrest,inthe same timewhich is neededby the samebodyto traversethe planeAC afterthe aforesaid fall. Drawthe horizontal layoffAE suchthat CB; BA+ zAC'.AC---AC'-A_E, offAR suchthat BA-_A_C= and lay EA-AkR. romR drawRX perpendicular DR;then,I say, F to X is the point sought. For sinceBAq-2AC'.AC=AC'_A_E, it follows,dividendo, that BA+AC-_AC=CE'akE. And since BA'_C=EA'A_R, have, componendo, AC'_A_C=ER: we BA+ RA. But BA+ AC'.AC=CE'AE, enceCE:EA=ER'_RA= h sum of the antecedents:um of the consequents=CR'aRE. s D ThusRE is seento bea meanproportionalbetween CRandRA. Mordover sinceit has been assumed that BA: XI RAC=EA'.AR, nd sinceby similartria followshat EA-_AR t =XA'JkR. Hence angleswe areequal. But =XA'a_-R, it A] / EA andXA have BA'JkC if weagree be represented ythe lengthRA,then b the timeoffallalong RCwillbe repre// that the timeof fallthroughRA shall sentedby the lengthRE whichis a !_ meanproportional between AandRC; R Fig. 69 likewise E willrepresenthe timeof A t descentalongACafterdescentalongRAor alongAN. But the timeof fallthroughXA is represented y the lengthXA, b whileRA represents timethroughRA. But it has been the shown XA andAE areequal. that O.F..F. PROBLE_ PRoPosrrIo_" III, XV Givena verticalline and a plane inclined it, it is reto quiredto finda lengthon the verticallinebelowitspoint ofinterseCtion willbe traversed thesametimeas which in the inclined lane,eachof thesemotions avingbeenprep h ceded a fallthroughthegivenverticalline. by LetABrepresent theverticalineandBCtheinclinedlane;it l p isrequired finda lengthontheperpendicular to belowitspoint ofinterseCtion, aftera fallfrom which Awillbetraversed the in same

THIRD DAY 203 sametimewhich isneeded forBCafteranidenticalallfrom f A. Drawthehorizontal AD,intersecCting theprolongation ofCBat D; letDE be a mean proportional between andDB;lay CD offBF equalto BE;alsoletAGbe a thirdproportional BA to andAF. Then, say,BGis thedistance hich body,after I w a falling through AB,willtraverse the in A same timewhich isneeded fortheplane "-'-'7 BC after the same preliminary fall. Forif weassume that the timeof fall alongAB is represented byAB,then/' thetimeforDBwillbe represented by DB. AndsinceDEis a meanproportionalbetween BDand DC,this same DE willrepresenthe timeof descent t alongtheentiredistance Cwhile D BE willrepresent thetimerequired forthe difference thesepaths,namely, C, of B provided eachcasethe fall is from in rest at D or at A. In likemannerwe mayinferthat BF represents time the Fig. 7 ofdescenthroughthe distance Gafterthesamepreliminary t B fall;but BF isequalto BE. Hencethe problem issolved. THEO_SM PROPOSITION XIII, XVI If a limitedinclined planeand a limitedverticallineare drawnfromthe samepoint,andif thetimerequired fora body,startingfromrest,to traverseeachof theseis the same, henabodyfalling anyhigher ltitude t from a willtraversetheinclined lanein lesstimethanisrequired p forthe verticaline. l Let EB be the verticalllneand CE the inclined plane,both startingfromthecommon ointE, andbothtraversed equal p in timesby a bodystartingfromrest at E; extendthe vertical line upwardsto any pointA, fromwhichfallingbodiesare allowedostart. Then,I saythat,afterthefallthrough t AE,the inclined planeEC willbe traversed lesstime thanthe perin pendicular

[233]

204 THE TWO NEW SCIENCESOF GALII.RO pendicular EB. JoinCB,drawthe horizontal AD,andprolong CEbackwards ntilit meetsthe latterinD; letDF bea mean u proportional betweenCD and DE whileAG is madea mean proportional betweenBAand AE. DrawFG and DG; then sincethetimesofdescentalongECandEB, startingfromrest at E, areequal,it follows, according Corollary ofProposito II tionVI that theangleat C isa Drightangle;but the angleatA is also a right angleand the anglesat the vertex E are equal;hence thetriangles ED A and CEBare equiangular and thesidesabouttheequalangles are proportional; ence BE: h EC= DE:EA. Consequently G the re&angle BE.EAis equal to the re&angleCE.ED;and c sincethe re&angle D.DEexC ceedsthe rectangle E.EDby C thesquare ofED,andsince the re&angle BA.AEexceedsthe rectangle E.EAbythesquare B ofEA,it followshattheexcess t of the re&angle CD.DEover the re&angle A_kE, r what B o isthe samething,theexcess f o the square of FD over the Fig.71 squareof AG,willbe equalto theexcess fthe square o ofDE overthesquare ofAE,whichexcessis equalto the squareof AD. ThereforeF-D* G--A_-t = -_)*=G--D*. enceDF is equalto DG, and the angleDGF H is equalto the angle DFGwhile angle the EGF islessthanthe angleEFG,and theopposite sideEF islessthan the opposite sideEG. If nowwe agreeto represent timeoffallthrough the AE bythelength AE,thenthetimealong DEwillberepresented byDE. Andsince isa meanproportional AG between BAand AE,

[234]

THIRD DAY 2o5 AE,it followshat AGwillrepresent t thetimeoffallthrough the totaldistance .B,andthe difference A EGwillrepresent time the offall,fromrest at A, through thedifference ofpathEB. In likemannerEF representshe timeofdescent longEC, t a startingfromrestatDorfalling fromrestatA. But ithasbeen shown thatEF islessthanEG;hence follows thetheorem. COROLLARY Fromthisand the preceding proposition, is clearthat the it verticaldistance covered a freelyfalling by body,aftera preliminary fall,and duringthe time-interval required traverse to an inclinedplane,is greaterthan the lengthof the inclined plane, but less than the distancetraversedon the inclined planeduringan equaltime,withoutanypreliminaryall. For f since havejust shown bodies we that fallingfroman elevated pointA willtraversethe planeECin Fig.7I in a shortertime thanthe verticalEB, it is evidentthat the distancealong EB whichwillbe traversed duringa timeequalto that ofdescent along ECwillbe lessthanthewhole ofEB. But nowin order to show that thisverticaldistance isgreaterthanthelengthof the inclined planeEC, we reproduce A I_ Fig.70 of the preceding theoremin ' / whichtheverticalengthBGistrayl ersedin the sametimeas BC aftera ./ preliminary throughAB. That fall / BG is greaterthan BC is shownas /" follows:inceBE and FB are equal _ s _ F whileBA is lessthan BD,it follows _" " thatFBwillbeartoBAa greater ratio than EB bearsto BD; and,componendo,FA willbear to BA a greater ratiothan El) to DB; but FA'.AB -@ GF:FB (since is a meanproporAF Fig. 2 7 tionalbetween BAandAG)andin likcmanner D:BE) CE: E = EB. HenceGBbearsto BF a greaterratiothanCBbearsto BE;therefore B isgreaterthanBC. G

[23s]

_6

THE TWO NEW SCIENCESOF GALILEO

PROBLEM PROPOSITIONVII IV, X

Givena verticallineand an inclined plane,it is required to layoffa distance long a thegiven planewhich betravwill ersedby a body,afterfallalongthe perpendicular, the in sametime-interval hichis neededfor this bodyto fall w fromrestthrough thegiven perpendicular. LetABbe theverticallineandBE the inclined lane. The p problemis to determine n BE a distancesuchthat a body, o __ Dafterfallingthrough AB,willtraverseit . l_/ B_// theperpendicular ABitself, tartingfrom s rest. ina timeequaltothat requiredo traverse t planeuntil it meetsthis linein D. Lay Drawthehorizontal ADandextend the

E suchthat BD:FD=DF'.DE. Then,I say,the timeof descentalongBE, after _/ offFB equalto BA;and choose thepoint fallthrough AB,isequaltothetimeoffall, Fig.73 fromrest at A, through AB. For,if we assume that the lengthABrepresentshe timeof fallthrough t AB,thenthetimeoffallthrough DBwillberepresented the by timeDB;and sinceBD:FD=DF:DE,it followshat DF will t represent he timeof descentalongthe entireplaneDE while t BF representshe timethroughthe portionBE startingfrom t restat D; but the timeof descentalongBEafterthe preliminary descentalong is the sameasthat after a preliminary DB fallthrough AB. Hencethetimeofdescent longBEafterAB a willbeBF whichof course isequalto the timeof fallthrough ABfromrestatA/ O.E.F.

[zs6]
PROBLEM V, PROPOSITION X-VIII

Given distance which the through abodyill vertically w {all from during rest a given tlme-lnterval, a|so andg_ven a smaller time-interval, to it s i required another locate [equal] vertical

THIRD DAY 2o7 verticaldistance whichthe bodywilltraverseduringthis given smallerime-interval. t Let the verticallinebe drawnthrough andon this linelay A, offthe distance ABwhichistraversed a bodyfallingfrom by rest at A, duringa timewhichmay alsobe A represented AB. Draw the horizontal ./Ix by lineCBE,andonit lay off BCto represent _ [\ the givenintervalof timewhichis shorter (" _ BI \ C thanAB. It is requiredto locate,in the \ perpendicular abovementioned, distance a whichis equalto ABand whichwillbedescribedna timeequaloBC. Jointhepoints i t AandC;then,since BC<BA,t followshat x i t the angleBAC<angle CA. ConstrucCt B the \_ angleCAEequalto BCAand let E be the x x point here w AEintersectshehorizontal t line; _x_ drawEl) at rightangles AE,cutting to the vertical tD; layoffDFequaloBA. Then, a t I say,thatFD is that portion ofthevertical Fig.74 which bodystarting a fromrestatAwilltraverse uring d theassignedtime-lnterval C. For, if in the right-angled B triangle AEDa perpendicular bedrawnfromthe right-angletE to the a opposite sideAD,thenAEwillbea meanproportional between DAandABwhileBEwillbe a meanproportional between D B and BA,or between andAB (seeing FA isequalto FA that DB);and sinceit hasbeenagreed representhetimeof fall to t through ABbythe distance B,itfollowshat AE,orEC,will A t representhetimeoffallthrough t theentiredistance D,while A EBwillrepresenthe timethrough t AF. Consequently rethe mainder Cwillrepresent timeof fallthrough B the theremainingdistanceFD. Q.E.F.

[_371 PROm,EM VI,PRoPosrrro_ XIX

Giventhedistance through whicha bodyfallsina vertical llnefromrest andgivenalsothe timeoffall,it isrequired to findthetimeinwhich same the bodywill,later,traverse
an

zo8 THE TWO NEW SCIENCESOF GALIT._O an equaldistancechosen anywhere the samevertical in line. OntheverticalineAB,layoffACequaltothedistance l fallen fromrest at A, alsolocateat random equaldistance an DB. ._ Let the timeof fallthroughACbe represented ythe lengthAC.It is b .Cto traverseDB afterfallfromrest at A. Aboutthe entirelengthAB describe semicircle EB;from required findthe timenecessary the to A C drawCE perpendicularo AB; t AE willbe longerthanEC; layoff EF equalto EC. Then,I say,the difference FAwillrepresent thetime requiredfor fallthroughDB. For Bsince is a meanproportional AE between BAandACandsince ACrepjoin the pointsA and E; the AC, resentsthetimeoffallthroughline

thetimethrough theentiredistance AB. AndsinceCE is a meanproC_ it followsthat AE willrepresent portional etween b DAandAC (see1_ ing that DA=BC) it follows that Fig.75 CE, that is,EF, willrepresent he t timeoffallthrough AD. Hence difference.A_F the willrepresent the timeoffallthroughthe difference B. D Q._. D. COROLI ARY Henceit is inferred if the timeof fallfromrestthrough that any givendistanceis represented y that distance b itself,then thetimeoffall,afterthegivendistance hasbeenincreasedy a b certainamount, illbe represented y the excess f the mean w b o proportional betweenthe increased distanceand the original distanceoverthe meanproportional between original isthe d tanceand the increment.Thus,for instance,f weagreethat i AB

THIRD DAY 2o9 ABrepresentshetimeof fall,fromrestat A, through dist the tanceAB,and that ASis the increment, thetimerequiredto traverse_AB, afterfallthrough willbetheexcess fthe SA, o meanproportional betweenSBand BAoverthe mean proportional between BAandAS.
PROBLEMVII, PROPOSITION

[z:38]

"<

Given distance and any whatever aportion laid of it offfrom the pointat whichmotionbegins,t isrei quiredto findanother ortion p which liesat theother endofthedistance ndwhich a istraversed inthesame timeasthefirstgiven portion. Fig. 76 Let the givendistancebe CBand letCDbe that part ofit c whichis laid off fromthe beginning f motion.It is o requiredto find anotherpart, at the end B, whichis traversedin the same time as the assigned portion D CD. Let BA be a mean proportional etweenBC b and CD; also let CE be a third proportionalo BC t and CA. Then,I say,EB willbe the distancewhich, after fall fromC, willbe traversedin the sametime as CD itself. For if we agree that CB shallrepresentthe timethroughthe entiredistance CB,thenBA ._ (which, f course, s a meanproportional o i between BC and CD)willrepresenthe time alongCD; and since t CAis a meanproportional between BCand CE,it followsthat CA willbe the time throughCE; but the total lengthCB represents time throughthe total the distanceCB. Therefore difference Awillbe the the B Fig.77timealongthe difference distances, afterfalling of EB, fromC; but this sameBAwasthe timeof fallthroughCD. Consequently thedistances DandEBaretraversed, C fromrest at A,in equaltimes, q._..F.

THEO_XIV, RoPosmo_ P XXI

If, onthe pathof a bodyfalling vertically fromrest,one laysoffa portion whichis traversed inanytimeyouplease and

.zto THE TWO NEW.SCIFNCESOF GALTT,_O andwhose upperterminus coincides withthe pointwhere the motionbegins, nd if thisfallis followed y a motion a b deflegtedlong a anyinclined lane, henthespace p t traversed along theinclinedlane, uring time-interval p d a equaltothat occupied the previous in verticalfall,willbe greaterthan twice,and lessthanthreetimes,the lengthofthe vertical fall. LetABbe a verticalinedrawn l downwards fromthehorizontal lineAE,and letit representhepathofa bodyfalling t from rest at A; choose portionACof this path. ThroughC any drawany inclined plane,CG,alongwhichthe motionis continuedafterfallthroughAC. Then,I say,that the distance traversedalongthis planeCG,duringthe time-interval equal to that of the fall through AC, is morethantwice,but less A B thanthreetimes, his same t A _ CF equal to AC,and extendthe planeGC untilit meetsthe horizontalnE; i EF =EF:EG. If nowwe assumethat the time of distanceACisrepresented A fallalong C. Letuslayoff bythelength AC,thenCE willrepresent he time of t !_ chooseG suchCE, while descentalong that CE: CF, or CA,willrepresent Fig. 8 7 the timeof descentalong CG. It nowremains be shown to thatthedistance CGismore than twice,and lessthan threetimes,the distanceCAitself. SinceCE:EF=EF_G, it follows CE:EF=CF:FG;but that EC<EF; therefore willbe lessthan FG and GC willbe CF morethantwiceFC,orAC. Againsince FE<_EC(forEC is greaterthanCA,or CF),wehaveGFlessthantwiceFC, and alsoGClessthanthreetimesCF,or CA. Q.E.D. Thisproposition maybe statedina moregeneral orm;since f what p c_///_ G

[539]

THIRD DAY 211 whathasbeen proven thecase for ofavertical andinclined plane holds equally inthe well casefmotion o along aplane of anyinclination by motion followed along anyplane ofless steepness, beseenrom adjoining Themethod as can f the figure. ofproofthe is same. [_o] PRo_._VIII,RoPosmo_ P XXII Given twounequal time-intervals, also distance the through whicha bodywillfallalonga verticalline,fromrest, during the shorterofthese intervals,it is requiredto passthrough the highest point of this vertical line a plane so inclined that the time of descentalongit willbe equalto the longer of the givenintervals. Let A representthe longerand B the shorterof the twounequal time-intervals,also let CD represent the length of the

J 'IS Fig.79 vertical fall,fromrest, duringthe time ]3. It is requiredto pass throughthe point C a planeof such a slopethat it willbe traversedin the time A. Draw _romthepoint C to thehorizontala lineCX of sucha length that B:A =CD:CX. It is clear that CX is the plane alongwhich a body will descendin the giventime A. For it has been shownthat the time of descentalongan inclinedplane bears to the time of fall through its vertical heightthe same ratio whichthe lengthof the plane bearsto its vertical height. Thereforethe time alongCX is to the time alongCD as the lengthCX is to the lengthCD, that is, as the time-intervalA tO is

212 THE TWONEW SCIENCESOF GALII.E0 to thetime-interval B:butB isthetimerequired traversehe to t vertical istance, d CD,startingfrom rest;therefore isthe time A requiredordescent long f a theplane CX. PROBT _MIX, PROPOSmON XXlII Given thetimeemployedya bodyinfalling b through cera tain distancealonga verticaline, it is requiredto pass l through lowerterminus fthis vertical all,a planeso the o f inclined this bodywill,afterits vertical all,traverse that f on this plane,duringa time-interval equalto thatof the verticalfall,a distance equalto anyassignedistance, rod p videdthisassigned istance d ismorethantwice andlessthan threetimes,theverticalfall. Let AS be any verticalline,and let AC denoteboth the lengthof the verticalfall,fromrest at A, and alsothe time _ r _ _ _. required this fall. Let for IR bea distance orethan m '-IA _._'l_,twice and less than three [_ times,AC. It is required _ to passa planethrough the _' [ pointC so inclined that a - " [ body,afterfallthrough AC, will,duringthe time AC, traverse distance qualto a e ,1,._ IR. Lay offRN andNM Fig.8o eachequaltoAC. Through thepointC,drawa plane CEmeeting thehorizontal, atsuch A.E, a pointthat IM:MN--AC:CE.Extendthe planeto O, and layoffCF,FG andGOequalto RN,NM, andMI respectively. Then,I say, the timealongthe inclined planeCO,afterfall through AC,isequalto thetimeoffall,fromrestat A,through AC. For since OG:GF=FC:CE,it follows,componendo, that OF:FG_-OF:FC=FE:EC, sincean antecedentis and to its consequent asthesumof theantecedents the sumof isto the consequents, have OE'.EF=EF:EC. Thus EF is a we meanproportional betweenOE and EC. Havingagreedto represent

[2411

THIRD DAY 213 represent thetimeoffallthrough ACbythelength ACit follows that ECwillrepresent thetimealong EC,andEF thetimealong theentiredistance O,while E thedifference CFwillrepresent the time alongthe difference O; but CF=CA; thereforethe C problem issolved.ForthetimeCAisthetimeoffall,fromrest at A, through CAwhileCF (which equalto CA)is thetime is requiredto traverseCOafter descent longEC or afterfall a throughAC. Q.E.F. It is to be remarked alsothat the samesolution oldsif the h antecedent motion takesplace, otalong vertical, n a butalong an inclined plane. This caseis illustrated the following in figure wherethe antecedent otionis alongthe inclined m planeAS [242] underneath horizontal Theproofisidentical iththe the AE. w preceding. SCHOLIUM Oncareful ttention, t willbeclearthat,thenearerthegiven a i lineIR approaches threetimesthe length to AC,thenearer the I M l_ l_.
I ' '| I i ,I

Fig. I 8 inclinedplane, CO, alongwhich the secondmotiontakes place,approacheshe perpendicular longwhich the space t a traversed, uring timeAC,willbe threetimesthe distance d the AC. Fori/IR be takennearly equalto threetimesAC,then IM willbe almostequalto MN; and since, y constrnd_don, b

/a

IM:

2i4 THE TWONEW SCIENCESOF GALII.EO IM-.MN =AC:CE, t followshat CE is but littlegreaterthan i t CA:consequently thepointE willlienearthepointA,andthe linesCOand CS,forming veryacuteangle, illalmost oina w c cide. But,ontheotherhand,if thegivenline,IR, beonlythe leastbit longerthantwiceAC,the lineIM willbeveryshort; fromwhichit followshat ACwillbeverysmalln comparison t i withCEwhichis nowsolongthat it almostcoincides withthe horizontal linedrawnthrough Hencewe caninferthat, if, C. afterdescent long a theinclinedlaneACoftheadjoiningigure, p f themotioniscontinued along horizontal a line,suchasCT,the distancetraversed a body,duringa timeequalto the time by offallthrough AC,willbeexactly twicethe distance C. The A argument ereemployed h isthe same asthepreceding.Forit is clear,sinceOE:EF=EF'.EC,that FC measures time of the descent long a CO. But, ifthehorizontal lineTCwhichistwice as longas CA,be dividedintotwoequalpartsat V thenthis linemustbe extended indefinitely the direction in ofX before it will intersectthe line AE produced;and accordinglyhe t ratioofthe infinite lengthTX to theinfinite lengthVX isthe sameas the ratioof the infinitedistanceVX to the infinite distance CX. The sameresultmaybe obtained anothermethodof apby proach, amely, yreturning n b tothesameline ofargument which wasemployed the proofof the first proposition.Let us in considerhe triangle t ABC,which, ylinesdrawnparallelo its b t base,representsorusa velocity f increasingnproportion the i to time;if theselinesareinfinite number, ust as the pointsin in j the lineAC are infiniteor as the numberof instantsin any interval ftimeisinfinite,heywillformtheareaofthetriangle. o t Letus nowsuppose themaximum that velocity attained--that representedy thelineBC--tobe continued, b withoutaccelerationandatconstant aluethrough v anothernterval ftimeequal i o to thefirst. Fromthesevelocities illbe builtup, in a similar w manner,the area of the parallelogram ADBC,whichis twice that of the triangle ABC;accordinglyhe distancetraversed t withthesevelocities uringany givenintervaloftimewillbe d twice

[z431

THIRD DAY zI5 twice that traversed with the velocitiesrepresentedby the triangleduringan equalintervalof time. But alonga horizontal plane the motion is uniformsince here it experiences neither accelerationnor retardation; thereforewe con-l). A cludethat thedistanceCDtraversedduringatime-_ ; interval equal to AC is twice the distanceAC; _ .for the latter is coveredby a motion, starting : : from rest and increasingin speed in proportion : to the parallel lines in the triangle, whilethe -: former is traversed by a motion representedby L: the parallel lines of the parallelogramwhich, beingalsoinfinitein number,yieldan area twice_ _ that of the triangle. Fig.82 Furthermorewemay remarkthat anyvelocityonceimparted to a moving body will be rigidlymaintained as long as the external causesof accelerationor retardation are removed,a conditionwhich is foundonly on horizontalplanes;for in the caseof planeswhichslopedownwardshere is alreadypresent a t causeof acceleration,whileon planes slopingupward there is retardation;from this it followsthat motionalonga horizontal plane is perpetual; for, if the velocitybe uniform,it cannotbe diminishedor slackened,much less destroyed. Further, although any velocitywhicha bodymay have acquiredthrough natural fall is permanentlymaintainedso far as its ownnature [suaptenaturalis concerned,yet it must be remembered that if, after descent along a plane inclineddownwards,the body is deflectedto a plane inclined upward,there is alreadyexistingin this latter plane a causeof retardation;for in any such plane this samebody is subjecCt a natural acceleration to downwards. Accordinglywe have here the superpositionof two different states, namely,the velocityacquiredduringthe precedingfall whichif actingalonewouldcarry the bodyat a uniformrate to infinity, and the velocitywhichresultsfroma natural acceleration downwardscommonto all bodies. It seemsaltogether reasonable,therefore,ifwewish to tracethe future history ofa body whichhas descendedalongsomeinclinedplane and has been deflected along someplane inclinedupwards,for us to
a3$11ine

2x6 THE TWO NEW SCIENCESOF GALII.EO assumethat the maximumspeedacquiredduringdescentis permanentlymaintainedduringthe ascent. In the ascent, however,there supervenes naturalinclination a downwards, namely, motionwhich, tartingfromrest,is accelerated the a s at usualrate. If perhapsthis discussion a littleobscure,the is following willhelpto makeitclearer. figure Let us suppose that the descenthas beenmade alongthc downwardloping s planeAB,fromwhichthe bodyisdefle_ed so as to continueits motionalongthe upwardsloping plane BC;andfirstlettheseplanes beofequallength andplaced soas to makeequalangleswith the horizontal lineGH. Nowit is wellknown a body,starting that fromrestatA, anddescending alongAB,acquires speedwhichis proportional the tlme, a to _ _, A. which maximum isa .... at B, and whichis ........ "" maintainedby the causesof fresh acceleration orretardabody so longas all I4. tion are removed; Fig.83 the acceleration to whichI refer is that to whichthe bodywouldbe subje_ if its motionwere continued alongthe planeAB extended, whilethe retardationis that whichthe bodywouldencounter if its motionweredeflected alongthe planeBC inclined upwards;but, upon the horizontal lane GH, the bodywould p maintaina uniformvelocityequalto that whichit had acquiredat B afterfall fromA;moreoverhis velocityis such t that, duringan intervalof time equalto the time of descent through AB,the bodywilltraversea horizontal distance equal to twice AB. Nowletus imaginehis same t bodyto movewith the sameuniform speedalongthe planeBC so that herealso duringa time-interval equalto that ofdescent long a AB,it will traversealongBC extendeda distancetwiceAB; but let us supposehat, at theveryinstant .]_e t _ bodybeginsts ascentit is i subje6ted, y its very--nature_ the same influences b m which surrounded

[z44]

THIRD DAY 217 surroundedt duringits descentfromA along i _A_B, namely,it descends fromrestunderthesameacceleration asthatwhich was effecdve AB, and it traverses, uringan equalintervalof in d time,the samedistance alongthis second laneas it didalong p AB;it isclearthat,bythussuperposing uponthebodyauniform motionofascent ndanaccelerated otion a m ofdescent,twillbe i carried along theplane BCasfarasthepointCwhere thesetwo velocities becomequal. e If nowwe assume anytwopointsD and E, equally distant {tomthe vertex]3,we maythen inferthat the descentalong BDtakesplace thesametimeastheascent long in a BE. Draw DF parallelto BC;weknowthat, afterdescent long a AD,the bodywillascendalongDF; or, if,on reaching , the bodyis D carriedalong horizontal it willreach withthe same the DE, E momentum [impetus] withwhichit left D; hencefromE the bodywillascendasfar asC,proving the velocity E is that at thesame asthat atD. Fromthiswemaylogically inferthat a bodywhichdescends along inclined laneandcontinuests motion any p i along plane a inclined upwards ill,on account f the momentum w o acquired, ascendto an equalheightabovethe horizontal; sothat if the descent is along D _ A E ABthe bodywill -x \ / / be carriedup the _ _ / / planeBCasfaras thehorizontal line ACD:and this is true whetherthe B inclinations ofthe Fig.84 planesare the sameor different, s in the caseof the planes a AB and BD. But by a previous postulate [p.184]the speeds acquiredby fall alongvariouslyinclinedplaneshavingthe sameverticalheightare the same. If therefore planes the EB and BD have the sameslope,he descentalongEBwill t be ableto drive the bodyalong BDas far as D; and since this propulsion comesfromthe speedacquiredon reaching the

[ 4s]

218 THE TWO NEW SCIENCESOF GALILI_O the pointB,it followshat this speedat B isthesamewhether t thebodyhasmadeitsdescent long a ABorEB. Evidently then the bodywillbe carriedup BDwhether descenthas been the madealongABor alongEB. Thetimeof ascentalongBD is however reaterthan that alongBC,just as the descentalong g EB occupies oretime than that alongAB;moreovert has m i beendemonstrated the ratiobetween that thelengthsofthese timesisthesameasthat between thelengths ftheplanes.We o must next discover what ratio existsbetweenthe distances traversed equaltimesalongplanesofdh/erent in slope,butof the sameelevation,hat is, alongplaneswhichare included t betweenthe sameparallelhorizontal ines. This is doneas l follows:
THEOREM XV, PROPOSITION

Given parallel planes avertical two horizontaland line connecting them;givenalsoan inclined lanepassing p through thelowerextremity ofthisverticalline;then,if a bodyfall freelyalongthe verticallineandhaveits motionreflected alongthe inclined lane,the distance hichit willtraverse p w alongthis plane,duringa timeequalto that oftheverticalfall,isgreater thanoncebut lessthantwicethevertical line. Let BC andHG be the twohorizontal planes,connecCted by theperpendicular alsoletEB representhe inclined .A.E; t plane B A C
i , |1

1_ G Fig.85 alongwhichthe motiontakesplaceafterthe bodyhas fallen along and hasbeenreflecCted AE fromE towards Then,I B. say,that,duringa timeequalto that offallalong AE,thebody will ascendthe inclinedplane througha distancewhich is greater

THIRD DAY 219 greaterhanAEbut lessthantwiceAE. LayoffEl) equalto t AE and choose so that EB:BD=BD-S3F. F Firstwe shall showthat F is the point to whichthe moving bodywillbe carried afterrefle_ion fromE towards during timeequalto B a that offallalong AE;andnextweshallshow thedistance that EF isgreater thanEAbutless thantwice thatquantity. Let us agreeto representhe timeof fallalong t AEby the lengthAE,thenthe timeofdescentalong orwhatis the BE, samething,ascent long a EBwillberepresentedythedistance b EB. Now,since is a meanproportional DB between andBF, EB and sinceBEis thetimeofdescentfortheentiredistance E, B it followshat BDwillbethetimeofdescenthrough t t BF,while theremainder DEwillbethetimeofdescent long a theremainder FE. But thetimeofdescent long a thefallfromrestat Bisthe sameasthe timeofascentfromE to F afterreflectionromE f withthe speedacquired duringfalleitherthrough AEorBE. Therefore representshe timeoccupied the body in DE t by passing fromE to F, afterfallfromAto E andafterrefie_ion alongEB. But by construgfion D is equalto AE. This E concludes thefirstpartofourdemonstration. Nowsincethe wholeof EB is to the wholeof BD as the portionDB is to the portionBF,wehavethe whole ofEB is to the wholeof BD asthe remainder l) is to the remainder E DF; but EB>BD and henceED>DF, and EF is lessthan twiceDE or AE. Q.E.D. Thesameis truewhenthe initialmotionoccurs, otalong n a perpendicular, uponan inclined lane:the proofisalsothe but p same provided theupward slopinglaneislesssteep, .e.,longer, p i thanthedownwardloping s plane. THEOREM XVI,P_OPOSmON XXV If descentalongany inclined laneis followedy motion p b alonga horizontal lane,the timeofdescentalong inp the dinedplanebearsto the timerequired traverse asto any signed lengthofthe horizontal planethe sameratiowhich twice

[245]

zzo THE TWO NEW SCIENCESOF GALILEO twicethe lengthof the inclined planebearsto the given horizontal length. Let CBbeanyhorizontal lineandABan inclined lane;after p descentalong ABlet themotioncontinuehrough assigned t the A horizontal distance BD. Then, ' ABbearsto the _imespentin traversing the same ratio BD I say,thetimeof bearstoaBD. c, D, z . B whichtwiceAB descent long Fig.86 For,layoff; equalto twice BC ABthenit follows, froma previous roposition, thetimeof p that descentalong ABisequalto thetimerequired traverseBC; to but the timealongBC is to the timealongDB as the length CBis to the lengthBD. Hencethe timeofdescentalong AB [z47] isto the timealong astwicethe distance BD ABisto the distanceBD. q. _. D.
PROBLEM PROPOSITION X, XX-VI

Givena vertical eight oining h j twohorizontal parallelines; l givenalsoa distance greaterthanonceand lessthantwice this verticalheight, t is required passthroughthefoot i to ofthegivenperpendicular aninclined lanesuchthat,after p fall throughthe givenverticalheight,a bodywhosemotionis defle6ted alongthe planewilltraversetheassigned distance timeequalto thetimeofvertical all. ina f Let AB be the verticaldistanceseparating two parallel horizontal AO andBC;alsoletFE begreater lines thanonceand lessthantwiceBA. Theproblem passa planethrough isto B, extending the upperhorizontal ine,and suchthat a body, to l afterhavingfallenfromA to B, will,if its motionbedeflecCted alongthe inclined plane,traversea distanceequalto EF in a timeequalto that offallalong AB. Lay off; D equalto AB; E then the remainder F willbe lessthanAB sincethe entire D length islessthantwicethisquantity;alsolayoff; I equal EF D to DF, and choosethe point X such that EI:ID=DF:FX; fromB,drawtheplaneBOequalin length EX. Then,I say, to that

THIRD DAY 221 thatthe plane BOis theonealong which, fterfallthrough a AB, a bodywilltraverseheassigned istance E ina timeequalto t d F the timeof fallthrough AB. LayoffBR andRSequalto ED and DF respectively; sinceEI:ID=DF:FX, we have, then componendo, El) =DX-Y_=ED-J3F =EX.ZZD =BO:OR = o A

Fig.87 RO:OS.If we representhe timeof fall alongAB by the t lengthAB,thenOBwillrepresenthe timeof descentalong t [248] OB,andRO willstand for the timealongOS,whilethe remainder Rwillrepresent B thetimerequired bodystarting fora fromrest atO to traverse remaining istance B. But the the d S timeofdescentalong SBstartingfromrestat Ois equalto the timeofascentfromB to S afterfallthrough AB. Hence BOis that plane,passing throughB, alongwhicha body,afterfall through AB,willtraverse thedistance S,equalto theassigned B distance in thetime-interval EF, BRorB_A. Q.E.F.

XVlI, oPosiTio P XXVlI

If a bodydescends longtwoinclined a planesofdifferent lengths ut ofthesamevertical eight, hedistance b h t which it willtraverse,nthelower artofthelonger lane,during i p p a time-interval qualto that of descent verthe shorter e o plane,is equalto the lengthof the shorterplaneplusa portionof it to whichthe shorterplanebearsthe same ratiowhichthe longerplanebearsto the excessof the longer vertheshorter o plane. Let AC be the longerplane,AB,the shorter,andAD the common elevation; the lowerpartofAClay offCE equal on
to

222 THE TWO NEW SCIENCESOF GALII,EO to AB. ChooseF such that CA'akE =CA:CA-AB CE'.EF. = Then,I say,that FCisthat distance hichwill,afterfallfrom w A,betraversed duringa time-interval equalto that required for A descentalongAB. For since that the remainder A: the E remainder AF = CA:AE. Therefore E is a meanproA porfionalbetweenAC and AF. Accordingly follows ifthelength _*_ CA'.AE =CE'.EF,it DABis employed measure to Fig. 8 8 thetimeoffallalong AB,then the distance ACwillmeasure thetimeofdescentthrough AC; but the timeof descenthrough t AFismeasured the length by AE, and that throughFC by EC. NowEC=AB;andhence follows theproposition. [249] PROBLEM PROPOSITION XI, XXVIII LetAGbeanyhorizontal linetouching circle;etABbethe a l diameter assing p through pointof contagt; ndletAE and the a EB represent twochords. The problemis to determine any whatratiothetime offallthrough A G AB bears to the time of descent ,._-.,. ' / overbothAEandEB. ExtendBE till it meetsthe tangentat G, and drawAF so as to bisegtthe angle _'\".../ : BAE.Then,I say,thetimethrough ABisto thesumof the timesalong AEand EB asthe length isto AE the sumofthe lengths AEandEF. For sincethe angleFABis equal to the angleFAE,whilethe angle .B EAGis equalto the angleABF it Fig. 9 8 follows thattheentireangle GAFisequaltothesumoftheangles FABandABF. But theangleGFAisalsoequalto thesumof thesetwoangles. Hencethe lengthGF is equalto the length GA

THIRD DAY 2z3 GA;and sincethe re6tangle G.GEis equalto the squareof B GA, it willalsobe equalto the squareofGF, or BG:GF = GF._E. If nowwe agreeto representhe time of descent t alpngAEby the length AE,thenthelength willrepresent GE the timeofdescentalong GE,while GFwillstandforthe time of descentthroughthe entiredistanceGB; so also EF will denotethe timethrough afterfallfromGor fromAalong EB AE. Consequently timealongAE,orAB,is to the time the along AEandEBasthelength AEistoAE+EF. Q.E.D. A shorter methodis to layoffGFequaltoGA,thusmaking GF a meanproportional between and GE. The rest of BG theproofisasabove. THEOm_M XVIII, PROPOSmON XXIX Given a limitedhorizontaline,at one end ofwhichis l erecteda limited verticallinewhose length equalo oneis t haKthe given horizontal line;thena body,falling through this givenheightand havingits motiondeflecCted intoa horizontal irecCtion, d willtraverse given the horizontal distance andverticaline l in less time than it o willany otherverti-A cal distanceplusthe N givenhorizontal isd _/ LetBCbethegiven dis-_ c tance in a horizontal plane;at the endB erecCt a perpendicular, onwhich 'c. ....'D.. lay offBA equalto half Fig. 9 BC. Then,I say,that the timerequired a body,starting for fromrest at A, to traversethe twodistances, and BC,is AB the leastof allpossible timesin whichthis samedistance BC togetherwith a verticalportion, hethergreateror lessthan w AB,canbetraversed. LayoffEB greaterthanAB,as in the firstfigure,and less than

[25ol

L ,

2z THE TWO NEW SCIENCESOF GALILEO 4 than.A_B, in the second.It must be shownthat the time as required traversethe distanceEB plusBC is greaterthan to that requiredfor ABplusBC. Let us agreethat the length ABshallrepresent timealong the AB,thenthetimeoccupiedn i traversing horizontal ortionBC willalsobe AB, seeing the p that BC=2AB;consequentlyhe timerequiredfor bothAB t andBCwillbe twiceAB. Choosehe point0 suchthat EB: t B0 =BO:BA,henB0 willrepresent he time of fall through t t EB. Againlay offthe horizontal istance equalto twice d BD BE; whence is clearthat BOrepresentshe time along it t BD afterfallthroughEB. SeleCt pointN suchthat DB:BC-a EB:BA =OB_N. Now sincethe horizontal otionis unim formandsince0B is the timeoccupiedn traversing i BD,after fallfromE, it followshat NBwillbe the timealongBCafter t fallthrough same the heightEB. Henceit isclear 0B plus that BN represents timeof traversing plusBC;and, since the EB twiceBAisthe timealong ABplusBCo remainso be shown it t that OB+BN>2BA. But sinceEB'.BO=B0_A,it followsthat EB_A--0--Bh _--A_. Moreoverince s EB_A =OB :BNit follows thatOB:BN= 0-B_:B--ABut 0B:BN--(OB:BA)(BA:BN), therefore _. and AB:BN=OB :BA,that is,BAis a meanproportional between B0 and BiN. ConsequentlyB+BN>2BA. 0 q. _. D. THEOIU_M PRoPosmoNXX X]TK, X A perpendicular let fallfromany pointin a horizontal is line;it is requiredto passthroughany otherpointin this samehorizontal planewhich linea shallcuttheperpendicularandalong which bodywilldescendotheperpendicular a t intheshortest ossibleime. Sucha planewillcutfromthe p t perpendicular portionequalto the distanceof the asa sumedpoint in the horizontalromthe upperend of the f perpendicular. LetACbeanyhorizontal lineandBanypointinitfromwhich is droppedthe verticallineBE). Choose point C in the any horizontal ine and l_iyoff,on the Vertical,the distanceBE l equal

[2s ]

THIRD DAY 225 equalto BC; joinC and E. Then,I say,that of allinclined planesthat canbepassed through C,cutting theperpendicular, CE isthat onealongwhich descento the perpendicular the t is accomplished the shortesttime. For,drawthe planeCF in cuttingthevertical bove and a E, theplaneCGcutting thevertical below anddrawIK,a parallel E; verticalline,touching C a cirat x cledescribed ithBCas radius. w LetEK bedrawn parallelo CF,_ t c_ and extendedto meetthe tangent,aftercutting thecircle L. at r, Nowit is clearthat the timeof fallalong LEisequalto thetime alongCE; but the time along KE is greaterthan alongLE; therefore time alongKE is the greater thanalong CE. But the G[ time alongKE is equalto the timealongCF, sincetheyhave the samelengthand the same slope; nd,in likemanner,tfola i lowsthat the planes CGandIE, havingthe samelengthandthe Fig. I 9 sameslope, illbe traversed w inequaltimes. Also,sinceHE< IE, the time alongHE willbe lessthan the time alongIE. Therefore lsothetimealong a CE(equal o the timealong t HE), willbeshorter thanthetimealong IE. Q._.D.

.I

THFoU PgoPosmoN XX, XXXI

If a straightlineisinclined t anyangleto thehorizontal a and if,fromanyassigned ointinthehorizontal, planeof p a quickestdescentis to be drawnto the inclined line,that planewillbe the one whichbisecCts anglecontained the [ 52] between linesdrawnfromthe givenpoint,one pertwo pendicular

zz6 THE TWO NEW SCIENCESOF GALILEO pendicularo the horizontal ine,the otherperpendicular t l to theinclinedine. l Let CDbe a lineinclined any angleto the horizontal at AB; andfromany assigned pointA in the horizontal drawACperpendicularo A_B, nd AE perpendicular CD; drawFA so t a to asto biseCthe angleCAE. Then,I say,that ofallthe planes t whichcanbe drawnthrough pointA, cuttingthe lineCD the at any pointswhatsoever AF is the oneof quickest descent[in quo tempore omnium brevissimo defiat _ scensus].Draw FG parallelto AE; the alternate _ GFAand FAEwill _.be angles also the angle equal; EAF isequalto the angle FAG. Thereforehe sides t GFandGAof thetriangle / FGAare equal. Accordinglyifwedescribe circle a A li_ aboutGascenter, ithGA w as radius,this circlewill Fig. _ 9 passthroughthe pointF, and willtouchthe horizontal t the pointA and the inclined a lineat F; forGFCisa rightangle,since GFandAEareparallel. It iscleartherefore alllinesdrawnfromA to the inclined that line,with the singleexception FA, willextendbeyond of the circumference thecircle, of thusrequiring more timetotraverse anyofthemthanisneededforFA. q.E.D. LEMMA If two circlesone lyingwithinthe otherare in contaCt, and if any straightline be drawntangentto the inner circle,cuttingthe outercircle,and if threelinesbe drawn fromthe pointat whichthe circles arein contaCto three t pointsonthe tangential traightline,namely,hepointof s t tangency the innercircle the twopointswherethe on and straight

THIRD DAY 227 straightlineextendedcutsthe outercircle,then these three lineswillcontainequalanglesat the pointof contacCt. Let the two circlestouch eachother at the point A, the center of the smallerbeingat B, the centerof the largerat C. Draw the straight lineFG touchingthe innercircleat H, and cutting the outer at the points F and G; alsodraw the three linesAF, All, and AG. Then, I say, the anglescontainedby theselines, FAIl and GAll,are equal. ProX longAH to the circumferencet a I; from the centersof the circles, draw BH and CI; jointhe centers / B and C and extendthe lineuntil _ {_ it reachesthe point of contacCt r-N<, at A and cuts the circles at the -\ points O and N. But nowthe _ H linesBH and CI are parallel,becausethe anglesICN and HBO are equal, each being twice the angleIAN. AndsinceBH,drawn from the center to the point of Fig.93 contacCtsperpendicularto FG,it followsthat CI will alsobe i perpendicularo FG and that the arc FI is equalto the arc IG; t consequently angleFAI is equalto the angle LAG.Q.E.D. the

[2s3l

XXI, PRoPosmo XXXlI

If in a horizontalline any two points are chosen and if throughoneofthesepointsa linebedrawninclined towards the other, and if from this other point a straight line is drawn to the inclinedline in sucha direcCtion it cuts that off from the inclinedline a portion equal to the distance betweenthe two chosenpointson the horizontalline,then the timeofdescentalongthelineso drawnislessthan along any other straight linedrawnfrom the samepoint to the same inclinedline. Alongother lineswhich make equal anglesonoppositesidesof thisline,the timesofdescentare the same. Let

2z8 THE TWO NEW SCIENCESOF GALII.KO LetA andB be anytwopointsona horizontaline:through l B drawan inclined straightlineBC,and fromB layoffa distanceBD equalto BA;jointhe pointsA andD. Then,I say, the timeof descentalong islessthan alonganyotherline AD drawnfrom to the inclined A lineBC. Fromthe pointA draw AE perpendicular BA;and fromthe pointD drawDE perto pendicularo BD,intersecting E at E. Sincein the isosceles t A _k triangleABD,we have the anglesBADandBDAequal, [254] theircomplements DAEand EDA are equal. Hence if, with E as centerandEA as radius, e describe circleit w a willpassthrough andwill D touchthe linesBA and BD at the points AandD. Now since AistheendoftheverticallineAE,thedescent long a ADwilloccupy lesstimethan Fig. 4 9 alongany other line drawn fromthe extremity to the lineBCand extendingeyond A b the circumference circle; hichconcludes ofthe w thefirstpartofthe proposition. If however, weprolong theperpendicular lineAE,andchoose anypointF uponit, aboutwhichascenter,wedescribe circle a of radiusFA, this circle,AGC,willcut the tangentlinein the pointsGandC. Drawthe lines AGandACwhich willaccordingto the preceding lemma,deviateby equalanglesfromthe medianlineAD. The time of descentalongeitherof these linesisthe same, incetheystartfromthe highest ointA, and s p terminate onthecircurrfference circle ofthe AGC. PROBT.vM PRoPosrrlo_ XXIII XII, X Givena limited verticallineandan inclined laneofequal p height,havinga common upperterminal;it is required tofinda pointonthe verticalline,extended upwards, from which

THIRD DAY 229 which bodywillfalland,when a deflectedlong a theinclined plane,willtraverseit in the sametime-interval whichis requiredorfall,fromrest,through f thegiven vertical eight. h Let AB be the givenlimitedverticalline and AC an indinedplanehavingthe samealtitude. It isrequired findon to thevertical A,extended B above pointfromwhich falling A,a a bodywilltraversethe distance ACin the sametimewhichis spentin falling,romrest at A, through given f the verticaline l AB. Drawthe lineDCEat rightangles AC,andlayoffCD to equalto AB;alsojointhepoints andD; thentheangle A ADC willbe greater thantheangleCAD,since thesideCAisgreater thaneitherABorCD. Maketheangle DAEequalto theangle [z55] ADE,anddrawEF perpendicularAE; thenEF willcutthe to inclinedplane, ex[ / tended bothways, t a It F. Lay off AI and AGeach equaltoCF; A throughG drawthe xlg _ //[

ho zontal GH.K'line

Then,Isay, H isthe / _ / " / [ pointsought. _ _ / / [ For,if we agree to _ Cx_/ [ [B let the length AB _ _/_ [ representhetimeof V t _ / fall alongthe verfi- J_. _ / calAB,thenACwill/ Not likewise represent the timeofdescent romx f restat A, alongAC; Fig. 5 9 andsince, ntheright-angled i triangleAEF, thelineEChasbeen drawnfromtherightangle perpendicular atE tothebaseAF,it follows AEwillbeameanproportional that between AandAC, F while CEwillbe a meanproportional between CandCF,that A isbetween CAand AI. Now,since ACrepresentshe timeof t descentfromA alongAC,it followshat AEwillbe the time t along entiredistance F, andECthe timealong the A AI. But since

/I

230 THE TWO NEW SCIENCESOF GALILEO sincein the isoscelesriangle t AEDthe sideEA is equalto the sideED it followshat ED willrepresenthe timeoffallalong t t AF,while isthe timeof fallaIong EC AI. Therefore D,that C is A.B,willrepresent he timeoffall,fromrest at A, alongIF; t which isthesameassaying thatABisthetimeoffall,fromGor fromH, alongAC. E.F. PROBLEMXIII, PROPOSITION i_ Givena limitedinclined planeand a verticallinehaving their highestpoint in common,t is requiredto find a i pointin the verticallineextended suchthat a bodywill fallfromit andthentraverse theinclined lanein thesame p timewhichisrequired traverse inclined to the planealone startingfromrestat thetopofsaidplane. Let AC and AB be an inclinedplaneand a verticalline respeCtively, havinga common highestpointat A. It is requiredto finda pointin theverticalline,above suchthat a A, body,falling fromit andafterwards avingits motiondireCted h alongA_B, willtraverseboth the assigned of the vertical part [256] lineand the planeAB in the sametimewhichis required for theplaneABalone,startingfromrestat A. DrawBCa horizontallineand layoffANequalto .A_C; choose pointL so the that AB'.BN --AL:LC,andlayoffAI equalto AL; choose the pointE suchthat CE,laidoffonthe vertical Cproduced, A will be a thirdproportional ACandBI. Then,I say,CEis the to distancesought;so that, if the verticallineis extended above Aandif a portion islaidoffequalto CE,thena bodyfalling AX fromX willtraverseboth the distances, and AB,in the XA sametime as that required, henstartingfromA,to traverse w ABalone. DrawXR parallelto BC and intersecCting BAproducedin R; nextdrawEr) parallelto BCand meeting BAproduced in D; onAD as diameter describe semicircle; a fromB drawBF perpendicularo AD, and prolong till it meetsthe circumt it ferenceof the circle;evidentlyFB is a mean proportional between ABandBD,while FAisa meanproportional between DA

THIRD DAY 231 DAandAB. TakeBSequalto BI andFHequalto FB. Now sinceAB_D =AC:CEand sinceBF is a meanproportional between ABand BD,whileBI isa meanproportional between AC and CE, it follows that BA:AC=FB:BS, sinceBA: and AC=BA_N=FB_S we shall have, convertendo, BF-.FS= AB:BN =AL'.LC. Consequently rectangle the formed FB by

[ sT]

lg I

IJ

Fig. 6 9 and CL is equalto the recCtangle whosesidesare ALand SF; moreover, rectangle this _A_L.SF excess the recCtangle is the of AL.FB, rAI.BF,overtherecCtangle orM.IB. Butthe o AI.BS, recCtangle FB.LCisthe excess ftherectangle C.BF o A overthe re,angleAL.BF; ndmoreover a therecCtangle isequalto AC.BF the reccangle AB.BIsinceBA-_A_C=FB:BI; hencethe excess ofthe recCtangle AB.BIoverthe recCtangle AI.BF,orAI.FH,is equalto the excess ofthe rectangle I.FHoverthe rectangle A AI.IB;therefore twicethe recCtangle is equalto the sum M.FH of

23z THE TWO NEW SCIENCESOF GALII.F.O of the redtangles B.BI and AI.IB, or 2AI.FH=2AI.IB+ A B-I . Add_ to eachside,then_2AI.IB+BI +_ =_= _ 2 zAI.FH+AI_. Againadd B-F_ eachside,thenAB2q-BF_= to _ = 2AI.FH+ _2 + _-_ = zAI.FH+ __T+ F--H. But _ _ _* =zAH.HF+ _" + H-_'_;and hence zAI.FH+ _2 + F--H2= zAH.I-IF _2 + _i_. Subtracering from each + _ sidewe have 2AI.FH+_* =2AH.HF+AH Sincenow_FH _. is a fadtorcommon both re,angles, it follows AH is to that equalto AI; for if AH wereeithergreateror smallerhanAI, t thenthetworecCtangles AH.HFplusthesquare ofHAwould be eitherlargeror smallerhanthe tworecCtangles t AI.FHplusthe squareof IA, a resultwhichis contraryto whatwe havejust demonstrated. If nowwe agreeto represent he timeof descentalongAB t by the lengthAB,thenthe timethrough_AC willlikewise e b measured byAC;andIB,which meanproportional isa between AC and CE,willrepresent he timethroughCE, orX_A, t from rest at X. Now,since isa meanproportional AF between DA andAB,or between and AB,and since RB BF,whichisequal to FH, is a meanproportional between and BD, that is AB between and AR,it follows,roma preceding AB f proposition [Proposition XIX, corollary], the difference represents that AH the timeofdescentalong ABeitherfromrest at R or afterfall fromX, whilethe timeofdescentalongAB,fromrest at A, is measured the lengthAB. But as hasjust beenshown,he by t timeof fall through ismeasured IB, whilethe timeof XA by descent long a A_B, fterfall,through a RAorthrough XA,isIA. Thereforehe timeofdescentthrough plusABismeasured t XA by the lengthAB,which,of course, lsomeasures timeof a the descent, romrest at A, along f ABalone. Q.F..F.

[zSs]
PROBLEMXIV, PROPOSITION XX.XV

Givenninclined and a plane alimited line, revertical is it quired adistance inclinedwhich to find onthe plane abody, starting rest, traverse same from will in the time that as neededtraversethe ertical inclined to both v and the plane. Let

THIRD DAY 233 Let ABbe theverticallineandBCtheinclined plane. It is required layoffonBCa distance hich body,startingfrom to w a rest,willtraversein a timeequalto that which occupied y is b fallthrough thevertical Bandbydescent ftheplane. Draw A o thehorizontal AD,whichintersecCts theprolongation line at E of the inclined planeCB;layoffBF equalto BA,andaboutE as center,withEF as radiusdescribehe circle t FIG. Prolong FE untilit interse_sthe circumference G. Choose pointH at a suchthat GB'J3F =BH'_-I_F. Drawthe lineHI tangentto the
I" D

Fig. 7 9 circle I. AtB drawthelineBKperpendicular FC,cutting at to thelineEILat L;alsodrawLMperpendicular andcutting toEL BCatM. Then,I say,BMisthedistance hich body,startw a ingfromrest at B,willtraverse the sametimewhichis rein quiredto descend fromrest at A throughbothdistances, AB and BM. Lay off EN equal to EL; then sinceGB:BF-BH'.HF,we shall have, permutando, GB:BH=BF'MF,and, dividendo, H:BH--BH'_-I-F. G Consequently the recCtangle GH.HFisequalto the square onBH; but this samerectangle is alsoequalto the square onHI; therefore BHis equalto HI. Since,in the quadrilateralLBH,the sidesliB and HI are I equal

'4

z34 THE TWO NEW SCIENCESOF GALI!.EO equalandsince angles B andI arefightangles,t follows the at i that the sidesBLand LI arealsoequal:but EI --EF;therefore thetotal lengthT,E, rNE, is equalto the sumofLBandEF. o If wesubtracCt common artEF, the remainder N willbe the p F equalto LB: but, by construcdon, FB=BA and, therefore, LB=AB4-BN. If againweagreeto representhe timeof fall t through ABby the lengthA_B, thenthe timeofdescentalong EBwillbemeasured byEB;moreoverince is a meanpros EN portionalbetweenME and EB it willrepresentthe time of descentalongthe wholedistance EM; therefore difference the ofthesedistances, BM,willbetraversed, fterfallfromEB,or a AB,in a timewhichisrepresentedyBN. But havingalready b assumed thedistance Basa measure A ofthetimeoffallthrough AB,thetimeofdescent long a ABandBMismeasured yAB4b BN. Since EBmeasureshe timeof fall,fromrest at E, along t EB, the timefromrest at B alongBM willbe the meanproportional etween b BEand BM,namely,BL. The timethereA 13foreforthe pathAB+ BM,startingfromrest at A is AB4-BN;but thetimeforBMalone, starting fromrestat 13, isBL; and sinceit has already been shown that BL= AB+ BN, theproposition follows. Anotherand shorter proofis the following: Fig.98 Let BCbetheinclined planeand BAthe vertical;at B draw a perpendicular EC, to extending both ways;lay off BH equal to the excessof it BE overBA;makethe angleHEL equalto the angleBHE; prolong ELuntilitcutsBKinL; atL drawLMperpendicular to EL andextendit till it meetsBC inM; then,I say,BM isthe portionof BC sought. For, sincethe angleMLE is a fight angle,BL willbe a meanproportional between and BE, MB while

[ s9]

THIRD DAY 235 while LEis a mean proportional between EandBE;layoff M EN equalo I,E;thenNE=EL=LH,andHB=NE-BL.But t alsoHB=NE-(NB+BA);therefore BN+BA=BL. If now we assumehelength asa measurefthetimeof descent t EB o along EB,thetimeofdescent, fromrestatB, along Mwillbe B represented byBL;but,ifthedescent along BMisfromrestat E oratA,thenthetimeofdescent bemeasured will byBN;and AB willmeasurehe timealongAB. Thereforehe timeret t quired traverse Band to A BM,namely, thesum ofthedistances ABandBN, is equalto thetimeof descent, fromrestat B, along BMalone. Q._. F. [26o] LEMMA LetDCbedrawn perpendicular thediameter A;from tO B the extremityB drawthe line BED at /k_ random;drawthe lineFB. Then,I say, FB is a mean proportional betweenDB and BE. Jointhe points E and F. Through B, draw the tangent Gwhich B willbeparallel to CD. Now,sincethe angleDBGis C equalo theangle t FDB,andsince the alternateangleof GBD is equal o t EFB, it followsthat the triangles FDBand FEBaresimilar andhence _i _s BD:BF=FB:BE. Fig.99 ' LEMMA Let ACbe a linewhichis longerhanDF, andletthe ratio t of ABto BCbe greaterthanthat ofDE to EF. Then,I say, AB is greaterthan DE. For, if AB _ . _ bearsto BCa ratiogreaterhanthatof t __ p DE toEF, thenDE willbearto some , , , _ length shorterthanEF, thesameratio Fig.ioo which ABbears toBC. Callthislength EG;thensinceAB:BC=DE=EG, follows, it componendo etconvertendo,

z36 THE TWO NEW SCIENCESOF GALILEO vertendo, CA'./kB that =GD'.DE.But sinceCAis greaterthan GD,it followshat BAisgreater t thanDE. #B LFAVIMA LetACIBbethe quadrantofa circle; romBdrawBEparallelo f t AC;about any point in the line $ BEdescribe circle a BOES,oucht ingABat B and intersecdng the circumference the quadrantat of I. JointhepointsC andB; draw the line CI, prolonging to S. it Then,I say,the lineCI is always lessthan CO. Drawthe lineAI touchingthe circleBOE. Then, [z6I] if the lineDI be drawn,it willbe equal to DB; but, since DB toucheshe quadrant, I willalso t D betangent o it andwillbeatright t angle_ AI; thusAI touches to the circleBOE at I. And sincethe angle AICisgreater thantheangle /k _,,. ABC,subtendingas it does a SIN isalsogreaterthanthe angle ABC. Whereforehe arc IES is t greaterthan the arc BO,and the line CS,beingnearerthe center, :: is longerhanCB. Consequently t CB--OC:CI. COis greaterthan CI, sinceSC: This result wouldbe all the Fig.ioi moremarkedif, as in the second figure,the arc RICwerelessthan a quadrant. For the perpendicular DBwould thencut thecircle CIB;andsoalsowould DI

1rtoo t t e

THIRD DAY 237 DI whichis equalto BD;the angleDIAwould beobtuseand thereforehelineAINwould t cutthecircle BIE. Since theangle ABCislessthanthe angle AIC,whichisequalto SIN,andstill lessthantheangle whichthetangentat I would makewiththe lineSI, it followshat the arcSEI is fargreaterthanthe arc t BO;whence, etc. Q.E.D. Ta_.ORSM XXII,PROPOSITION XXXVI If fromthe lowestpoint of a verticalcircle,a chordis drawnsubtending arc not greaterthan a quadrant, an and if fromthe twoendsof this chordtwootherchords be drawn anypointonthearc,thetimeofdescent long to a the twolatter chords willbe shorterthanalongthe first, and shorter also,bythe same amount,hanalong t thelower ofthesetwolatterchords. Let CBDbe an arc, notexceeding quadrant,takenfroma a verticalcirclewhoselowestpoint is C; let CD be the chord [planum elevatum]ub-m s B & tendingthis arc,andletwt" ' fA" .... ..A_' there be two otherI f//j, chordsdrawn from C/ _ {/ ]/. and D to any point B/ _[ /// onthearc. Then,I say,[ /o \/_/'-thetimeofdescent long / v_. a [ the two chordsplana][/_ [ DB and BC is shorter Id_--1" thanalongDCalone, r[ o along BCalone, tarting s from restatB. Through the point D, drawthe _" G _ l_ 1_ horizontalline MDA cuttingCB extended at Fig. lO2 A:drawDN andMC at rightangles MD,and BNat right to angles BD;aboutthe right-angled to triangle DBNdescribehe t semicircle FBN,cuttingDCat F. Choose D thepointO such that DOwillbea meanproportional between DandDF;inlike C manner

[262]

238 THE TWO NEW SCIENCESOF GALII.EO manner elecc s Vsothat AVisa meanproportional between CA and AB. Let thelength represent PS thetimeofdescentalong thewhole distance orBC,bothof whichrequirethe same DC time. LayoffPR suchthat CD'.DO =timePS.timePR. Then PRwillrepresent thetimeinwhich body,startingfromD,will a traversehedistance F,while t D RSwillmeasurehetimeinwhich t the remaining istance, C,willbe traversed.But since is d F PS alsothe time ofdescent,fromrest at B,alongBC,and if we choose suchthat BC:CD=PS:_ thenPT willmeasure T the timeofdescentromAto C,forwehavealready f shown [Lemma] that DC isa meanproportional between ACand CB. Finally choose thepointGsuchthatCA:AV _lYI'_G,thenPGwillbe the timeofdescent romA toB,whileGTwillbe the residual f timeofdescentalongBCfollowing descentfromAto B. But, sincethe diameter, DN,ofthe circle DFNisa verticalline,the chords F andDBwillbetraversed equaltimes;wherefore D in if one can provethat a bodywilltraverseBC, afterdescent along DB,in a shortertimethanit willFC afterdescent long a DF he willhaveprovedthe theorem. But a bodydescending fromD along DBwilltraverse BCin the sametimeasif it had comefromA along seeing thebodyacquireshe same AB, that t momentum descending longDB as alongAB. Henceit in a remains onlyto show descentalong that BCafterABisquicker thanalongFCafterDF. But wehavealreadyshown GT that representshe timealongBCafterAB;alsothat RSmeasures t the time alongFC afterDF. Accordingly mustbe shown it that RS is greaterthan GT, whichmaybe doneas follows: SinceSP_R =CD'.DO,it follows,nvertendot convertendo, i e that RS:SP=OC:CD; lso we have SP.'_=DC:CA. And a sinceTP:PG=CA'_AV, follows,nvertendo, PT:TG= it i that AC:CV, herefore,ex wquali, S:GT=OC:CV.But, as we t R shallpresently show,OCis greaterthan CV;hencethe time RSisgreaterthanthe timeGT,whichwasto be shown.Now, since[Lemma] CFis greaterthanCBandFD smallerhanBA, t it followsthat CD:DF>CA-.AB.But CD'J)F=CO.)F, seeingthat CDff)O =DO'.DF; nd CA'dkB_*:_". Therea = fore

[2631

THIRD DAY z39 foreCO)F>CV:VB, and,according the preceding to lemma, CO>CV. Besideshisit is clearthatthetimeofdescentalong t DC is to the time alongDBCas DOCis to the sumof DO and CV. SCHOLIUM Fromthe preceding is possible inferthat the path of it to quickest descent[lationem omniumvelocissimam] one from pointto anotheris not the shortestpath, namely,a straight line,but the arc ofa circle.*In the quadrant AEC,having B the sideBC vertical,dividethe arc ACinto any numberof equalparts,AD,DE,EF, FG,GC,andfromC drawstraight linesto the pointsA, D, E, F, G; _ A drawalsothe straightlinesAD, DE, EF, FG,GC. Evidently descentalong thepathADC isquicker [264] D thanalongACaloneor along DC fromrestat D. But a body,startingfrom rest at A, willtraverse DC morequicklythan the path ADC;while,if it startsfromrest atA, itwilltraverse thepathDEC 1_ in a shortertimethanDC alone. C Hence descentalong the three Fig. lO3 chords, DEC,willtakelesstimethan along twochords A the ADC. Similarly, followingdescent alongADE, thetimerequired to traverse EFCislessthanthat needed forECalone. Thereforedescentis morerapidalong thefourchords DEFCthan A along threeADEC. Andfinally body,afterdescent long the a a ADEF,willtraversethe twochords, GC,morequickly F than FC alone. Therefore, alongthefivechords, ADEFGC, descent willbe morerapidthanalong thefour,ADEFC. Consequently
* It is wellknownthat the first correctsolutionfor the problemof quickestdescent,under the condition a constantforcewas givenby of John Bernoulli 667-I748). [Trans.] 0

z4o THE TWO NEW SCIENCESOF GALII.EO the nearerthe inscribedolygon p approaches circle a theshorter is thetimerequired fordescent rom f Ato C. Whathas been provenfor the'quadrantholdstrue alsofor smaller rcs;the reasoningsthesame. a i PROBLEM XV,PRoPosmos XXVII X Givena limitedverticalineandan inclined laneofequal l p altitude;it is requiredto finda distanceon the inclined plane whichis equal to the verticalline and whichis traversed an intervalequalto the timeoffallalongthe in verticalline. LetABbetheverticalineandACtheinclined lane. Wemust l p locate,on the inclined plane,a distanceequalto the vertical A_line AB and whichwill be from rest at A in the same timeneededforfallalongthe vertical line. Lay off AD equal to AB,and bisect the remainder C at I. Choose D traversedby a bodystarting 1_the pointE suchthat AC:CI Fig.Io4 --CI:AE and lay off DG equalto AE. Clearly isequaltoAD,andalsoto AB. And EG further,I say that EG isthat distance whichwillbe traversed by a body,startingfromrest at A, in the sametimewhichis requiredfor that bodyto fall throughthe distance AB. For since AC:CI--CI:AE--ID:DG,we have, convertecuto, CA: AI =DI:IG. Andsince wholeofCAis to the wholeofAI the as the portionCI is to the portionIG, it followshat the ret [z6S] mainderIAisto the remainder Gasthewhole fCAisto the A o wholeof AI. ThusAI is seento be a meanproportional betweenCAandAG,while CIisa meanproportional between CA andAE. Iftherefore thetimeoffallalong ABisrepresentedy b the lengthAB,the timealongACwillbe represented yAC, b whileCI,or ID,willmeasure thetimealong AE. Since isa AI mean proportional betweenCA and AG, and sinceCA is a measure

THIRD DAY 241 measure timelong entire AC,it ollows of the a the distance f that AI isthetimealong AG,andthedifference ICisthetimealong the difference _; but DI wasthe time alongAE. ConseC quently lengths I andICmeasurehetimesalong and the D t AE CGrespe_ively.Thereforehe remainder Arepresentshe t D t timealongEG,whichofcourse equalto thetimealong is AB.
Q. E. F.

COROT,T.ARY Fromthisit is clearthat thedistance sought s bounded i at eachend byportions theinclined lanewhich of p aretraversed inequaltimes. PRObLeM XVI,PRoPosrrloN XXXVlII Giventwohorizontal planescut by a verticalline,it is required finda pointonthe upperpartof the vertical to linefrom which bodies ayfalltothehorizontal m planes and there, havingtheir motiondefle_edinto a horizontal dire_ion, ill,duringan interval qualto thetimeoffall, w e traverse istances hichbearto eachotherany assigned d w ratioofa smalleruantity a larger. q to Let CD and BEbethe horizontal planescut by the vertical ACB,and let the ratio of the smaller quantityto the larger be that of N to FG. It is required findin the upper to part of thevertical ine,AB,a pointfromwhicha bodyfalling l to the planeCD andtherehavingitsmotiondefle_edalong this plane, illtraverse, w during interval qualto itstimeoffalla an e distance suchthat ifanother ody,falling b fromthissame point to the plane BE, there haveits motiondefle_edalongthis planeandcontinued during interval qualtoitstimeoffall, an e willtraversea distance which bearsto the former istance d the [266] ratioofFG tO N. LayoffGHequalto N, and sele_the point L sothatFHa_IG =BC:CL. Then,I say,L isthepointsought. For,if we lay offCM equalto twiceCL, and drawthe line L'V[ utting planeBEat O,then]30willbeequalto twice c the BL

.....

~ " ""

7 .....

24z THE TWO NEW SCIENCES GALIt,F.O OF BL. And since FH:HG=BC:CL,we have, componendo et convertendo, HG._F=N._F=CL.q,B=CM',BO. It is clear that, sinceCM is doublethe di_stance the spaceCM is LC, that whicha bodyfallingfromL throughLCwilltraversein the pIaneCD; and,forthe samereason, inceBOis twicethe s distanceBL,it is clearthat BOis the distancewhicha body, A L

: !f L t

'N -I

Fig. o5 I afterfallthroughLB,willtraverseduringanintervalequalto the timeof itsfallthrough LB. Q.v.F. SAox. ndeed,I thinkwemayconcede ourAcademician, I to withoutflattery,hisclaimthat in the principleprincipio, e., [ i. accelerated otion]aiddown thistreatisehehasestablished m l in a newscience dealing witha veryoldsubjecc. Observing ith w whateaseand dearnesshe deducesroma single f principle the proofsof so manytheorems,I wondernot a little howsuch a questionescapedthe attentionof Archimedes, Apollonius, Euclid and so many other mathematicians illustrious and philosophers, especially sinceso manyponderousomeshave t beendevoted thesubjedt fmotion. to o [_671 S_v. Thereisa fragment ofEudid whichtreatsofmotion, but

THIRD DAY z43 but init there isnoindication thatheeverbegan investigate to thepropertyf acceleration o andthemanner inwhich varies it withslope. Sothatwemaysaythedoorisnowopened, forthe firsttime,to a new method fraught withnumerous andwonderfulresults hichinfuture w years willcommand theattention of otherminds. SACg. reallybelievethat just as, for instance,he few I t propertiesf the circle o proven byEuclid intheThird Bookof hisElementseadto many l others morerecondite, the prinso ciples hicharesetforthin this littletreatise when w will, taken upbyspeculative minds, leadtomany another remarkable more result; nd istobebelieved a it thatitwillbesoonaccount ofthe nobility fthesubjec*c, issuperior o which toanyother innature. During longandlaborious this day,I haveenjoyed these simple theorems orethantheirproofs, anyof which, or m m f their completeomprehension, require thananhour c would more each;this study,if youwillbegoodenougho leave t thebook inmyhands, isonewhich mean takeupat myleisure fter I to a wehaveread theremaining portion which withthemotion deals of projedkiles; andthisif agreeable youweshalltakeuptoto
morrow.

SALV.shall otfailto bewithyou. I n

END OF THE THIRD DAY.

[26s1

FOURTH DAY
_ALVIATI. Oncemore,Simplicio hereon is ;l_]time; so let us withoutdelaytakeup the '_/_=_ isasfollows: i.__. question ofmotion.ThetextofourAuthor THE MOTIONOF PROJECrILES Intheprecedingages p wehavediscussed the properties f uniform o motionand of motionnaturallyaccelerated alongplanesof all inclinations.I nowproposeto set forththoseproperties hichbelong a bodywhose w to motionis compounded oftwoother motions, namely, neuniform o andone naturally accelerated; theseproperties, ellworthknowing, w I propose demonstraten a rigid to i manner.Thisis the kindof motionseenin a moving projectile; itsorigin conceiveo beas I t follows: Imagine anyparticleprojected along horizontal a planewithout friction;then we know,from what has been morefully explained the preceding in pages,that this particlewillmove alongthis sameplanewith a motionwhichis uniformand perpetual, rovided planehasnolimits. But if theplaneis p the limitedand elevated,henthemoving t particle, hich w weimagineto bea heavyone,willonpassing overtheedgeoftheplane acquire,in additionto its previousuniformand perpetual motion,a downward ropensity p dueto itsownweight; o that s the resulting motionwhichI callprojection [projectio],comis pounded ofonewhichis uniform andhorizontal ofanother and whichisvertical ndnaturallyaccelerated. enowproceed a W to demonstrate

FOURTHDAY 245 demonstrate some its properties, of thefirstofwhichis as follows: THEOREM I, PROPOSITION I Aprojectile hich w iscarried uniform bya horizontal motion compounded a naturally with acceleratederticalmotion v describes pathwhichis a seml-parabola. a SAGR. Here,Salviati,it willbe necessary stop a little to whilefor my sakeand,I believe, lsofor the benefit f Sima o plicio;for it so happensthat I havenotgoneveryfar in my studyofApollonius ammerelyawareof the fact that he and treatsofthe parabolaandotherconicsections, without n una .derstanding whichI hardlythinkonewillbe ableto follow of the proofof otherpropositions depending ponthem. Since u evenin thisfirstbeautifulheorem t theauthorfindsit necessary to provethat thepathofa projecCtile parabola, isa andsince, s a I imagine, shallhaveto dealwithonlythiskindofcurves, we it willbeabsolutely necessary tohavea thorough acquaintance, if notwithalltheproperties hich w Apollonius hasdemonstrated for thesefigures, t leastwiththosewhichare needed the a for presenttreatment. SALV. are quitetoo modest,pretending You ignorance f o facts whichnot longagoyou acknowledged aswellknown--I mean at the time whenwe werediscussing strengthof the materialsand neededto use a certaintheoremofApollonius which gaveyounotrouble. SACl_.mayhavechanced know ormaypossibly ave I to it h assumed solongasneeded, it, forthat discussion; butnowwhen wehavetofollow allthesedemonstrations aboutsuchcurves e w oughtnot,astheysay,toswallow whole, it andthuswastetime andenergy. SIMP. eventhough Now Sagredos,asI believe, equipped i well forallhisneeds, donotunderstandventheelementary I e terms; foralthoughour philosophers havetreatedthe motionof projedkiles, do not recalltheirhavingdescribed pathof a I the proje_ileexcept statein a general aythat it is always to w a curved

[269]

246 THE TWO NEW SCIENCES GATJLFD OF curved line,unlessthe projecCtion vertically be upwards.But if the littleEuclidwhichI havelearhed since ourprevious discussion doesnot enableme to understand demonstrations the whichareto follow, henI shallbe obliged acceptthe thet to orems onfaithwithoutfullycomprehending them. SALv. the contrary,I desire On that youshould understand themfromtheAuthorhimself, whenhe allowed eto see who, m this workof his,wasgoodenough provefor metwoof the to principal roperties ftheparabola p o because didnothappen I to haveat handthebooks ofApollonius. bzeseroperties, T p which arethe onlyoneswe shallneed in the presentdiscussion, he provedin sucha waythat noprerequisite nowledge asrek w quired. Thesetheorems are,indeed, ivenby Apollonius, g but aftermanypreceding ones,to follow whichwouldtakea long while. I wishto shorten ourtaskbyderiving thefirstproperty purelyandsimplyfrom modeofgenthe erationof the parabolaand proving the second immediately fromthefirst. Beginning owwiththe first,imagine n a right cone,ereftedupon the circular baseibkcwithapexat 1. The seedonf o this conemadebya planedrawn parallel tothe sidelk isthecurvewhichis called a parabola. he baseof thisparabola c T b ,i .]_cutsatrightanglesthediameterk ofthe i circle ibkc,and theaxisadis parallelto theside/k;nowhavingtakenanypointf in thecurvebfadrawthe straight inefe l Fig./o6 parallelto bd; then, I say, the square of bdis to the squareoffe in the sameratioas the axis ad is to the portion02. Throughthe point e passa planeparallel to thecircle ibkc,producing inthe conea circular e6tion s whose diameters thelinegeh. Since i bdisat rightangles ik in the to circle ibk,thesquareofbdisequalto the recCtangle by k/ formed and dk; so also in the uppercirclewhichpassesthroughthe pointsgfhthe squareoffe isequalto the re_Langle formed by ge

[7o]

FOURTH DAY z47 geand eh;hencethe squareof bdis to the square offe asthe teetangle id.dkisto thereccangle ge.eh.Andsince thelineedis parallelto hk, the lineeh,beingparallelto dk, is equalto it; therefore reftangleid.dkisto the recmngle as/d isto the ge.eh [2711 ge,that is,asdaisto a; hencelsotherectangle/d.dk the w a isto rectangle ge.eh, is,the square that ofbdis tothe square ofre,as the axisdaisto the portionae. Q._. D. Theotherpropositionecessaryorthisdiscussion n f wedemonstrateas follows.Letus drawa parabola hose w axiscaisprolonged upwardso a pointd;fromanypointbdrawthelinebe t parallelo thebaseoftheparabola; t ifnowthepointd ischosen so that da= ca, then, I say, the straight line drawn throughthe d pointsb and d willbe tangentto the parabola b. Forimagine, at if possible, thatthislinecutstheparabolaaboveor that its prolongationcutsitbelow, ndthrough a any pointg in it drawthe straightline fge. And sincethe squareoffe is a greaterthanthe squareofge,the square offewillbeara greateratio r to the square ofbcthanthesquare ofgeto thatofbc;andsince, bythe preceding proposition, square the offeisto that ofbcasthelineeais to ca,it follows that the lineea willbear to the line caa greater ratiothan the square ofgeto that of bc,or,than the squareofedto thatofcd(thesides ofthetriangles deg and dcb beingproportional). Fig. o7 x But the lineeaisto ca,orda,inthesame ratioasfourtimesthe reccangle is to fourtimesthesquare ea.ad ofad,or,whatisthe same,he square t ofcd,sincethisis fourtimesthesquareofad; hencefourtimesthe recmngle bearsto the squareof cd ea.ad a

248 THE TWONEW SCIENCESOF GALn,_o a greaterratiothan the squareofed to the squareof cd;but that wouldmakefour timesthe rectangle ea.adgreaterthan the squareof ed;whichis false,thefac't eingjust the oppob site,because two portions the eaand adof the lineed are not equal. Therefore line db touchesthe parabolawithout the cuttingit. Q.E.D. Sire,.Yourdemonstration proceedsoorapidly t and,it seems to me,you keeponassuminghat allofEuclid'stheorems re t a as familiar availableto me as his firstaxioms, hichis and w farfromtrue. Andnowthisfa_ whichyou spring uponus, that fourtimesthe rec_ngleea.adis lessthan the square of debecausethe two portions a and adof the linede arenot e equalbringsmelittlecomposure ofmind,but ratherleaves me in suspense. SALv. Indeed,allrealmathematicians assume onthe partof the readerperfecCt familiarity with at least the elements of Euclid;and hereit is necessary yourcaseonlyto recalla in proposition fthe Second o Bookin which heproves whena that lineis cut intoequaland alsointotwounequalparts,the rectangleformedon the unequalparts is lessthan that formed on theequal(i.e.,lessthanthe squareonhalfthe line),by an amountwhichis the square ofthe difference between theequal and unequal egments.Fromthis it is clearthat the squareof s the wholelinewhichis equalto fourtimesthe squareof the half is greaterthan four times the recCtangle the unequal of parts. In orderto understand following ortionsof this the p treatiseit willbe necessaryo keepin mindthe twoelemental t theoremsromconicsecCtions wehavejust demonstrated; f which and thesetwo theorems indeedthe only oneswhichthe are Authoruses. We can nowresumethe text and seehowhe demonstrates is first proposition whichhe showsthat a h in bodyfalling witha motioncompounded uniform ofa horizontal and a naturallyacceleratednaturale [ descendente] describes one a semi-parabola. Let us imagine elevated an horizontalineor planeabalong l whicha bodymoveswithuniform speedfroma to b. Suppose this

[272]

FOURTHDAY 249 t_s planeto endabruptly b;thenat thispointthebodywill, at on account f its weight, cquirealsoa naturalmotiondowno a wardsalong perpendicular/m. rawthe linebealongthe the D planebato representheflow,ormeasure, t oftime;divide this lineinto a numberof segments, c,cd,de,representingqual b e intervals ftime;from o thepointsb,c,d,e,letfalllines which are parallelto the perpendicular Onthe bn. d e_ _ d_ first of these lay off _ _ 0 anydistancei,onthe j_, second distance a c four /_ timesas[273] on / long,df; the third, one nine timesas long, h;and e soon,in proportion to thesquaresof cb,db, eb,or,wemaysay,in Fig. IO8 the squared ratioof thesesamelines. Accordingly weseethat while thebodymovesrombto cwithuniform f speed,t alsofalls i perpendicularly throughthe distance ci,andat the endof the time-interval bcfindsitselfatthepointi. In likemanner t the a endofthetime-interval which bd, isthedouble ofbc,thevertical fallwillbefourtimesthefirstdistance ci;forithasbeenshown in a previous discussion thedistance that traversed a freely by falling bodyvaries asthe square ofthetime;inlikemannerhe t spaceehtraversedduringthe time bewillbe ninetimesci; thus it is evidentthat the distances h,dr, ciwillbe to one e anotheras the squaresof the linesbe,bd,bc. Nowfromthe pointsi,f, h drawthe straightinesio,fg,hlparallelo be;these l t lineshl,fg, ioareequalto eb,dbandcb,respeCtively; soalsoare thelinesbo, g,blrespeCtively tod, dr,andeh. Thesquare b equal ofhlisto that offg asthelinelbisto bg;andthe square is offg to that ofioasgbis to bo;therefore points h,lieonone the _,f, andthe sameparabola. In likemannerit maybe shown that, ifwe takeequaltime-intervals ofany sizewhatever, andif we imagineheparticle t tobecarried similar ompound the bya c motion,

25o THE TWONEWSCIENCES F GALII,FO O thepositions ofthisparticle, theends fthese at o time-intervals, willlieononeandthesame parabola, q._. D. S_v. Thisconclusion follows fromthe converse first ofthe of the twopropositions above. For,having given drawn a parabola throughhepoints and t b h,anyother twopoints, and f i, notfalling ntheparabola lieeither ithin o must w orwithout; consequently fg iseitheronger theline l orshorterhantheline t which terminates ontheparabola.Therefore thesquarefM o willnotbear tothesquareffgthesame o ratioasthelinelbto bg, utagreater b orsmaller; thefactis,however, thatthesquare ofMdoes ear b thissame to the square ffg. Hence ratio o the pointfdoesieontheparabola, l andsodoalltheothers. SACX. Onecannot thattheargument deny isnew, ubtle nd s a conclusive, resting it does as uponthishypothesis, namely, that thehorizontal motion remains niform, thevertical u that motionontinues beaccelerated c to downwards inproportion to thesquare fthetime,andthatsuchmotions o andvelocities as thesecombine without ltering, a disturbing, hindering or each other,*othatasthemotion roceeds s p thepathoftheproje_ile does notchange intoa different curve: butthis,inmyopinion, [z74] is impossible. Forthe axisofthe parabola long a which we imagine thenatural motionfafallingody o b totakeplace stands perpendiculara horizontal to surfacendendsat thecenter f a o theearth;andsince theparabola eviates andmore d more from itsaxisnoproje_ile anever c reach thecenter ftheearthor,if o it does, sseems ecessary, a n thenthepathoftheproje_ile ust m transform intosome itself othercurve verydifferent fromthe parabola. Srm,. othese T difficulties, Imayaddothers.One ofthese is thatwesuppose horizontal which the plane, slopes either p n u nordown, toberepresented straight bya lineasifeach point n o thislinewereequally istant romthecenter, d f which isnotthe case;forasonestartsfromthemiddle [ofthe line]andgoes towardeitherend,he departsfartherand fartherfromthe center[ofthe earth]andisthereforeonstantlyoing c g uphill. Whence follows it that the motioncannotremain uniform *Avery approach near toNewton's Law Second ofMotion. [Trans.]

FOURTHDAY 25i through distance hatever, any w butmustcontinuallyiminish. d Besides, donotseehowit ispossibleoavoidtheresistance I t of themedium whichmustdestroy theuniformityfthehorizono tal motionand changehe lawof acceleration t offalling bodies. Thesevariousdifficulties render highlyimprobable a it that result erived d fromsuchunreliable hypotheses should holdtrue in pracCtice. SALv. thesedifficultiesnd objecCtions All a whichyou urge areso wellfounded it is impossible removehem;and, that to t asfor me,I am readyto admitthemall,which indeed think I ourAuthor would alsodo. I grantthattheseconclusions proved in the abstracCt willbe different henapplied the concrete w in andwillbefallacious tothisextent,hat neither illthehorizont w tal motionbe uniform the naturalacceleration in the nor be ratioassumed, or the path of the pro_ecCtile n a parabola, tc. e But,on theotherhand,I askyounottobegrudge urAuthor o thatwhichothereminent enhaveassumed m evenifnotstri_ly true. Theauthority-of Archimedes willsatisfy alone everybody. In hisMechanics andin hisfirstquadraturefthe parabola e o h takesforgranted that the beamofa balance orsteelyard a is straightline, everypoint of whichis equidistant romthe f common enter c ofallheavy bodies, ndthatthecords a bywhich heavy bodies aresuspendedreparalleloeachother. a t Someconsider assumptionermissible this p because,n praci tice,ourinstruments andthe distancesnvolved resosmallin i a comparison iththe enormous w distance fromthecenterofthe earththatwemayconsider minute farcona greatcirclesa a o a straightine,andmayregard l theperpendiculars letfallfromits twoextremities asparallel.Forifin accrual pracCtice onehadto consideruchsmallquantities, would necessaryirstofall s it be f tocriticise architecCts the whopresume, byuseofa plumbline, to ere_ hightowers withparallel ides. I mayadd that, in all s theirdiscussions, Archimedes andthe othersconsidered themselvesas locatedat an infinite distanceromthe center the f of earth,in whichcasetheirassumptions werenot false,and therefore theirconclusions wereabsolutelyorrecCt. c When' we wish

[ZTS]

zSz THE TWO NEW SCIENCESOF GALII.KO wishto apply ourproven conclusions distances hich, hough to w t finite,areverylarge,it isnecessary_for infer, nthebasisof usto o demonstrated truth, whatcorrection bemadefor the fact isto that our distancefromthe centerof the earth is not really infinite, ut merelyvery great in comparison ith the small b w dimensions our apparatus.Thelargestof thesewillbe the of rangeof our proje_iles--and evenherewe needconsider nly o the artillery--which, however reat,willneverexceed g fourof thosemilesof whichas manythousandseparateus fromthe centerof the earth;and sihcethesepathsterminateuponthe surface ofthe earthonlyvery slightchanges antake placein c their parabolic figurewhich,it is conceded, wouldbe greatly alteredif theyterminatedatthecenteroftheearth. As to the perturbationarisingfromthe resistance the of medium ismoreconsiderable this anddoesnot,onaccount fits o manifoldforms,submitto fixedlawsand exa_ description. Thusifweconsider onlytheresistance hich w theairofferso the t motions tudied us,weshallseethat itdisturbsthemalland s by disturbs themin aninfinite varietyofwayscorresponding the to infinitevarietyin the form,weight,and velocityof the projec_iles.Foras to velocity,he greaterthis is,the greaterwill t be the resistance offered the air; a resistance hichwillbe by w greateras the movingbodiesbecome dense[mengram]. less Sothat although thefalling bodyoughtto be displacedandare [ accelerandosz] in proportiono the squareof the durationofits t motion,yet no matterhowheavythe body,if it fallsfroma veryconsiderable height, heresistanceftheairwillbesuchas t o to preventany increase speedand willrenderthe motion in . [2761 uniform;and in proportion the movingbodyis lessdense as [mengrave]his uniformity t willbe so muchthe morequickly attained andaftera shorter fall. Evenhorizontal motion which, ifnoimpediment offered, were would beuniform andconstant s i alteredby theresistance f the air andfinallyceases; nd here o a againthe less dense[pig leggiero] bodythe quickerthe the process. Of theseproperties[accidentz] weight,of velocity, of andalsoofform[figura], infinite innumber, t isnot possibleo i t give

FOURTH DAY z53 giveany exaCtdescription; ence,inorderto handlethis matter h in a scientific way,it is necessaryto cut loose fromthesedifficulties; and havingdiscoveredand demonstratedthe theorems,in the caseof noresistance,to use themandapplythemwith such limitationsas experience willteach. And the advantageof this method will not be small; for the material and shape of the projectile may be chosen,as dense and round as possible,so that it willencounterthe least resistancein the medium. Nor willthe spacesand velocitiesin generalbe so great but that we shallbe easilyableto correCtthemwith precision. In the caseof thoseprojectileswhichwe use,made of dense [grave] material and round in shape,or of lightermaterial and cylindricalin shape, such as arrows,thrown from a slingor crossbow,the deviationfrom an exacCt parabolicpath is quite insensible. Indeed, if youwillallowme a little greaterliberty, I can showyou, by twoexperiments, hat the dimensions four t o apparatus are so small that these externaland incidentalresistances,amongwhich that of the mediumis the most considerable,arescarcelyobservable. I now proceedto the considerationof motionsthroughthe air, sinceit is with thesethat we are nowespeciallyconcerned; the resistanceof the air exhibits itselfin two ways: first by offeringgreater impedanceto less dense than to very dense bodies,and secondlyby offeringgreaterresistance a bodyin to rapid motionthan to the samebodyin slowmotion. Regarding the first of these, considerthe caseof two balls having the same dimensions, ut one weighingten or twelve b times as much as the other; one, say, of lead,the other of oak, both allowedto fallfroman elevation I5oor2oocubits. of Experimentshowsthat they willreach the earth with slight differencein speed,showing that in both casestheretardation us causedby the air is small;for if both balls start at the same moment and at the sameelevation,and if the leadenone be slightlyretarded and thewoodenonegreatlyretarded,thenthe former ought to reach the earth a considerabledistance in advance of the latter, sinceit is ten times as heavy. But this [z771 does

:z54 THE TWO NEW SCIENCESOFGALtt.F.O doesnot happen;indeed,the gainin distanceof oneoverthe otherdoesnot amountto the hundredth partofthe entirefall. Andinthe caseofa ballofstone weighingnlya thirdor halfas o muchasoneoflead,thedifference intheirtimesofreaching the earthwillbe scarcely noticeable.Nowsincethe speed[irnpeto] acquired leaden bya ballin falling froma height fzoocubits o is sogreatthat if themotionremained niform ballwould, n u the i anintervaloftimeequalto thatofthe fall,traverse 4oocubits, andsincethisspeedis soconsiderable comparison iththose in w which, y useofbows b orothermachines xcept irearms,weare e f ableto giveto our proje_iles,it followshat wemay,without t sensible error,regardasabsolutelyruethosepropositions t which weareaboutto provewithoutconsidering resistance fthe the o medium. Passing nowto the second case,wherewehaveto show that the resistance the air lcor rapidlymoving of a bodyisnot very muchgreaterthan foronemoving slowly, ampleproofis given by the followingxperiment.Attachto two threadsof equal e length--say four or five yards--twoequal leadenballs and suspendthemfromthe ceiling; owpullthemasidefromthe n perpendicular, one through8oor moredegrees, he other the t throughnot morethanfouror fivedegrees; o that, whenset s free, the one falls,passesthroughthe perpendicular, deand scribesargebut slowly l decreasingrcsof I6o,I5o, I4Odegrees, a etc.;theotherswinging throughsmallandalsoslowly diminishingarcsofIO,8,6, degrees, etc. In the first placeit must be remarked that one pendulum passesthrough arcsofI8o0,I6O0, its etc.,inthesametimethat theother swingshroughits Io0,8,etc.,fromwhich follows t it that the speedofthe firstball is I6 and I8 timesgreaterthan that ofthesecond.Accordingly, theairoffers oreresistance if m to thehighspeedthanto the low,the frequency fvibrationin o the largearcsof I8o or I6o0,etc.,oughttobe lessthan in the smallarcsof zoo, ,40,etc.,andevenlessthanin arcsofz0,or 8 I; but this prediction not verified experiment; is by becausef i twopersons startto countthevibrations,he onethe large,the t other the small,they willdiscoverthat after countingtens and

FOURTH DAY 255 andevenhundredsheywillnot differ ya single t b vibration, noteven byafra6tionfone. o [2781 Thisobservation justifies twofollowing the propositions, namely, thatvibrations ofverylarge andverysmall mplitude a alloccupy sametimeandthat the resistance the air the of doesnotaffect otionsfhighspeed m o more thanthoseoflow speed, ontrary the opinionitherto c to h generally entertained. SAoR. Onthe contrary, ince s wecannot enythat the air d hinders bothofthese motions, bothbecoming andfinally slower vanishing, wehaveto admithattheretardation t occursnthe i same proportion ineachcase. Buthow?How,ndeed, i could the resistance offered the onebodybe greaterhan that to t offeredotheotherexcept t bytheimpartation ofmore momentumandspeed[impetovelocitlz]thefastbodythanto the e to slow?And ifthisissothespeed withwhich abody movessat i once thecause andmeasurecagione [ emisura] ofthe resistance which meets.Therefore, it allmotions, fastor slow, rehina dered anddiminished inthesame proportion;result,t seems a i tome, fnosmallmportance. o i S_v. Weareable,therefore, inthissecond tosaythat case theerrors, egle_ing which n those areaccidental, intheresults which weareabout odemonstrate t aresmallnthecase i ofour machines where thevelocities employed aremostly verygreat and the distances egligible comparison the semin in with diameter oftheearth oroneofitsgreatcircles. SIMP.would I liketo hearyourreason forputting theprojec2iles offirearms,. e.,those i using powder, different ina class fromtheprojec2iles employed inbows, slings, andcrossbows, on the ground oftheirnotbeing equally subje6to change t and resistance fromtheair. S_v. I amledtothisview bytheexcessive and,soto speak, supernatural violence withwhich proje6t_iles such arelaunched; for,ndeed, i itappears tomethatwithout exaggeration onemight saythatthespeed ballfired ofa eitherfrom musket a orfrom apiece fordnance o issupernatural. Forifsuch ball eallowed a b to fallfromsome greatelevation speed its will,owing the to resistance

z56 THE TWO NEW SCIENCESOF GALII,KO resistance ftheair,notgoonincreasing o indefinitely; which that happensto bodiesof smalldensityin fallingthroughshort .distances--I meanthereducdon oftheirmotion uniformity-to willalsohappento a ballof ironorleadafterit hasfallena few thousand cubits;this terminal r finalspeed[termina_ o veloci_] isthe maximum whichsucha heavybodycannaturallyacquire [z79] in fallingthroughthe air. ThisspeedI estimateto be much smallerhanthat impressed ponthe ballby the burning t u powder. An appropriateexperiment iUserveto demonstrate w this faCt. Froma heightof onehundredor morecubitsfirea gun [archibuso] loadedwith a lead bullet,verticallydownwards upon a stonepavement;with the samegun shoot againsta similarstonefroma distanceofoneortwocubits,andobserve whichofthe twoballsis the moreflattened. Nowif the ball whichhas comefromthe greaterelevation foundto be the is lessflattenedof the two, this willshowthat the air has hinderedand diminishedhe speedinitially t impartedto the bullet by the powder, nd that the air willnot permita bulletto aca quiresogreata speed,nomatterfromwhatheightit falls;forif the speedimpressed ponthe ball by the firedoesnot exceed u that acquired byit infalling freely[naturalmente] thenits downwardblowoughtto begreaterratherthanless. Thisexperiment havenotperformed, amoftheopinion I butI that a musket-ball r cannon-shot, o fallingfroma height as greatasyouplease, illnot deliver w sostronga blowas it would iffiredintoa wallonlya fewcubitsdistant,i. e.,at sucha short range the splitting that orrending oftheairwillnotbesufficient torobtheshotofthat excess fsupernatural o violenceivenitby g thepowder. Theenormous omentum m [impel.o] oftheseviolentshotsmay cause some deformationfthetrajectory,making o thebeginning oftheparabolalatterandlesscurvedthantheend;but,sofaras f our Authoris concerned, is a matterof smallconsequence this in praCtical operations, themainoneofwhichisthepreparation of a tableof ranges shotsof highelevation, for givingthe distance

FOURTHDAY 257 tanceattainedby the ball as a funcdonf the angleof elevao tion;andsince shotsofthiskindarefiredfrommortars[mortart] usingsmallcharges andimparting nosupernatural momentum [impetoopranaturale] s theyfollowtheirprescribed paths very exacCtly. But nowlet us proceedwith the discussionn whichthe i Authorinvitesus to the studyand investigation ofthe motion ofa body[impetorel a mobile] whenthatmotioniscompounded of twoothers;andfirstthe caseinwhichthetwoareuniform,he t onehorizontal, theothervertical. THEOREMPROPOSITION II, II W'hen themotionofa bodyistheresultant ftwouniform o motions, onehorizontal, theotherperpendicular, thesquare of the resultantmomentumis equalto the sum of the squares ofthetwocomponent momenta.* Let us imagine anybodyurgedbytwouniform motions and let abrepresentthe verticaldisplacement, whilebcrepresents thedisplacement which, nthesameinterval i of time,takesplacein a horizontal irecd tion. If then the distances aband bcare traversed,duringthe sametime-interval, _ L with uniformmotionsthe corresponding Fig.lO9 momenta illbeto eachotherasthedistances bandbcare to w a eachother;but the bodywhichisurgedby thesetwomotions describeshediagonal c; its momentum t a isproportional ac. to Alsothe squareof acis equalto the sumof the squares ab of andbc. Hencethe squareoftheresultantmomentum equal is to thesumofthe squares ofthetwomomenta bandbc.Q.E.D. a Shay._At his point thereis just one slightdifficulty t which needsto be cleared for it seemsto me that the conclusion up; *Intheoriginal thistheorem asfollows: reads "Si aliquod duplici _equabili mobile motu moveatur, orizontali nempe et perpendiculari, seumomentum impetus lationis utroque comex motu po.ri_ potentia erit cequalis ambobus moment# priorumotuum." m Forthejustification ofthistranslation oftheword "potentia" and oftheuseoftheadjective "resultant" seep. 266below.[Trans.]

[280]

258 THE TWO NEW SCIENCESOF GALILEO just reachedcontradi_sa previous proposition in whichit is * claimed thespeed[impel.o] bodycoming that ofa froma to bis equalto that in comingfroma to c_whilenowyou conclude that thespeed[impao] tc isgreater a thanthat atb. SALV. propositions, Both Simplicio, true, yet therei8 a are great difference etweenthem. Here we are speakingof a b bodyurgedby a singlemotionwhichi8 the resultantof two uniform motions, hiletherewe werespeaking twobodies w of eachurgedwith naturallyaccelerated otions, ne alongthe m o verticalabthe other alongthe inclined planeac. Besides the time-intervals weretherenot supposed be equal,that along to the inclineac beinggreaterthan that alongthe verticalab; but the motionsof whichwe nowspeak,thosealongab,bc, ac,areuniform andsimultaneous. SIMP. Pardonme;I amsatisfied; praygoon. SALV. OurAuthornextundertakeso explain t whathappens whena bodyisurged a motioncompounded by ofonewhichi8 horizontal nd uniformand of anotherwhichis verticalbut a naturallyaccelerated; fromthesetwo components resultsthe path of a projecCtile, whichis a parabola. The problem to is determine speed[impeto] the projecCtile eachpoint. the of at Withthispurposein viewourAuthor setsforthasfollowshe t manner,orratherthemethod, fmeasuringuchspeed[impeto] o s alongthe pathwhichis takenby a heavybodystarting from restandfalling witha naturally accelerated motion.

[28 ]

Ta o= III,PgoPosi=o III
Let the motiontake placealongthe lineab, startingfrom restat a,and in this linechoose anypointc. Let acrepresent the time,or the measure the time,requiredforthe bodyto of fall throughthe spaceac; let ac also representthe velocity [impetus seumomentum] c acquiredby a fall throughthe at distance In thelineabsele_ anyotherpointb. Theprobac. lemnowis to determine velocityat b acquired a body the by in fallingthroughthe distance andto express ab thisin terms ofthe velocity themeasure fwhichisthelengthac. Take atc, o
See p. 16 above. [Trans.] 9

FOURTHDAY 259 as a meanproportional betweenac and ab. We shallprove that the velocity b is to that at c as the lengthas isto the at lengthac. Drawthe horizontal linecd,havingtwicethe length .a of ac,and be,havingtwicethe __ .... c lengthof ba. It then follows, from the precedingtheorems, s that a bodyfalling throughthe e [ distanceac, and turned so as _-.... j , [m to movealongthe horizontal cd Fig.IiO with a uniformspeedequalto that acquiredon reachingc [28z1 willtraversethe distancecd in the sameintervalof time as that required fallwithaccelerated otionfroma to c. Liketo m wisebewillbetraversed the sametimeas ha. But the time in of descentthroughabis as; hencethe horizontal istancebe d is alsotraversedin the timeas. Takea pointl suchthat the timeas is to the timeac asbeis to bl;sincethe motionalong beisuniform,the distancebl,if traversed withthe speed[momentum celeritatis] acquired b, willoccupythe timeac;but at in this sametime-interval, ac,thedistancecdis traversed with the speedacquired c. Nowtwospeeds in areto eachotheras the distancesraversed equalintervals time. Hencethe t in of speedat c is to the speedat b as cdis to bl. But sincedcisto beas theirhalves,namely,as cais to ha,and sincebeis to bl as baisto sa;it followshat dcisto blascaisto sa. In other t words, he speedat c isto that at bas caisto sa,that is,asthe t timeoffallthrough ab. Themethod ofmeasuring thespeedofa bodyalong thedirection of its fall is thus clear;the speedis assumed increase to direcCtly time. asthe But beforeweproceed further,sincethis discussion to is dealwiththe motioncompounded a uniform of horizontal one and oneaccelerated ertically v downwards--the of a propath je&ile, amely,a parabola--it snecessaryhat wedefine n i t some common standardby whichwe mayestimatethe velocity,or momentum [velocitatem, impetum momentum] both moseu of tions

z6o THE TWO NEW SCIENCESOF GALII.V.O tions;and sincefromthe innumerable niformvelocities ne u o only,andthat notsele&ed random,sto becompounded at i with a velocityacquiredby naturallyaccelerated motion,I can thinkofnosimpler ayofsele&ingndmeasuringhisthanto w a t assume anotherof the samekind.* For the sakeof clearness, drawthe verticalline_ to meetthe hor_ntal linebc..dc is the heightand bcthe amplitude f the semi-parabola o ab,which isthe resultantof the twomotions, nethat of a bodyfalling o from rest at a, throughthe distanceac, wlth naturallyaccelerated motion,he othera uniform t motionalong thehorizontal ad. The speedacquired c by a fall at .9throughthe distanceac is determined by the heightac;for the speedof a bodyfallingfromthe sameelevation alwaysone is andthe same;but alongthehorizontal ne o maygivea bodyan infinite number f unio _; _ a_ formspeeds. However, inorderthat I may rate it fromthe rest in a perfectly definite manner, willextend height aupwards I the c to e just as faras is necessary andwillcall this distance the "sublimity." Imagine ae a bodyto fallfromrest at e;it isclearthat we maymakeits this multitudeanda the [_ sele&one out of terminalspeedat sepa_: same as that with whichthe samebody Fig.ix1 travels alongthe horizontalline ad; this speedwillbe suchthat, in the timeofdescentalongea,it will describea horizontal istancetwicethe lengthof ea. This d preliminary remarkseems necessary. Thereaderisremindedhat above havecalled t I thehorizontal linecbthe "amplitude"of the semi-parabola the axisac ab; ofthis parabola, havecalled "altitude";but thelineeathe I its fall alongwhichdetermineshe horizontal peedI have called t s the "sublimity." These matters havingbeen explained,I proceed withthedemonstration. *Galileo proposes here toemploy asastandardfvelocity o theterminal speed fa body o fallingreelyrom given eight.[Trans.] f f a h

[283]

FOURTHDAY 26x SAGR. Allowme,please,to interruptin orderthat I may pointoutthe beautifulagreement between thoughtof the this Authorand the viewsof Plato concerninghe originof the t various niform u speeds ithwhichtheheavenly odies w b revolve. The latter chanced uponthe ideathat a bodycouldnot pass fromrest to anygivenspeedand maintain uniformly it except bypassing through allthedegrees fspeed o intermediate bet_veen thegivenspeedandrest. Platothoughtthat God,afterhaving createdthe heavenlybodies,assigned them the properand uniformspeeds withwhichtheywereforeverto revolve;and that He madethemstartfromrestandmoveoverdefinite distances undera natural and recCtilinear acceleration such as governs motionof terrestrialbodies. He addedthat once the thesebodies hadgained theirproperandpermanent speed,heir t redtilinear otionwasconverted m intoa circular ne,theonly o motioncapableofmaintaining niformity, motionin which u a the bodyrevolves ithouteitherrecedingromor approaching w f its desired goal. Thisconception istrulyworthyofPlato;and it isto be allthemorehighlyprized sinceitsunderlying rincip plesremained hiddenuntil discoveredy our Authorwhoreb moved fromthemthemaskandpoetical ressandsetforththe d ideain corre_historical erspe&ive.In viewofthe fa_ that p astronomical cience s furnishes suchcomplete us information concerning sizeof the planetaryorbits,the distances the of thesebodiesromtheircenters f ofrevolution, andtheirvelocities, I cannothelpthinkingthat our Author(towhomthis ideaof Platowasnotunknown) ad some h curiosityo discover hether t w or nota definite "sublimity"mightbe assignedo eachplanet, t suchthat, if it wereto startfromrestat thisparticularheight andtofallwithnaturally'accelerated motionalong straightline, a and werelaterto change speedthusacquired the intouniform motion,he sizeofitsorbitanditsperiod t ofrevolution ould w be thosea_uallyobserved.

[2841

madethe computationndfounda satisfactoryorrespondence a c withobservation.But he did notwishto speakof it, lestin view

SALv. I remember I think hishaving methatheonce told

z6z THE TWO NEW SCIENCESOF C_T,IT,I_O viewofthe odiumwhichhismanynewdiscoveries already had brought ponhim,this mightbe addingfuelto the fire. But u if anyonedesires suchinformation ecanobtainit forhimself h fromthetheorysetforthinthepresen_ treatment. Wenowproceed withthematterin hand,whichis to prove: PROBLEM I, PROPOSITION IV Todeterminehemomentum projecCtile t ofa ateachparticular pointin its givenparabolic ath. p Let beebe the semi-parabola whoseamplitudeis cd and whose heightisdb,whichlatterextended upwards utsthe tanc gentof the parabolaca in a. Throughthe vertexdrawthe horizontal inebi parallelto cd. Nowif the amplitudecd is l equalto the entireheightda, then bi willbe equalto baand alsoto bd;and if we take abas the measureof the time requiredfor fallthroughthe distanceaband alsoof themomentumacquired b in consequence fallfromrestat a, then at ofits if.we turn intoa horizontal irectionthe momentum d acquired by fallthrough ab[impetum ab]thespacetraversed the same in interval ftimewillberepresented ydcwhichistwicebi. But o b a bodywhichfallsfromrest at b alongthe lineM willduring the sametime-intervalallthroughthe heightof the parabola f [z851 bd. Hencea bodyfalling fromrest ata,turnedintoa horizontal directionwith the speedabwilltraversea spaceequalto dc. Nowifonesuperposes uponthismotiona fallalong bd,traversingthe heightbdwhilethe parabolabcis described, thenthe momentum ofthe bodyat the terminal ointc is the resultant p of a uniform horizontal omentum, m whose valueis represented by ab,and ofanothermomentum acquired fallfromb to the by terminalpointd or c;thesetwomomenta equal. If, thereare fore,we take abto be the measure one of thesemomenta, of say,the uniform horizontal ne, thenbf,whichis equalto bd, o willrepresent he momentum t acquiredat d or c; and ia will representhe resultantofthesetwomomenta,hat is,the total t t momentum withwhichthe proje_ile,travelling alongthe parabola, trikes c. s at

With

FOURTHDAY . z63 Withthis inmindletus take anypointontheparabola, ay s e, and determine momentum the with whichthe projectile passesthat point. Drawthe horizontal efand takebga mean proportional betweenbdand bf. Nowsinceab,or bd, is assumedtObe the measureof the _ time and of the momentum[tooAI rnentum velocitatis] acquired byfallJ/[ ingfromrest at b throughthedis// [ tance bd, it followsthat bgwill ,/ / I measurethe time and also the o ___.._.._g momentum [impetus] acquired atf J_-]. by fallfromb. If therefore elay w _ [7 offbo,equaltobg,the diagonal line _" ]q joininga and o willrepresent he C" / t q' momentum thepointe;because / at [d the lengthabhasbeenassumed to represent the momentumat b Fig.ii2 which, fterdiversion a intoa horizontal irection, d remains constant; and becausebomeasures momentum or e, acthe atf quiredby fall,fromrest at b,throughthe heightbf. But the square aoequalsthe sumofthe squares fabandbo. Hence of o the theoremsought. SAcg.The mannerin whichyou compoundhesedifferent t momentato obtaintheir resultantstrikesme as so novelthat my mind is left in no smallconfusion.I do not referto the compositionf twouniform o motions, venwhenunequal,and e when one takes placealonga horizontal,he other alonga t verticaldire_ion;becausein this caseI am thoroughly convinced theresultantsa motion that i whose square isequalto the sum of the squaresof the two components.The confusion ariseswhenoneundertakes compound uniform to a horizontal motionwith a verticalonewhichis naturallyaccelerated.I trust, therefore, e W maypursuethis discussion moreat length. [286] Sn_P.AndI needthisevenmorethanyousinceI amnotyet as clearin my mindasI oughtto be concerning thosefundamentalpropositions uponwhichthe othersrest. Evenin the case

264 THE TWO NEWSCIENCES OF GALTT.F.O caseof the two uniformmotions,one horizontal, he other t perpendicular,wishto understand I betterthe mannerinwhich youobtaintheresultantfromthecomponents.Now,Salviati, youunderstand whatweneedandwhatwedesire. SALv. Your requestis altogetherreasonable nd I willsee a whether mylongconsiderationfthesematterswillenableme o to makethemclearto you. But youmustexcuse e if in the m explanation repeatmanythingsalreadysaidby theAuthor. I Concerning otionsand theirvelocities r momenta[mov/m o mentie lorvelocit_ imperil hetheruniform naturallyaco w or celerated, ne cannotspeakdefinitely o untilhe has established a measureorsuchvelocitiesnd alsofortime. Asfortimewe f a havethealready widely adopted hours,firstminutes andsecond minutes. Soforvelocities, asforintervals just oftime,thereis need of a common standardwhichshallbe understoodand accepted everyone, by andwhichshallbethe sameforall. As hasalreadybeenstated,theAuthorconsidershevelocity t ofa freelyfallingbodyadaptedto this purpose, incethis velocity s increases ccording the samelawin allparts of the world; a to thusforinstance speedacquired leaden the bya ballofa pound weightstartingfrom rest and fallingverticallythroughthe heightof,say,a spear'slengthis the samein allplaces;it is therefore excellently adaptedfor representing momentum the [impeto] acquired thecaseofnaturalfall. in It still remainsfor us to discover methodof measuring a momentum thecaseofuniform in motionin sucha waythat all whodiscussthe subje&willformthe sameconception its of sizeand velocity[grander._ veloci@.This willpreventone e person fromimaginingt larger,anothersmaller,han it really i t is; so that in the compositionf a givenuniform o motionwith onewhichis acceleratedifferent enmaynotobtaindifferent d m valuesfor the resultant. In orderto determine nd represent a sucha momentum andparticular peed[imt_eto s e velocith particolate]our Authorhas foundno bettermethodthan to use the momentum acquired a bodyinnaturallyaccelerated otion. by m [2871 The speedof a bodywhichhas in this manneracquiredany momentum

FOURTHDAY z65 momentum whatever ill,whenconverted w intouniform motion, retainpreciselyucha speedas,duringa time-interval s equalto that ofthe fall,willcarrythebodythrough distance a equalto twicethat of the fall. But sincethis matter is onewhichis fundamentaln our discussion is wellthat we makeit peri it fecdylear c bymeans ofsome particular xample. e Letus considerhespeedandmomentum t acquired a body by fallknghrough height,say,of a spear[picca] standard t the asa whichwe may use in the measurement other speedsand of momentaas occasion demands;assume instancethat the for timeofsucha fallisfourseconds [minuti econdi s d'ora]; owin n orderto measure speedacquiredfroma fallthroughany the otherheight,whethergreateror less,one must not conclude that thesespeedsbearto one anotherthe sameratioas the heights offall;forinstance,tisnottruethat a fallthrough i four timesa givenheightconfers speedfourtimesasgreatasthat a acquiredby descentthroughthe givenheight;becausethe speedof a naturallyaccelerated otiondoesnotvary in prom portionto the time. Ashasbeenshown above,theratioofthe spacesisequalto thesquare oftheratioofthetimes. ILthen, as is often donefor the sakeof brevity,we take the samelimitedstraightlineas themeasure ofthe speed,and of the time,and alsoofthe space traversed duringthat .a time,it follows the durationof fall andthe speed that acquired the samebodyin passingoverany other by distance,snotrepresented ythissecond i b distance, but f' by a mean proportional etweenthe two distances. b This I canbetterillustrate an example.In theverby ticallineac,lay offtheportionabto representhe dist tancetraversedby a bodyfallingfreelywith acceleratedmotion: hetimeof fallmayberepresentedyany t b limited straightline,but forthesakeofbrevity, shall we representt bythe samelengthab;thislengthmayalso i beemployed a measure as ofthemomentum speed II 3 and Fig. acquiredduringthe motion;in short,let abbe a measureof thevarious hysical uantities hich p q w enterthisdiscussion. Havingagreedarbitrarilyuponab as a measureof these three

z66 THE TWO NEW SCIENCESOF GALILFO threedifferent uantities, amely, pace,time,and momentum, q n s our next task is to findthe time requiredfor fall througha givenverticaldistanceac,alsothe momentum acquiredat the terminal ointc, bothofwhichare to be expressed termsof p in thetimeandmomentum representedyab. These b tworequired quantitiesare obtainedby layingoff ad, a meanproportional between abandac;inotherwords, he timeoffallfroma to c is t represented adon the samescaleon whichwe agreedthat by the timeoffallfroma to bshould representedyab. In like be b mannerwe may say that the momentum[impeto grado o di veloci_] cquired c isrelatedto that acquired a at atb,inthesame mannerthat thelineadis relatedto ab,sincethevelocity varies direcdy the time,a conclusion, as whichalthoughemployed as a postulate in PropositionIII, is hereamplifiedby the Author. This point beingclearand well-established passto the we considerationf the momentum[impeto]n the caseof two o i compound otions, neof whichis compoundedf a uniform m o o horizontaland a uniformverticalmotion,whilethe other is compounded uniform ofa horizontal anda naturally accelerated verticalmotion. If both componentsre uniform, nd one at a a rightangles theother,wehavealready to seenthat the square of the resultantis obtained addingthe squaresof the compoby nents[19. 257]aswillbeclearfromthe following illustration. Let us imagine bodyto movealongthe verticalabwith a a uniformmomentum [impeto] 3, and on reaching to move of b a toward c with a momentum[veloc#R ed intervalit willtraverse3 cubitsalongthe verticaland4 alongthe horizontal. But a irnpeto] of c [_particle 4, 8othat duringthe sametimewhichmoves withthe resultantveFig.II 4 locity[velocitY] will,in the sametime,traversethediagonal c,whoselengthis not 7 cubits--thesumof a ab(3) and bc (4)wbut 5,whichis in_otenza equalto the sum of3 and4, that is,the squares and4 whenaddedmakezS, of3 whichisthe square ofac,and isequalto thesumofthe squares

[z88]

of

FOURTHDAY 267 ofaband bc. Henceacis represented thesidc orwemay by saytheroot--ofa square hose w areais25,namely SAsa fixed andcertain ruleforobtaininghemomentum t which [2891 results.fromtwo uniformmomenta,one vertical,the other horizontal, wehavetherefore following: the takethe squareof each,addthesetogether, ndextra(tthesquare a rootofthesum, whichwillbethe momentum resulting fromthe two. Thus,in the aboveexample,the bodywhichin virtue of its vertical motionwouldstrikethe horizontal lanewith a momentum p [for_]of3,would owing itshorizontal to motion alonestrikeat cwitha momentum of4;but if thebodystrikes witha momentumwhich isthe resukant fthesetwo,itsblow o willbethatofa bodymoving witha momentum [velocita eforza] of5;andsucha blowwillbe the sameat allpointsofthe diagonal c,sinceits a components arealwayshe same t andneverincrease rdiminish. o Letus nowpassto the consideration uniform ofa horizontal motioncompounded withthe vertical otionofa freelyfalling m bodystartingfromrest. It is at onceclearthat the diagonal whichrepresentshe motioncompounded t ofthesetwois nota straightline,but, as has beendemonstrated, semi-parabola, a in whichthe momentum [impeto] always is increasing because the speed[velocita] theverticalcomponent always of is increasing. Wherefore, determine momentum to the [impeto] any at givenpoint in the parabolic diagonal, t is necessary to i first fixuponthe uniform horizontal omentum m [impeto] then, and treatingthe body as one fallingfreely,to find the vertical momentum the givenpoint;this latter can be determined at onlybytakingintoaccounthe duration t offall,a consideration whichdoesnotenterintothe composition oftwouniform motionswherethe velocities nd momenta always same; a are the but herewhereone of the component otionshas an initial m valueofzeroandincreases itsspeed [velodta] indirecCt proportion to the time,it followshat the timemustdetermine speed t the [veloci_] the assigned at point. It onlyremains obtainthe to momentum resulting fromthesetwocomponentsasinthe case ( ofuniform motions) yplacing square b the oftheresultantequal to

268 THE TWO NEW SCIENCESOF GALII.F.O to the sumof the squaresof the two components.But here againit isbetterto illustrate bymeansofanexample. On the verticalaclayoff: poetion any abwhichweshallemployasa measure ofthe space traversed bodyfalling bya freely alongthe perpendicular, likewise a measure the timeand as of alsoof the speed_grado velocitY] di or,wemay say,ofthe momenta[impet,].It is at onceclearthat if themomentum a of [29o] bodyat b, afterhavingfallenfromrest at a, be divertedalong thehorizontal irec'tion d bd,withuniform motion,its speedwill be suchthat, duringthe time-interval b, it willtraversea a distance hichisrepresented ythelineMandwhichistwiceas w b great as ab. Now choosea point c, suchthat bcshallbe draw the line ce equal and parallelto bd; throughthe pointsto and e drawthe paequal b ab, and throughc rabola bei. Andsince, uring d the time-interval ab,the horizontaldistance orce,double M thelength ab,istraversed with othe momentum ab,and since

J d _--./_ .1 e

d bcistrav_ ' thevertical istance duringan equaltime-interval ersed,thebodyacquiring t c a './_ . _ momentumrepresented a by Fig.II5 the same horizontal, bd,k followsthat during timeabthebodywillpassfrombto e along the the parabola be,andwillreache witha momentum compounded of twomomentaeachequalto ab. Andsinceone of theseis horizontal andtheothervertical,thesquare oftheresultantmomentumisequalto the sumofthe squaresofthesetwo components,i. e.,equalto twiceeitheroneof them. Therefore, ifwe layoff:he distancebf,equalto ha,and draw t the diagonal f, it follows a that the momentum[impeto pere cossa] te willexceed ofa bodyat bafterhavingfallenfrom a that a,

FOURTHDAY z69 a, orwhatisthe samething,willexceed horizontal the momentum [percossa dell'impeto] alongbd,in the ratioofaf to ab. Suppose owwe choose the heightof fall a distancebo n for whichis notequalto but greaterthanab,andsuppose bg that represents meanproportional a between baandbo; hen,stillret tainingbaas a measure ofthedistance fallenthrough, romrest f at a, to b,alsoasa measure ofthe timeand ofthemomentum whichthe fallingbodyacquires b,it followshat bgwillbe at t the measure the timeand alsoof themomentum of whichthe bodyacquiresnfalling i frombtoo. Likewise justasthemomentumabduring timeabcarried bodya distance the the along the horizontal qualto twiceab,so now,duringthe time-interval e bg,the bodywillbe carriedin a horizontal irecCtion d through a distancewhichis greaterin the ratioof bgto ba. Lay offlb equalto bgand drawthe diagonal /, fromwhichwe havea a quantitycompoundedftwovelocitiesimperil o [ onehorizontal., theothervertical;hesedetermineheparabola.Thehorizontal t t anduniform velocity isthat acquired b infalling at froma; the otheristhat acquired o,or,wemaysay,at i, bya bodyfalling at throughthe distance bo,duringa timemeasured thelinebg, by [29 z] whichllnebgalsorepresentshe momentum t ofthe body. And in likemannerwemay,by takinga meanproportional between the two heights,determinethe momentum[bnpeto] t the a extreme of the parabolawherethe heightis lessthanthe end sublimity ab;this meanproportional to be drawnalong is the horizontaln placeof bf,and alsoanotherdiagonaln placeof i i af,whichdiagonal illrepresenthemomentum the extreme w t at endoftheparabola. To what has hithertobeen said concerninghe momenta, t blowsor shocksof projecCtiles, mustadd anothervery imwe portantconsideration; determineheforceand energy to t ofthe shock[for_ edenergia ellapercossa]isnot sufficient cond it to sideronlythespeedoftheprojecCtiles, butwemustalsotakeinto accountthe nature and condition the targetwhich,in no of smalldegree, eterminesheefficiency d t ofthe blow. Firstof all it is wellknownthat the targetsuffers iolenceromthe speed v f [veloci_]

z7o THE TWO NEW SCIENCESOF _ALIT,EO [veloci_] the projectile proportion it partlyor entirely of in as stopsthemotion; ecause b ifthe blow fallsuponanobjectwhich yields theimpulse to [velocit_ delperc_ziente] withoutresistance sucha blowwillbe ofno effect;likewise whenone attackshis enemy witha spearandovertakes himat aninstantwhenheis fleeing with equal speedtherewillbe no blowbut merelya harmless touch. But if the shockfallsuponan objectwhich yieldsonlyin part thenthe blowwillnot have ks fulleffe&, but the damage willbe in proportiono the excess fthe speed t o oftheprojectile overthat ofthereceding body;thus,forexample,if the shotreaches targetwitha speedof IOwhilethe the latter recedeswith a speedof 4, the momentum shock and [impeto percossa] e willbe represented 6. Finallythe blow by willbea maximum, sofarastheproje&ilesconcerned, in i when the targetdoesnot recedeat allbut if possible ompletely c resistsand stopsthe motionofthe proje&ile.I havesaidin so far as theproje_ileis concerned becauseif the target should approach projecetile shockofcollisioncolpo l'incontro] the the [ e would greaterin proportion be asthe sumof the twospeeds is greaterthanthatoftheprojectile lone. a Moreovert isto beobserved the amountofyielding i that in the targetdepends onlyuponthe qualityof the material, not as regards hardness, hether be ofiron,lead,wool,etc.,but w it alsouponits position. If the positionis suchthat the shot strikesit at rightangles,hemomentum t impartedby theblow [impeto delcolpo] illbe a maximum; ut if the motionbe w b oblique,hat is to say slanting, he blowwillbe weaker;and t t moreandmoresoin proportiono theobliquity; or,nomatter t f howhard the materialof the targetthus situated,the entire momentum [impeto moto] f the shotwillnot be spent and e o stopped; he projecCtile t willslideby and will,to someextent, continueits motionalongthe surfaceof the opposing body. Allthathasbeensaidabove concerning theamount fmomeno tum in the projecCtilethe extremity the parabola at of mustbe understoodorefer a blowreceivedna lineat rightangles t to o to thisparabola oralongthe tangentto theparabola t thegiven a point

[z9z]

FOURTHDAY z7, point;for, eventhoughthe motionhas two components , one horizontal, othervertical, eitherwillthemomentum the n along the horizontalnor that upon a plane perpendicular the to horizontal a maximum, be sinceeachof thesewillbe received obliquely. SAcR. Yourhavingmentionedheseblows t andshocks recalls to myminda problem, r rather a question,n mechanics f o i o whichno authorhas givena solution said anything or which diminishes y astonishment evenpartlyrelieves y mind. m or m My difficulty surpriseconsistin not beingableto see and whence uponwhatprinciples derived and i theenergyand immense force[energia eforza immensa] which makes itsappearance in a blow;for instance seethe simple we blowof a hammer, weighing morethan 8 or IOlbs.,overcomingesistances not r which, ithouta blow,would w notyieldto theweightofa body producing impetusby pressurealone,eventhoughthat body weighed manyhundreds fpounds. I would o liketo discover a methodofmeasuring theforce[forza] ofsucha percussion. can I hardlythinkit infinite, utincline b ratherto theview it has that its limit and canbe counterbalanced measured other and by forces, uchasweights, s orbylevers rscrews rothermechanical o o instruments hichareusedtomultiply w forcesn a manner hich i w I satisfacCtorily understand. SAnv. ouarenot alonein yoursurprise this effecCt in Y at or obscurity asto thecause ofthisremarkable property. I studied this mattermyself ora while vain;butmyconfusion erely f in m increased ntilfinally u meeting ourAcademicianreceived I from [293] himgreatconsolation. irsthetoldmethat he alsohad fora F longtimebeengroping inthe dark;butlaterhesaidthat,after havingspentsomethousands f hoursin speculatingnd cono a templating thereon,he had arrivedat somenotions whichare far removed fromour earlierideasand whichare remarkable fortheirnovelty. Andsince nowI know youwould that gladly hearwhatthesenovelideasareI shallnotwaitfor youto ask but promisethat, as soonas our discussion projectiles of is completed, willexplain thesefantasies, r if you please, I all o vagaries

z7z THE TWO NEW SCIENCES GALILEO OF vagaries, s far as I can recallthem fromthe wordsof our a Academician. n the meantime proceedwiththe proposiI we tionsoftheauthor.
PROPOSITION, PROBLEM V

Having aparabola, the given find point, itsxis xtended in a e upwards, which from aparticlefall must in orderdescribe to this parabola. same Let abbe the givenparabola, bits amplitude, nd heits axis h a extended.Theproblem findthepointefromwhicha body isto mustfallinorderthat,afterthemomentum whichit acquires t a a has beendivertedintoa horizontal dire&ion, t willdescribe i the parabolaab. Drawthe horizontal g,parallelto bh, and a e havinglaidoff af equal to ah, I drawthe straightlinebfwhich ] willbe a tangentto the parabJ it" la at b, and willinterse&the o /[_" horizontal agatg:choose such e [ that agwillbe a meanproporeq/ ! tionalbetween a/and ae. Now __.._..._ a I saythat e is the pointabove ] sought.That is,if a bodyfalls /_ _'_It fromrest at this point e,and if "" the momentum acquiredat the Fig.If6 pointa be divertedintoa horizontal dire&ion,and compounded ith the momentumacw quiredat h in falling fromrest at a, thenthe bodywilldescribe the parabolaab. Forif weunderstand eato be the measure of the timeoffallfrome to a, andalsoofthemomentum acquired at a,thenag(which meanproportional isa between eaand af) willrepresentthe time and momentum fall fromf to a or, of whatisthe same thing,froma toh;andsince bodyfalling a from e, during timeea,will,owing the momentum the to acquired at a, traverse at uniformspeeda horizontaldistancewhich is twace ea,it followshat, the bodywillif impelled y the same t b momentum, uringthe time-intervalag traversea distance d equalto twiceagwhichis the halfofbh. This istrue because, in

FOURTHDAY 273 in the caseof uniformmotion,the spacestraversedvary direcCtly thetimes. Andlikewisef themotionbeverticaland as i start fromrest, the bodywilldescribe distance in the the ah [2941 timeag. Hencethe amplitude hand the altitudeaharetravb ersedby a bodyin the sametime. Thereforehe parabolaab t willbe described a bodyfallingfromthe sublimity e. by of
Q. E. F.

COROLLARY Henceit followshat halfthe base,oramplitude, t ofthesemiparabola(whichsone-quarter i oftheentireamplitude) mean isa proportional between altitudeandthe sublimityromwhich its f a falling bodywilldescribe thissame parabola. PRoPosrrioNI, PROBLEM V Giventhe sublimity thealtitudeof a parabola, o find and t itsamplitude. Let the lineac,in whichlie the givenaltitudecband sublimity ab, be perpendicularo t F a thehorizontal linecd. Theproblem is to find the amplitude, alongthe horizontalcd,of the semi-parabola whichis described with the sublimitybaand altitude bc. Lay off cd equal to b twicethe meanproportional betweencband ba. Then cdwill be the amplitudesought,as is evident romthe precedingrop-d f p -: c_ osition. Fig. xx7
THEOREM. PROPOSITIONII V

I_ projecdles describe semi-parabolas the sameampliof tude,themomentum requiredo describe onewhose t that amplitude doubleits altitudeis lessthan that required is foranyother. Let

_74 THE TWO NEW SCIENCESOF GALILFO Let bdbe a semi-parabola hoseamplitudecd is doubleits w altitudecb;onits axisextended upwards layoffbaequalto ks ahitudebc. Drawthe linead whichwillbe a tangentto the parabolaat d and willcut the horizontalinebeat the point l e,makingbeequalto bcand alsoto ha. It isevidentthat this parabola illbedescribed ya projedtile w b whose uniform horizontalmomentum isthat which would it acquire binfalling at from restat a andwhose naturallyacceleratederticalmomentum v is that ofthebodyfalling c,fromrestat b. Fromthisit follows to

C Fig. I:8

that the momentum the terminalpoint d, compoundedf at o thesetwo,is represented the diagonal e,whosesquareis by a equalto the sumof the squares ofthe two components. ow N letgdbe any otherparabolawhatever havingthe sameamplitudecd,but whosealtitudecgis eithergreateror lessthanthe altitudebc. Let /u/ be the tangent cutting the horizontal through at k. Sele_a pointl suchthat hg:gk gk:gl.Then g = froma preceding proposition [V],k followshat glwillbe the t height

[29S1

FOURTHDAY 275 heightfromwhicha bodymust fallin orderto describethe parabola gd. Let gmbe a meanproportional between aband gl;thengm will[Prop.IV]representhe timeand momentum t acquired at g.bya fallfroml; forabhasbeenassumed measure asa ofboth ume and momentum.Againlet gnbe a meanproportional between bcand cg;it willthenrepresenthe timeand moment tumwhichthe bodyacquires c in falling at fromg. If nowwe joinm andn, this linemnwillrepresenthemomentum d of t at the projecdle traversing paraboladg;which the momentum is, I say,greaterthan that of the projeccile travelling alongthe parabolabdwhose measure wasgivenbyae. Forsincegnhas beentakenasa meanproportional between bcandgc;andsince bcis equalto beandalsoto kg (eachof thembeingthehalfof dc__) it follows__ cg:gn that =gn:gk, ascgor (hg)is to gk so__is and ng_ to gk2:but by constru_ionhg:gk=gk:gl.Henceng2: gk_=gk:gl. But gk:gl=-gk_:_,since gmis a meanproporgm tionalbetween and gl. Therefore threesquares kg the ng,kg, mgforma continued proportion, _ =gk2: in. Andthe g-n_: g _ sumof the twoextremes hichis equalto the square mnis w of greaterthan twicethe squareof gk;but the squareof aeis double thesquareofgk. Hencethesquareofmnisgreater than thesquareofaeandthelengthmnisgreater thanthelengthae. Q.E.D. [z96] COROI JARY Converselyt is evidentthat lessmomentum i willbe required to senda proje6tileromtheterminal ointd along f p theparabola bdthanalonganyotherparabola having elevation an greater r o lessthan that of the parabola bd,for whichthe tangentat d makesan angleof 45 with the horizontal.From whichit followshat ifprojecdles t arefiredfromthe terminal pointd, all having thesamespeed, uteachhavinga different b elevation, the maximum range,i. e.,amplitude fthe semi-parabola ofthe o or entireparabola, illbe obtained w whentheelevation is45:the other

Z76 THE TWO NEW SCIENCESOF GALII,V.O othershots,firedat anglesgreateror lesswillhavea shorter range. SACR. Theforceof rigiddemonstrataons suchas occuronly in mathematicsillsme with wonderand delight. From acf countsgivenbygunners,I wasalreadyawareofthe facethat t in the useof cannon mortars,the mnx_rnum and range,that is the one in whichthe shotgoesfarthest,is obtainedwhenthe elevation is45 or, asthey say,at the sixthpointof the quadrant; but to understand whythis happensfar outweighshe t mereinformation btainedby the testimony othersor even o of byrepeated experiment. S_u,v. hatyousayisverytrue. Theknowledge a single W of fa_ acquired througha discovery its causesprepares of the mindto understandnd ascertain a otherfac%s ithoutneed w of recourseo experiment, t precisely in the present ase, as c where byargumentation theAuthor alone proves withcertainty thatthemaximum range ccurs hentheelevation o w is45. He thusdemonstrates whathas perhaps everbeenobserved n in experience, namely,hat of othershotsthosewhich t exceedor fallshortof 45 by equalamounts aveequalranges; that h so if theballshavebeenfiredoneat anelevationf 7 points,the o otherat 5, theywillstrike levelat the samedistance:he the t sameistrueif theshots arefired 8 andat4 points, t9 andat at a 3,etc. Nowletushear thedemonstration ofthis. [2971 TaEOr.E_.PxoFosrnoN VIII The amplitudes two parabolas of described proje_iles by fired withthe samespeed, at angles felevation hich but o w exceedandfall shortof 45 by equalamounts, areequal to each other. In thetriangle cblet the horizontal m sidebcandthe vertical cm,which forma rightangleat c, beequalto eachother;then the angle mbcwillbea semi-right angle; thelinecmbeprolet longed d, sucha pointthat thetwoangles to atb, namely robe andmbd,oneabove andtheotherbelow thediagonal rob,shall beequal.It isnow tobeproved thatinthecase oftwoparabolas described

FOURTHDAY 277 described twoprojecCtiles by firedfromb withthe samespeed, oneat the angleofebc,the otherat the angleofdbc,theiramplitudeswillbe equal. Nowsincethe externalanglebmcis equalto thesumofthe internal nglesndb a r anddbm wemayalso equateto themthe anglembc;but ifwereplacethe angledbmby robe,henthissame t d anglembcis equalto the tworobeandbdc: / and if we subtraCtfrom each side of this equation theanglerobe, wehavetheremain[ derbdcequalto the remainder bc. Hence e m the two trianglesdcb and bceare similar. _ BiseCtthe straightlinesdc and ec in the /_/e pointsh andf: and drawthe lineshi andfg /JJ parallelto the horizontal b,and chooseI /Jj,_ c f suchthat dh:hl ih:hl. Thenthe triangle l _ = _'hl " willbesimilar i/M, nd alsotothe triangle r to a o ," _ andsinceih andgfareequal, achbeing e Fig.ii9 ofbc,itfollows h/is equalto fe and alsoto fc; and if that we addto eachofthesethecommon artfh,it willbe seenthat p chisequaltoil. Let us nowimagine parabola a describedhrough points t the handbwhosealtitudeishcandsublimity L Its amplitude ill h w becbwhichisdouble thelength hisince hiisa meanproportional between dh(orc/z) ndh/. Thelinedbistangentto theparabola a at b, sincechis equalto hd. If againwe imagine parabola a described throughthe pointsf and b, with a sublimity/and f altitude of whichthemeanproportional orone-half f fc, isfg, o cb,then, as before, illcbbe the amplitude the lineeba w and tangentat b;forefandfcareequal. [z98] But the twoanglesdbcand ebc,the anglesof elevation, iffer d byequalamounts froma 45 angle. Hencefollowshe proposi t tion. TIIEORE_. PROPOSrrION IX Theamplitudes parabolasqual their of two are when e altitudes andsublimities are nversely i proportional. Let

278 THE TWO NEW SCW.NCES GALII,I_.O OF .Let the altitudegf of the parabola bearto the altitudecb fh oftheparabola Mthesameratiowhichthesublimity abearsto b the sublimitye; thenI say the amplitude r hgis equalto the amplitude _. Forsincethefirstofthesequantities, a gf,bearsto _'the second cbthesame e ratiowhichthe third, ha,bearsto thefourth re, it follows that the area of the re&angle the re&_nglecb.ba; therefore squares which are equal to ___ g _.feisequalto that of these re&anglesare Fig.i2o equal to eachother. But [byProposition IIthe squareof halfofghisequalto the V re&angle f.fe;and the square halfof cdis equalto the recg of tangle cb.ba.Therefore thesesquaresand theirsidesand the doubles oftheirsidesareequal. But theselastare the amplitudes ghand cd. Hencefollowshe proposition. t L_M__A T_EVm.T WmORovosrrm_ FOg O P Ifa straightinebecutat anypointwhatever ndmeanprol a portionals between lineandeachofits partsbe taken, this the sumofthe squares f thesemeanproportionals equal o is to thesquare oftheentireline. Letthe lineabbe cutat c. ThenI saythat thesquareofthe mean proportional betweenaband acplus the squareofthemeanproportional between aband cbisequalto the squareofthewhole lineab. This is evidentas soonas we describea semicircle ponthe entirelineab,a u t ere&a perpendicular cdat c, and drawda Fig.i2i anddb. Fordaisa meanproportional between aband acwhile [299] db is a meanproportional betweenaband bc:and sincethe angleadb, nscribed a semicircle, a rightanglethe_sumof i in is

the

FOURTHDAY 279 thesquares ofthe linesdaanddbis equaltothe squareofthe entire lineab. Hencefollowsheproposition. t TaEOREM. PROPOSmON X The momentum [impetus momentum] seu acquired a by particleat theterminal ointofany semi-parabola p isequal tothat whichit wouldacquire infalling through vertical a distance equaltothesumofthesublimityndthealtitude a ofthesemi-parabola.* Letabbe a semi-parabola havinga sublimityaandan altitude d ac,thesumofwhichistheperpendicular Now dc. I saythemomentum oftheparticle atbisthesame d asthatwhich itwould acquirenfalling i freely from e d to c. Let us take the lengthof dc itselfasa 2 a measure timeand momentum, layoffcf of and equalto the meanproportional etween and b cd da;alsolay off cea meanproportional between cdand ca. Nowcfisthemeasure ofthetimeand of the momentum acquiredbyfall,fromrest at d, throughthe distanceda;whileceis the time j / and momentum fall,fromrest at a, through of _ the distanceca;alsothe diagonalefwillrepre-_'( _* c senta momentum whichisthe resultant f these Fig. 22 o 1 two,and is therefore momentum the terminalpointof the at theparabola, . b And sincedchas beencut at somepointa and sincecfand cearemeanproportionals between thewhole ofcdanditsparts, daand ac,it follows, fromthe preceding lemma, hat the sum t ofthe squares ofthesemeanproportionals isequalto thesquare of the whole:but the squareof ef is alsoequalto the sumof thesesamesquares; hence followshat the lineefisequalto w it t dc. Accordinglyhe momentum t acquiredat c by a particlein failingfromd is the sameas that acquired b by a particle at traversing parabolaab. the Q.E.V.
* In modemmechanics well-knownheoremassumes following this t the form: Thespeedof a projectile t anypointis thatproduced a byafall frora the directrix. [Trans.]

280 THE TWO NEW SCIENCESOF GALTI.EO COROT J.ARY Henceit followshat, in the caseof allparabolas herethe t w sumofthe sublimity altitudeis a constant, hemomentum and t at theterminalpointisa constant. PROBLEM.ROPOSITION P XI Given he t amplitude the and speed [impetus] teratthe minal of pointasemi-parabola, altitude. to its find Let given berepresented ertical and the speed bythe v llne ab, the amplitudeby the horizontal inebe;it is requiredto find l the sublimity the semi-parabola of whoseterminalspeedis ab and amplitudebe. From what precedes [Cor.Prop.V] it is clearthat halfthe amplitude bcisa meanproportional between thealtitude ndsublimityftheparabolafwhich a o o theterminal speed isequal,inaccordance withthepreceding proposition, to _t_ speedacquiredy a bodyinfalling the b //_ CTherefore a lineba mustbe cutat b. from restat throughthe distance a the a J_/f I".pintsuchthat thereCtangleormed f by A \ I its twopartswillbe equalto the square / \ _ [ ofhalfbc, amely Necessarily, n bd. there/ _N_] _ fore,bdmustnot exceedthe halfof ba; ofa straightlinetheoneof greatest rea a isobtainedwhenthe lineisdividedinto two equalparts. Let e be the middle /_ forof allthe rectanglesormedby parts f ,_ equalto the line ab;and nowif bdbe _ point ofbetheproblemissolved;for be willbe the altitudeand eathe sublimity Fig.I_3 of the parabola. (Incidentally may we observea consequence alreadydemonstrated, namely:of all parabolasdescribedwith any giventerminal speedthat for whichthe elevation 45 willhave the maximum is amplitude.) But supposethat bd is lessthan half of bawhichis to be divided

[3ool

FOURTH DAY 281 dividedin sucha waythat theredtang|e ponitspartsmaybe u equalto the square fbd. Uponeaasdiameter o describe semia circle efa,in whichdrawthe chordaf,equalto bd:joinfe and lay offthe distanceegequalto re. Thenthe rectangle bg.ga plusthe square ofegwillbeequalto thesquare ofea,andhence alsoto the sumofthe squares ofaf andre. If nowwesubtracCt the equalsquaresoffe andgethereremainshe recCtangle t bg.ga equalto the squareofaf, that is,of bd,a linewhichis a mean proportional between bgand ga;fromwhichit isevidentthat the semi-parabola whoseamplitudeis bc and whoseterminal speed[impetus] represented y bahas an altitudebgand a is b sublimity ga. If however elay offbiequalto ga,thenbiwillbe the altiw tudeofthesemi-parabola ic,andiawillbeitssublimity.From the preceding demonstration eareableto solve following w the problem.
PROBLEM. PROPOSITION XII

To computeand tabulate the amplitudes all semiof parabolas whichare described yproje6tiles b firedwiththe sameinitialspeed[impetus]. From the foregoing follows it that, wheneverhe sumof t the altitudeand sublimityis a constantverticalheightfor any set of parabolas,theseparabolasare describedby proje_iles havingthe sameinitialspeed;allverticalheightsthus obtainedare therefore included etween parallel orizontal b two h lines. Let cbrepresent horizontalineand aba verticalline a l ofequallength;drawthe diagonal ac;the angleacbwillbeone of45;letd bethemiddle pointoftheverticalineab. Thenthe l semi-parabola is the onewhichis determined y the subdc b limityad and the altitudedb,whileits terminalspeedat c is that whichwould beacquired bbya particlefalling at fromrest at a. If nowagbedrawnparallelo bc,thesumofthe altitude t and sublimityfor any other semi-parabola havingthe same terminalspeedwill,in the mannerexplained, equalto the be distance betweenthe parallelinesagandbc. Moreover, l since it

[3o ]

z8z THE TWO NEW SCIENCESOF GALILEO it has alreadybeenshownthat the amplitudes f two semio parabolas rethesame a whentheirangles ofelevation differ from 45bylikeamounts, followshat the samecomputation hich it t w isemployedorthelarger levation f e willserve alsoforthesmaller. Let us also assumeIOOOOas the greatestamplitudefor a parabola whoseangleofelevation is45; this thenwillbe the lengthof the line o baand the amplitudeof the semiparabolabc. This number,IOOOO, a'is sele_edbecausein thesecalculationsweemploy tableoftangents a _in which this is the value of the tangentof 45. And now,coming down business, to drawthe straight line ce makingan acute angleecb greaterthan acb:the problemnow is to draw the semi-parabolao t whichthe lineec is a tangentand tr _ forwhichthe sumofthe sublimity Fig.x24 and the altitudeis the distanceba. Take the lengthof the tangent* be fromthe table of tangents,usingthe anglebceas an argument:letf bethe middle pointof be;nextfind a third proportionalo bf and bi (the t half of bc),whichis of necessity greaterthanfa.t Callthis fo. Wehave nowdiscovered that, for the parabolainscribed in the triangleecbhavingthe tangentce and the amplitude cb,the altitudeis bfand the sublimityo. But thetotallength f of bo exceedsthe distancebetweenthe parallelsag and cb, whileour problemwas to keep it equal to this distance:for both the parabolasoughtand the paraboladc are described Thereader ill bserve w o thattheword "tangent" ishereusedina sense omewhat s different fromthat of the preceding sentence. he T "tangent c"isa linewhichouchesheparabolat c;butthe"tane t t a genteb"isthesideoftheright-angled triangle hich opposite w lies the an gleecb, line a whoseengthsproportional l -i tothenumerical of value thetangent ofthisangle.[Trans.] t This factisdemonstrated inthethirdparagraph below.[Tra_.l

[3ozl

FOURTHDAY 283 by proje_iles firedfromc withthesamespeed. Nowsincean infinitenumberof greaterand smallerparabolas, imilarto s eachother,maybedescribed ithinthe angle wemustfind w bce another arabola p which likecdhasforthesumofitsaltitudeand sublimityhe height a,equaltobc. t h Therefore layoffcrso that, ob:ba bc:cr; = thencrwillbe the amplitude fa semi-parabola o forwhichbce isthe angle ofelevationandfor whichthesumofthe altitudeand sublimitys the i distance etween b theparallels aandcb,asdesired.Theprocess g is thereforeas follows: ne drawsthe tangentof the given O anglebce; akeshalfofthistangent, ndaddstoit thequantity, t a fo,whichis a third proportional the halfofthis tangent nd to a the halfofbc;thedesiredamplitude ris thenfoundfromthe c following roportion =bc:cr.For example the angle p ob:ba let ecb beoneof5o;itstangentis11918, halfofwhich, amely n bf, is 5959; alfofbcis5o00;thethirdproportional h ofthesehalves is4195, hich w addedto bfgives thevalue10154orbo. Further, f asobisto ab,that is,as 10154sto IOO_, oisbc,or IOOOO i s (each beingthe tangentof45) to cr,whichis the amplitudeought s and whichhas the value9848 the maximum , amplitude being bc,orIOOOO. Theamplitudes oftheentireparabolasredouble a these,namely,19696 nd2cx:x:x). a Thisisalsotheamplitude fa o parabola hoseangleofelevation w is40, sinceit deviates y an b equalamountfromoneof45. [303] SAca.In order thoroughly to understand thisdemonstration I needto beshown thethird how proportional ofbfandbiis,as theAuthor indicates, necessarily greater thanfa. SAT.v. Thisresult an,I think,beobtaineds follows.The c a square fthemean o proportional between twolinessequalo the i t reccangle formed y thesetwolines. Therefore squareof b the /z/(orof bdwhichisequaltob0mustbeequal othe re(tangle t formed fb andthe desired by thirdproportional. third This proportional isnecessarily greater fa because reccangle than the formed bfandfa is lessthanthesquare fbdbyanamount by o equal o thesquare t ofdr,asshown inEuclid, II.1. Besidest is i tobeobserved thatthepoint , which f isthemiddle oint f the p o tangent

284 THE TWO NEW SCIENCES OF GALII,V.O tangent eb,fallsin generalabovea and only onceat a; in which casesit is self-evidentthat the third proportionalto the half ofthe tangent and to the sublimityb/lies whollyabovea. But the Author has taken a case where it is not evident that the third proportionalis alwaysgreater than fa, so that when laid offabovethe pointf it extendsbeyondthe parallelag. Now let us proceed. It willbeworth while,by the use ofthis table, to compute another givingthe altitudes of these semiparabolasdescribedby projectileshavingthe sameinitialspeed. The construfftion asfollows: is [3o41 Amplitudes ofseml-parabolas Altitudes f semi-parabolas o dedescribed iththe same w inscribedwiththe sameinitial itialspeed, speed. Angle of Angle of Elevation Elevation 45 1oooo 46 9994 44 47 9976 43 48 9945 42 49 99oz 41 50 9848 40 51 9782 39 5z 9704 38 53 9612 37 54 9511 36 55 9396 35 56 927 z 34 57 9136 33 59 58 8829 31 8989 32 60 8659 30 61 8481 29 62 829o 28 63 8o9o 27 64 7880 26 65 766o 25 66 7431 24 67 7191 23 68 6944 22 69 6692 21 Angle of Angle of Elevation Elevation 1 3 460 5173 2 I3 47 5346 3 28 48 5523 4 50 49 5698 5 76 50 5868 6 lO8 51 6o38 7 15o 5z 62o7 8 194 53 6379 9 245 54 6546 Io 3oz 55 67IO 11 365 5_; 6873 12 432 57 7033 13 5o6 58 719 15 14 670 585 _ 7348 75o2 16 76o 61 7649 17 855 62 7796 18 63 7939 29 I_ 64 8078 2o H7o 65 8214 zz 2285 66 8346 zz I4oz 67 8474 23 I527 68 8597 24 I685 69 8715 25 1786 70 8830

FOURTH DAY
Amplitudesof semi-parabolas Altitudes of semi-parabolasdedescribedwith the same in- scribed with the same initial itial speed, speed. Angleof Elevation 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 Angleof Elevation 6428 2o 6X57 I9 5878 x8 5592 x7 5300 I6 500 15 4694 14 4383 13 4067 12 3746 xl 342o IO 309 9 2756 8 2419 7 2079 6 1736 5 x39I 4 lO44 3 698 2 349 I Angleof Angleof Elevation Elevation 26 I922 7l 27 2o6I 72 28 2204 73 29 2351 74 3 2499 75 31 2653 76 32 281 o 77 33 2967 78 34 3128 79 35 3289 80 36 3456 81 37 362x 82 38 3793 83 39 3962 84 40 4132 85 41 43o2 86 42 4477 87 43 4654 88 44 4827 89 45 5oo0 9o

285

8940 9045 9144 9240 9330 9415 9493 9567 9636 9698 9755 9806 985x 9890 9924 9951 9972 9987 9998 xoooo

PROBLEM. PROPOSITION XIII

[305]

From the amplitudes f semi-parabolas o givenin the precedingtableto findthe altitudes ofeachof theparabolas described iththesameinitialspeed. w Let bcdenotethe givenamplitude; ndlet ob,the sumofthe a altitudeand sublknity, bethemeasure oftheinitialspeed which understoodto remainconstant. Next we must find and determinehealtitude, hich t w weshallaccomplish bysodividing that the recCtangle contained yitspartsshallbe equalto the b squareof halfthe amplitude, c. Letf denotethis pointof b division d and {be the middle and pointsof oband bcrespectively. The squareofib isequalto the rectangle f.fo; the b but squareof dois equalto the sumof the rectangle bf.foand the
square

z86 THE TWO NEW SCIV.NCES GALILEO OF squareoffd. If, therefore, romthe squareof dowe subtract f the square biwhichisequalto the recCtangle of bf.fo,therewill remainthe squareoffd. The altitudein question,bf, is now obtained addingto this length, d, thelinebe{.Theprocess by f isthenas follows: romthe squareof halfof bo F i_whichis known,subtractthe squareof bi which is alsoknown;take the squareroot of the remainder, add to it theknownlengthdb;then and youhavetherequired altitude,bf. Example.Tofindthealtitudeofa semi-parabola described ithanangle w ofelevationf55. From o the preceding table the amplitudeis seento be c. i _ 9396,of whichthe halfis 4698,and the square _ 22o7Izo4. 'When this is subtractedfrom the Fig. Iz5 square of the half of bo, which is always 25,ooo, ooo,the remainder s z928796, whichthe squareroot i of is approximately7IO. Addingthis to the halfof bo,namely I 5oo% wehave67Iofortheakitudeofbf. It willbeworthwhileto adda thirdtablegiving altitudes the and sublimitiesor parabolasin whichthe amplitude a conf is stant. SACg. shallbe very gladto seethis; for fromit I shall I learnthe difference speedandforce[degl'rnpeti delleorze] of i e f required fireprojectiles ver the samerangewithwhat we to o callmortarshots. This difference ill,I believe, ary greatly w v withtheelevation sothat if,forexample, onewished employ to anelevation 3 or 4, or 87 or 88 andyet givethe ballthe of samerange whichit hadwithanelevation of45 (where ehave w shownthe initialspeedto be a minimum)he excessof force t required ill,I think,beverygreat. w SALV. Youarequkefight,sir;andyouwillfindthat in order to perform thisoperation completely, allanglesofelevation, at youwillhaveto make greatstridestowardan infinitespeed. Wepassnowto the consideration ofthetable.

[3o6]

FOURTHDAY

[307]

287

Tablegivingthe altitudesand sublimitiesf parabolas o of constant amplitude, amelyIoooo,computed n for eachdegree ofelevation.
Angleof Elevation Altitude Sublimity I 87 286533 2 175 14245o 3 z6z 95802 4 349 71531 5 437 5714 z 6 525 47573 7 614 40716 8 7o2 35587 9 792 3H65 IO 88I 28367 II 972 25720 IZ XO63 23518 13 II54 21701 I4 I246 20056 15 1339 I8663 I6 I434 I7405 I7 I529 I6355 18 I624 15389 19 1722 14522 ZO X8ZO 13736 2X 1919 13O24 22 2ozo I2376 23 2123 ii778 24 2226 II230 25 2332 IO722 26 2439 10253 27 2547 9814 28 2658 9404 29 2772 920 3O 2887 8659 3I 3o8 8336 32 3124 8OOX 33 3247 7699 34 3373 7413 35 35oI 7141 36 3633 6882 37 3768 6635 38 39o6 6395 Angleof Elevation Altitude Sublimity 460 5177 4828 47 5363 4662 48 5553 452 49 5752 4345 5 5959 4196 51 6174 4048 5z 6399 39 6 53 6635 3765 54 6882 3632 55 7141 35 0 56 7413 3372 57 7699 3247 58 8002 3123 59 8332 34 60 8600 2887 6I 9020 277I 62 9403 2658 63 98X3 2547 64 XO25I 2438 65 10722 233I 66 II230 2226 67 II779 2122 68 i2375 2020 69 I3025 I919 7 I3237 1819 71 I452I I72I 72 I5388 1624 73 I6354 I528 74 I7437 I433 75 X8660 H39 76 20054 I246 77 21657 X_54 78 23523 lO62 79 _5723 972 8o 28356 881 81 31569 79 z 82 35577 7o2 83 40222 613

288 THE TWONEW SCIENCES F GALILEO O

Angleof Elevation Altitude 39 4049 40 4196 4x 4346 4z 45oz 43 4662 44 4828 45 5oo Sublimity 6x74 5959 575z 5553 5362 5x77 5oo Angleof Elevation Altitude Sublimity " 84 4757z 525 85 5715o 437 86 7x5o3 349 87 95405 z6z 88 14318I I74 89 286499 87 9o infinita

PROPOSITIONIV X

[308]

Tofind foreach degree felevation o thealtitudes andsublimities ofparabolas ofconstant amplitude. Theproblem easily is solved.Forif weassume constant a amplitude IOOOO, of thenhalfthe tangent t anyangle a of elevation illbethealtitude.Thus,to illustrate,parabola w a havingnangle felevation andanamplitude xoooo, a o of3o of willhave analtitude fz887, hich o w isapproximately one-half the tangent.Andnowthe altitude aving h beenfound,he t sublimity isderivedsfollows. incet hasbeen a S i proved that halftheamplitude semi-parabola ofa isthemean proportional between thealtitude ndsublimity, a andsincehe altitude as t h already found, been andsince thesemi-amplitude isaconstant, namely it follows 5oo% thatifwedividehesquare t ofthesemiamplitude bythealtitude eshall w obtain thesublimity sought. Thus inourexample thealtitude asfound bez887:the w to square of5ooo is25,ooo, which ooo, dividedy _887 b givesthe approximate ofthesublimity, value namely 8659. SALV. Herewesee,firstof all,howverytrueisthe statementmade above,hat,fordifferent t angles felevation, o the greater thedeviation themean, from whether orbelow, above the greaterhe initialspeed t [impeto violenza] e required to carry theproje_ile thesame over range. orsincehespeeds F t i the resultant two motions, of namely, onehorizontal and uniform, othervertical ndnaturallyccelerated; the a a and sincethe sumof the altitude andsublimity represents this speed, isseenfromthe preceding thatthissumis a it table minimum

FOURTH DAY _89 minimum an elevation 45 wherethe altitudeand subfor of ]imit_ are equal,namely,each 5ooo;and their sum ioooo. But ifwechoose greater a elevation, ,weshall indthealtisay5o f tude5959, ndthesublimity 196giving sumo{1o155; like a 4 , a in mannerweshallfindthat thisispreciselyhevalueofthespeed t at 40elevation, both anglesdeviating equally fromthemean. Secondly is to be notedthat, whileequalspeedsare reit quiredforeachof twoelevationshat areequidistant t fromthe mean,thereisthiscurious lternation, a namely,hatthealtitude t and sublimity the greaterelevation at correspond inversely to the sublimity altitudeat the lowerelevation.Thusin the and preceding example n elevation 5o gives altitudeof5959 a of an and a sublimity 4196;whilean elevation 4ocorresponds of of to an altitudeof4196and a sublimity of5959-Andthis holds true in general; ut it is to be remembered in orderto b that, escapetediouscalculations, o accounthas been taken of n fra_ionswhichare of littlemomentin comparison ith such w largenumbers. SAc_.I note alsoin regardto the two components the of inkialspeed[irnpeto] the higherthe shotthe lessis the that horizontal thegreaterthe vertical omponent; and c ontheother hand,at lowerelevations herethe shotreaches w onlya small heightthe horizontal omponent the initialspeedmust be c of great. In the caseof a projecCtile firedat an elevation of9o, I quiteunderstand allthe force[forza]n theworld that i would notbesufficient tomakeitdeviatea singleinger's readth{rom f b the perpendicular that it wouldnecessarily backinto and fall its initialposition; ut in the caseofzeroelevation, b whenthe shotis firedhorizontally, am notso certainthat someforce, I lessthaninfinite, ould w notcarrytheproje_ilesome distance; thUS evena cannoncanfirea shotin a pedecdy orizontal not h dired_don, oraswesay,pointblank,that is,withnoelevation at all. HereI admitthereissome room fordoubt. Thefacet do I not denyoutright,because anotherphenomenon of apparently no lessremarkable, ut yet onefor whichI have conclusive b evidence. Thisphenomenons the impossibilityf stretching i o
a

bog]

290 THE TWO NEW SCIENCESOF GALILEO a ropeinsuchawaythat it shallbeat oncestraightandparallel to the horizon; he faccsthat the cordalways t i sagsand bends and that no forceis sufficienta stretchit peffecCtly t straight. SALV. this caseof the ropethen, Sagredo, ou ceaseto In y wonder t thephenomenon a because ouhaveitsdemonstration; y but if we considert withmorecarewemaypossibly i discover some correspondence between thecaseofthegunandthatofthe string. The curvature thepathoftheshotfiredhorizontally of appears toresultfromtwoforces, one(thatoftheweapon) drives it horizontallyndtheother(itsownweight) rawsitvertically a d downward.Soin stretching theropeyouhavethe force which pullsit horizontallynd itsownweightwhicha&sdownwards. a Thecircumstances thesetwocases in are,therefore, verysimilar. If then you attribute to the weightof the ropea powerand energy[possanza edenergia] sufficient opposeand overcome to any stretchingforce,no matter howgreat, why deny this power the bullet to ? Besides musttell you something hichwillboth surprise I w and pleaseyou, namely,that a cord stretchedmoreor less tightlyassumes curve a whichclosely pproximates a theparabola. Thissimilarity isclearly seenif youdrawa parabolicurveona c verticalplaneandtheninvertit sothat theapexwilllieat the bottom and the base remainhorizontal; on hanginga for, chainbelow base,oneendattached eachextremity the to ofthe base,youwillobserve that,onslackening thechain moreorless, it bendsand fits itselfto the parabola; nd the coincidence a is moreexaccn proportion the parabola drawnwith less i as is curvature or,so to speak,morestretched; o that usingparabs olasdescribed with elevationsessthan 45 the chainfits its l parabola almostperfe&ly. SAQR. Then witha finechainonewouldbe ableto quickly drawmanyparabolicines l upona planesurface. SXLV. Certainly andwithnosmalladvantage I shallshow as youlater. SIMv.Butbeforegoing, further,I amanxiouso be convinced t at leastof that proposition whichyou say that there is a of
[310]

rigid

FOURTHDAY 29I rigiddemonstration;referto thestatement it isimpossible I that by anyforcewhatevero stretcha cord8othat it willliepert fecCtly straightandhorizontal. SAcR. willseeifI canrecall hedemonstration; inorder I t but to understand Simplicio, willbenecessaryoryouto take it, it f forgrantedconcerning machines hatisevident otalonefrom w n experiment ut alsofromtheoretical onsiderations, b c namely, that the velocity a moving of body[velocith delmovente], even whenits force[for_] is small,canovercome verygreatrea sistance xerted slowly oving e bya m body,whenever thevelocity ofthemoving bodybearsto that ofthe resistingodya greater b ratio than the resistance [resisten_a] the resisting of bodyto theforce[forza] ofthemoving body. SIMP. ThisI know verywellforit hasbeendemonstrated by Aristotle his _uestionsn Mechanics; is alsoclearlyseen in i it in the leverand the steelyard wherea counterpoise weighing notmorethan4 pounds willlift a weightof4ooprovided that the distanceof the counterpoiseromthe axis aboutwhich f the steelyard rotatesbe morethanonehundredtimesasgreat as the distancebetween axisandthe pointof supportfor this [3II] the largeweight. This is truebecause counterpoise its the in descenttraverses space a morethanonehundred timesasgreat as that movedoverby the largeweightin the sametime;in otherwordsthe small ounterpoise c moves itha velocity hich w w is morethan one hundredtimesas greatas that of the large weight. SAcg.Youarequiteright;youdonothesitate admitthat to however mallthe force[forza] fthemoving s o bodyit willovercomeany resistance,however reat,providedit gainsmore g in velocitythan it losesin forceand weight[_igore gravi_]. e Nowletusreturnto the case ofthecord. In theaccompanying figure abrepresents linepassing a through twofixed pointsa and b; at the extremities f this linehang,as you see,two large o weights and d, whichstretchit withgreatforceand keepit c truly straight,seeingthat it is merelya linewithoutweight. NowI wishto remarkthat iffromthe middle pointofthisline, which

z9z THE TWO NEW SCT_.NCES F GALTt._D O whichwemaycalle,you suspend :anysmallweight, ay h,the s lineabwillyieldtoward point andonaccount fitselongathe f o tionwillcompelhe twoheavyweights and d to rise. This I t c shall demonstratesfollows: a withthepointsa andb ascenters describehetwoquadrants, igandelm; t e nowsince thetwosemidiameters iand blareequalto aeandeb,theremainders and a fi fl aretheexcesses ofthelinesaf andfboveraeandeb;theythere-

....

t'

Fig.xe6 foredetermineheriseoftheweights andd,assumingfcourse t c o that the weighth has takenthe position. But the weighth f willtake the position wheneverhe lineefwhichrepresents f, t the descentof h bearsto the linefi--that is,to the riseof the weights and d--a ratiowhichis greaterthan the ratioof the c weight fthetwolargebodies thatofthebodyh. Evenwhen o to theweights fc andd areverygreatandthat ofhverysmallhis o t willhappen; ortheexcess ftheweights anddovertheweight f o c ofh canneverbe sogreatbut that the excess fthe tangentef o overthe segmenti maybe proportionallyreater. This may f g be

[zIz]

FOURTHDAY 293 be provedas follows: Drawa Circle f diameter o gai;drawthe linebosuchthat the ratioof its lengthto anotherlengthc, c>d,isthe sameasthe ratiooftheweights anddto theweight c h. Since c>d,theratioofboto d isgreaterthanthat ofboto c. Takebea thirdproportional obandd;prolong diameter to the gi to a pointf suchthat gi:if=oe:eb; fromthe pointf draw and the tangentfn; then sincewe alreadyhave oe:eb gi:if, we = ShallObtain, y compounding b ratios,ob:ebif:if. But d is a = meanproportional between obandbe;whilenf is a meanproportionalbetween andfi. Hencenf bearsto fi the same if ratioasthat ofcbtod,whichisgreater thanthat oftheWeights c and d to the weighth. Sincethenthe descent, r velocity, o ofthe Weight bearsto the rise,or velocity, f the weights h o c andd a greaterratiothantheweight fthebodiescandd bears o to the weightof h, it is clearthat the weighth willdescend and the lineabwillceaseto be straightandhorizontal. Andnowthiswhichhappensin the caseof a weightless cord abwhenany smallweight/_ isattachedat thepointe,happens als6whenthe Cordis madeof ponderable atterbut without m any attachedweight;because thiscasethematerial fwhich in o the cordis composed functionssa suspended eight. a w S_a,. I am fullysatisfied.So nowSalviaticanexplain, s a he pi_mised,the advantage sucha chainand, afterwards, of presentthe speculations ourAcademician the subjectof of on impulsiveorces f [forr_ _ercossa]. della SAa_v. the preceding Let discussionsuffice to-day;the s for hourisalreadylateandthetimeremaining illnotpermitusto w clearupthe subjeccs proposed; wemaytherefore postpone our meeting untilanotheraridmoreopp6rmne ccasion. o SAoR. concurin youropinion, ecause I b after variousconversations ithintimate w friends ofourAcademicianhaveconI cluded thisquestion that ofimpulsive forces isveryobscure, and I thinkthat, upto thepresent, oneofthose n whohavetreated

[313]

this subjecthavebeenableto clearup.its darkcorners which liealmostbeyondthe reachof humannnagination; mong a the various views whichI haveheardexpressedne,strangely o fantastic

294 THE TWONEWSCIENCES F GALILEO O tastic, emains r inmymemory, namely, impulsive are "that forces indeterminate, ifnotinfinite. etus,therefore, theconL await venience Salviati.Meanwhile of tellmewhatis thiswhich follows thediscussion ofproje_iles. S_v. These aresome theorems pertaining thecenters f to o gravity ofsolids, iscovered d byourAcademician inhisyouth, andundertaken byhimbecause heconsidered thetreatment of Federigo Comandino tobesomewhat incomplete. Thepropositionswhich youhavebefore ouwould, thought, the y he meet deficiencies Comandino's of book. The investigation was undertakent the instance fthe Illustrious a o Marquis Guid' Ubaldo DalMonte, verydistinguished a mathematician ofhis day,asisevidenced byhisvarious ublications. p TothisgentlemanourAcademician copy gavea ofthiswork,hopingo ext tendtheinvestigation toother solids ottreated n byComandino. Butalittlelatertherechanced tofallintohishandshebook t of the greatgeometrician, LucaValerio, where hefound subthe je&treated socompletely thatheleftoff: hisown investigations, although themethods hich w heemployed werequitedifferent from ofValerio. those SAcg. lease P begood enough leavethisvolume ithme to w untilournextmeetingothatI maybeableto readandstudy s these propositions intheorderinwhichheyarewritten. t SALv. isapleasure comply ithyourrequest only It to w andI hope thatthepropositions beofdeep will interest toyou.
END OF FOURTHDAY.

APPENDIX
Containing some theorems, and their proofs, dealing with centers of gravity of solid bodies, written by the same Author at an earlier date.* * Following exampleof the National Edition,this Appendixwhich the coversI8 pagesof the LeydenEditionof I638is hereomittedas being of minorinterest. [Trans.] [FINIS]

INDEX
Accelerandosl = displace, 52 2 Beamcarryinga constantload Accelerated motion,x6o whichmoves fromoneend to reduced to theother,I4O uniform motion,x73 Beamcarryinga constantload Acceleration, definitionof uni- in a fixed position, 4oet $q. 1 form,I69 Beams,similar, elativestrength r Acceleration gravitymeasured, of, I24 of I78 _, strength of,II5 Accelerazione, in the sense used , variation fstrength o with ofspeed, 67 I diameter,I9 I Achilles nd the Tortoise,proba , variation fstrength o with lemof,x64 length,I23 Acoustics, 95 Bendingtrength,II9 s Adhesionf plates,xI, IZ o Bernoulli, James,I49 Aggiunti, x , John,I83 Air, resistance to projectiles,Bones, roportionf, I3X of, p o 252,256 , ofbirds,I5o,239 Altitudeof parabola,determina-Brachistochrone, 225,239 tion of,285 Amplitudeof paraboladefined, Canne fathom = Capillarity, seenin shorteneffect 26o,273 tablesof, 284 ingofrope,2o Apollonius, 242,245,246 as seenin dropson theoremof,38,44 cabbageeaves, o l 7 Aproino, Paolo, x i Catenary, I49,29o Arcetri,xi,xviii Cause,speedconsidered a, 255 as Archimedes, 4I, IiO, I44,I47, Causes, v, searchafter,futile,I66 242,25x Cavalieri, 4I Archimedes' principle, 8I Chords ofcircles, timesofdescent Ariosto, uotation q from,I3I along,I88etsq. Aristotle, 2, I3,2o,26,49,64,76, Cicero, I xix 77,80,95,IIO,I25,I35 29I Cohesionf solids, I etsg.,I8 , o I Aristotle'sdeaoffalling i bodies, Comandino, 6I Federigo, 294 Arrighetti, x Compass, Galileo's military, 49 I Atmospheric pressure, 7 I Composition ofmotions, 244,257, Augmented Fourth,Io4 267 Axioms uniform of motion,I54 Continuity, mathematical, 24etsq.

298

INDEX

Continuous quantities, 34-36 Gra_ty,centers 3I, of,294,295 Contraction, theoryof, 51 .,experimental determinaCylinders,relative strength of tionof,178 similar,Iz4 Guevara (di)Giovani, I26 zo, Definitionfuniformly o accelerated Hammerblow,problem 27i of, motion, 61 I Heath,T. L., v, I45 Definition funiform o motion,x54 Huygens, v Definitions, mathematical, 28 Impenetrability matter,49,5I, of Density,seeSpecific ravity G 6I Diapason, 99 Diapente, 99 Impulsiveorces, 93 f 2 Diminishedifth,Io4 F Inclinedplane,principle I84 of, Diodati, lia,ix,xii E etsq. --, usedto "dilute'" Energia, 269,27I,290 gravity, 78 1 Elzevir, ouis,xi L Inclinedplanesof equalheight, Elzevirs, The,xviii speed acquired on,I69,184 Errors in parabolic trajectories, Indivisibility, 30,36 25Ietsq. Infinity, 6,30,3I, 37,39_ 2 Euclid,242,243,246,248,283 Isochronism the simplepenof Expansion, theoryof,51 dulum, 7 9 of pendulums f difo Fallingbodies,Aristotle's notion ferentmaterials, 5 8 of,6I, 64-68 Isoperimetric problems, 58 --., lawsof,265 --; sphere locus Jealousy investigators, as of, of 83 I9Z,I93 Featherandcoinresultpredicted, Lawoflever,II2 72 Lawofmotion, ewton'sirst,244 N f Fifth,a musical interval, oo i Newton'ssecond, Fire escape, evisedby kinsman 25o d ofGalileo, o I Lawofsimple pendulum, 96 Force, synonyms 1I4,267,269, Lawsofairresistance, for, 253 27I 286,29I, 293 , Lawsof falling bodies,I74et sq., Forza, n the senseof mechanical 265 i advantage, 24 1 Lever,lawof, II2 .,inthesense ofmomentum,Levityofair,77,78 267,269,271,286 Lift pump, heoryof,I6 t Frequenza frequency pen- Light,speedof,42 = of dulum, 7 9 Limiting speedof bodiesn vacuo, i 7z Geometrycompared with Logic, Lined,Academy ofthe,xx I37 Loaded eams,I4o b Gilding, thickness f, 53etsq. o Logic andgeometryompared, c I37

INDEX
Marsili,x Mathematicaldefinitions, 28 Mathematics,r61eof in physics, 276 Maximumrange projectile, 75 ofa 2 Media,effectof,on speedoffalling bodies,76 Medici,Antoniode', x , Prince Mattia de', xi Mersenne,M, v. Micanzio,xi Minimum momentum theorem, 273 Mirrors,spherical,41 Mole= volume,8o, 82 Moment,used in sense of importance,65 Momento-- force,183 Momentoin the sense of magnitude, 124 Momentum, 258, 259, z63, z66; z69,270 , measureof, 264 Monochord,99 Motion, axiomsof uniform,154 , natural,153 of fallingbodies,16o of projectiles,244 supernatural,255 uniform,lawsof, I54 Monte, G. del, ix,294 Motte, v Music,95 Naturally acceleratedmotion,16o "Natural" Motion, 153 Newton, v Newton's First Law of Motion, 215,216,244 SecondLaw of Motion, 250,257,263 Noailles,Count of, xii, xvii Octave,ratio involvedin, 99 Orbitsof planets, 261

299

Padua, ix,x, xii, 182 Parabola,description 148,246, of, 258 , quadratureof, 147 --., to find the sublimityof a given,272 Parallelogram velocities, 57 of 2 Pendulum motion, law of air resistance 254 to, Pendulums lead and cork,84 of Pendulums, ime of descent,95 t Pendulum with string striking againstnail,17o Periodo= periodof pendulum,97 Peripatetics, 48,61 Peso= specific gravity,69,72 Pieroni,G., xl Planetary orbits,261 Plato, 9o, 137,261 Point blank firing,289 Potenza,meaning of"in potenza," 257,266 Principleof virtualwork,183 Problem of Achilles and the Tortoise,164 of stretchedrope,290 Projectile,speedof, 279 Projectiles,motionof, 244 Pythagoreanproblem music,95, in lO3 Quadratureof parabola,147,251 Quickest escent,225,239 d Range,maximum,275 Range,minimum momentum requiredfor a given,273 Range tables, 2.56, 84 2 Resistanceof mr proportionalto speed,74 of air to projectiles, 252 Resonance f a bell,98 o of pendulum,97 of strings,99

3oo INDEX Resultant,257,258, 267 Strings,lawsof vibrating,Ioo Ripples, produced hy sounding "gul_limlty" de_ned,z60 a goblet,99 of any parabola, to Rope-making,heoryof, 8 et sq. t find,272 Ropes,shortening dueto moisture, Subsidence ofsmallparticles,87 20 _, stretched,problemof, z9o Tangent, used in two senses,28z Terminal speed, 9z, 256, 26o Sack,properdesignof, 56 Theoremof Apollonius,38, 44 Sacrobosco (John Holywood) 7 5 Tritono,lO4 Salusbury,Thomas,v Tubes,bendingstrength of, I5O Sarpi,Fra Paolo,ix, x Secondminute, as a unit of time, Ubaldi,ix z64, 265 Uniform acceleration, definition Siena,x of, 169 Space,describedin uniformlyac- Uniform motion,lawsof, 154 celeratedmotion,173et sq. Spearas unit oflength, 265 Vahrio,Luca, 30, 148,z94 Specificgravity, halls adjusted to Velocities,parallelogram 257 of, equal that of water, 69 Velocity,standardof, 26o Specificgravity of air, 78 tt sq. , uniform, definitionof, of aquatic 154 animals,13z , virtual, z9I Speedof projectile,z79 Venice,xi , standard of, 260, 265 Figore= force,z91 , terminal,9I Vinta, Belisario,x uniform,definitionof, 154 Virtual work, principle of, z9z , Statics, fundamentalprincipleof, Viviani,Vincenzio,x, I8o i
IIO

Steelyard,29I Strengthof materials,lO9

.Weighingin vacuo,81-83 Weston,Thomas,vi